CAD/CAM Industry Service August 13 and 14, 1984 Boston Marriott Buriington Hotel Buriington, Massachusetts

THIS PRESENTATION WAS NOT AVAILABLE AT PUBLICATION TIME.

IF A COPY IS MADE AVAILABLE TO DATAQUEST, WE WILL MAIL IT DIRECTLY TO YOU FOLLOWING THE CONFERENCE.

V D^aiquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 THIS PRESENTATION WAS NOT AVAILABLE IN TI.^E FOR PUBLICATION.

A COPY WILL BE MAILED TO YOU WHEN DATAQUEST RECEIVES IT FROM THE SPEAKER.

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Focus Conference ^^Workstations^^ The Evolution of an Industry

Sponsored by the CAD/CAM Industry Service August 13 and 14, 1984 Boston Marriott Buriington Hotel Buriington, Massachusetts

^-'•-=-« H;)aDataquest

1290 Ridder Park Drive San Jose, California 95131 (408) 971-9000 Telex: 171973

Sales/Service offices:

UNITED KINGDOM GERMANY DATAQUEST UK Limited DATAQUEST GmbH 144/146 New Bond Street In der Schneithohl 17 London WIY 9FD D-6242 Kronberg 2 United Kingdom West Germany (01) 409-1427 (06173) 6921 Tfelex: 266195 Telex: 410939

FRANCE JAPAN DATAQUEST SARL DATAQUEST Japan, Ltd. 41, rue Ybry Azabu Heights, Suite 711 92522 Neuilly-sur-Seine Cedex 1-5-10, Roppongi, Minato-ku France Tokyo 106, Japan (01) 758-1240 (03) 582-1441 Telex: 630842 Telex: J32768

The content of this report represents our interpretation and analysis of infor mation generally available to the public or released by responsible individuals in the subject companies, but is not guaranteed as to accuracy or completeness. It does not contain material provided to us in confidence by our clients. This information is not furnished in connection with a sale or offer to sell securities, or in connection with the soUcitation of an offer to buy securities. This firm and its parent and/or their officers, stockholders, or members of their families may, from time to time, have a long or short position in the securities mentioned and may sell or buy such securities. Printed in the United States of America. All rights reserved. No part of this publication may be reproduced, stored in retrieval systems, or transmitted, in any form or by any means—mechanical, electronic, photocopying, duplicating, microfilming, videotape, or otherwise—without the prior written permission of the publisher.

FOCUS CONFERENCE AGENDA "Workstations" The Evolution of an Industry Sponsored by the CAD/CAM Industry Service August 13 and 14, 1984 Boston Marriott Burlington Hotel Burlington, Massachusetts

MONDAY, AlUGUS T 13 TUESDAY, AUGUST 14

8:30 a.m. Workstation Industry Overview 8:30 a.m. Network Configurations James Newcomb, Director Peter Shaw, President Design and Manufacturing Automation Group Syte Technologies Dataquest Incorporated 9:20 a.m. Start-up: Building the 9:20 a.m. Architectures, Graphics, and Performance Management Team David Burdick, Industry Analyst Arthur Campbell, President CAD/CAM Industry Service Mosaic Technologies Dataquest Incorporated 10:10 a.m. Coffee Break 10:10 a.m. Coffee Break 10:30 a.m. Product Mix: The Integration of 10:30 a.m. Workstations, The Transfer of Technologies Internal and External David Nelson, Senior Vice President Technologies Apollo Computer Corp. Jerry Rotds, Sr. Vice President Sales and Marketing 11:30 a.m. Lunch Gould Incorporated

1:00 p.m. The Role of VLSI Technology in the 11:20 a.m. Marketing: The OEM Approach Workstation Marketplace Donald McDougall Vice Presiden James Clark, Chief Technical Officer and General Manager Silicon Graphics Incorporated Data General Corporation

1:50 p.m. The Role of Proprietary Architectures 12:10 p.m. Lunch in Workstatton Design David Folger, President 1:30 p.m. Workstations: Flexible Systems Ridge Computer Company Architecture in a High-Performance 2:40 p.m. Systems Architecture- Graphics Environment Utilizing Standards Dennis Peck, President Scott McNeally, President Saber Technology Sun Microsystems 2:20 p.m. Display Technologies: 3:30 p.m. The Utilization of Artificial Intelligence Impact on Workstations in Developing User Interfaces and Richard Fichera Next Generation Workstations Director of Marketing Richard Rifenburgh Raster Technologies Incorporated Chairman of the Board/CEO Perq Systems Corp. 3:10 p.m. Application-Specific Hardware— A Component Approach 4:20 p.m. Communications integration Arthur Collmeyer, President Thomas Bredt, Vice President and Director Weitek Corporation Telecommunications Industry Service Dataquest Incorporated 4:00 p.m. Conference Wrap-Up James Newcomb, Director 5:30 to Design and Manufacturing 7:00 p.m. Cocktail Party Automation Group Dataquest Incorporated

Dataquest Incorporated, A Subsidiary of A.G. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dataquest

CAD/CAM FOCUS CONFERENCE EVALUATION QUESTIONNAIRE Buriington, IVIassachusetts August 13-14, 1984

Thank you for attending our CAD/CAIVI Industry Focus Conference. Would you please assist us in planning our next conference by completing and returning this questionnaire?

1. Please rate each presentation on a scale of 1 to 10 (where 10 is highest in terms of your approval):

CONTENT DELIVERY COMMENTS (1 to 10) (1 to 10) (Use reverse side if necessary)

Newcomb, Workstation Industry Overview Burdick, Arcliitectures, Graphics, and Performance Nelson, Wbrlcstations, Transfer of Technologies Clark, Role of VLSI Technology in the Workstation Marketplace Folger, Role of Proprietary Architectures in Workstation Design McNeally, Systems Architecture—Utilizing Standards Rifenburgh, Utilization of Artificial Intelligence in Developing User Interfaces and Next Generation Workstations Bredt, Communications Integration Shaw, Network Configurations Campbell, Startup: Building the Management Team Rotds, Product Mix: Integration of Internal and External Technologies McDougall, Marketing: The OEM Approach Peck, Workstations: Flexible Systems Architecture in a HIgh-Performance Graphics Environment Fichera, Display Techndogies: Impact on Workstations Collmeyer, Application-Specific Hardware — A Component Approach

2. How would you rate the conference facilities (1 to 10)? Location Guest Rooms Meals Meeting Rooms Recreational Facilities

3. How would you rate the service quality and staff attitude (1 to 10)? The hotel staff DATAQUEST staff

(over)

Dataquest Incorporated, A Subsidiary of A.G. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 4. What did you like most about the Focus Conference?.

5. What did you like least about the Focus Conference?.

6. Suggestions for improving the Fbcus Conference:

7. Topics that would be of interest to you for the next Focus Conference:

8. Do you prefer longer, shorter, or the same length conference?

9. Do you prefer more, less, or the same amount of free time?

10. Your primary interest in the display and graphics terminal industry is as a: Manufacturer Service Vendor User Financial Analyst Other (Please Specify)

Name and Company (optional)

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dataquest

CAD/CAM INDUSTRY CONFERENCE

August 13 and 14, 1984

Burlington, Massachusetts

Final List of Attendees

AT&T Bell Labs Andy Nagle Supervisor

Basant R. Chawla Department Head

Adage, Inc. Clarence A. Tilger Director of Market Planning

Air Products and Chemicals Timothy Pickering Technology Assessment Analyst

American Micro Systems Gary Miles CAD Planning Manager

Analog Design Tools, Inc. Richard N. Lanham Vice President, Marketing & Sales

Apollo Computer David Nelson Senior Vice President

Debra Sheedy

John Cencioso

Lou Reynolds Product Group Manager

Stephen Weinberger Director Market Development

Win Burke Director of Product Marketing

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Applicon Francis DonIan Director of Engineering

Jeffery Simon Manager Systems Marketing

Richard Miller Director of Marketing

Arthur Young and Company George Southwick Partner

Bank of America Al Schumer Assistant Vice President

Dave Woonton Assistant Vice President

Bishop Graphics Loren T. Bersack Director Product Development

Boeing Aircraft Corporation Richard Hedlund Electronic CAD Manager

Burr, Egan, Deleage & Co. Esther B. Sharp Associate

CAD/CAM Specialists, Inc. Dan Garms Vice President

Michael W. McDonald President

CIE Terminals Roger Budris Manager Advanced Planning

CPT Corporation Pamela A. Jensen Senior Market Research Specialist

Cadmus Computer Systems Jeff Singer Vice President, Engineering

CalComp Craig Montgomery Product Line Manager

Mike Watterson Product Line Manager

Wendy Abramowitz

Janifer Jones

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Cericor Incorporated David Bailey President

William Shea Regional Sales Manager

Ciprico Incorporated Bruce Brunette Director, Marketing

Computervision Leslie Swanson Product Specialist

Robert Horn

Robert Resslhuber Product Specialist

Arnie Epstein Director, Systems Architecture

Hank Kellogg Marketing Analyst

James Cook Director of Technical Research

Jitendira Singh

Mark Krentzman

Neil Sawe

Owen Knox

Stephanie Marrus

Data General Corporation Donald McDougall Vice President & General Manager

Dataquest Incorporated Beth Tucker Industry Analyst

Bob Smith Marketing Manager

Brad Smith Associate Director

Dataquest incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Chuck Kanupke Vice President TCIS East Coast

Dave Burdick Industry Analyst

Don Johnson Research Associate

Eileen Earth Research Associate

Frank Florence Marketing Manager

H. Edward White Associate Director OAS

Irene Barth Librarian

Jim Newcoinb Director

Kelly Leininger Research Clerk

Pamela Shook Group Administrator

Thomas Bredt Vice President and Director

Wendy Ledamun Research Associate

Diconix Ernest Truax Manager Product Planning Digital Equipment Corporation Barry Reynolds Finance Business Manager

Dorothy Rado Business Manager Marketing

John Clarke

John Dimack Engineering Operations Manager

Libby Aston Engineering Supervisor

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Nick Johnson Technologist

Norm Tripp Product Manager

Prakash Bhalerao Engineering Group Manager

Roslyn Morvay Product Marketing Manager

Tim Yeaton Finance Manager

Susan Marie Marketing Analyst

Eastman Kodak Company John C. Batterson Planning Specialist

Edge Computer Corporation J. Douglas Crawford Director, Sales

Electronic Business Cindy A. Thames Senior Editor

Evans and Sutherland Edward C. Cheadle CAD/CAM Specialist

Fairchild John Andrews CAD Staff Engineer

GenRad Incorporated S. Russel Craig Director, Corporate Marketing

General Computer Company Kevin Curran President

Pax Goodson Marketing Manager

General Electric Company Dale W. Rowe General Manager 66TD

James Whitten Manager Design Engineering

Marshall Kidd Manager Technical Analyst

Robert W. Baeder St3ff Executive

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Gould Incorporated Dana A. Gauthier Manager Planning & Financial Services

Joseph Heaney

Hewlett-Packard Billy Abrams

Alan Benjamin Product Marketing Engineer

Jim Lyons Product Marketing Manager

IBM A. C. Ewing Graphics Adminstrator

L. Wilcox Product Planner

Richard Ma Senior Marketing Analyst

Steve Vaughn Advisory Engineer

ICM, Incorporated Brian W. Kirschner Vice President, Research & Development

Dr. Ii. L. Barinka Preisdent

IDS Growth Spectrum Advisors Michael F. Guilford Senior Analyst

Iropell Corporation Pat Strange Director

Intel Corporation Sudhir Bhagwan Assistant General Manager

Jennison Associates David T. Poiesz Vice President

Jupiter Systems, Inc. Jon G. Peddie President

LSI Logic Corporation Richard Perry Software Marketing Manager

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Lexidata Corporation Forrest Gasik Manager, Forecasting

Shelly Lowent±al Engineering Manager

Loomis, Sayles & Company Jaye E. Morency Securities Analyst

Matra Datavisicm John Keller Manufacturing M2mager

Mayfield Fund Grant Heidrich General Partner

McAuto Jeff Barnell Hardware Product Manager

Megatek Corporation Greg Elmassian Senior Engineer

Susan Harvey Corporate Planning Manager

Mosaic Technology Arthur Campbell President

David S. Lobel Board Memeber

Motorola Incorporated Arthur R. Miller Strategy and Bus. Dev., High End MPU

National Semiconductor John Payne Marketing Manager

N. T. Hsu Senior Engineering Manager

Nixdorf Lorraine Chouinard Product Line Manager

Oak Management Corporation Carla Haugen Associate

Olivetti Research & Development Charles Erickson Manager

Omnicad Corporaticm Edward J. Morton

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Pacific Telesis Group Bill R. Frazer Director

Paragon Technology Philip Monego President

Parallax Graphics Jim A. Summers President

Perq Systems Richard Cox Director, Marketing Planning

Jim Gay Vice President, Product Management

Prime Computer Corpocaticm Barry Wolman Director, Systems Development

Celia Bland Manager Workstation Marketing

Edward Vravlick Systems Consultant

Marvin Miller Product Integration Manager

Prudential Bache Securities Laura Conigliaro Vice President

Christine Chien

Michael Weisberg Director, Technology Research

Pyramid Technology Corporation Frank Madren Vice President, Marketing

William Gimple Vice President, Research and Dev.

RCA J.W. Scherer Manager Product Marketing

Racal-Redac R. Davis

Ray Spencer

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Pari< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Raster Technologies Richard Fichera Director o£ Marketing

Resolutions Systems Incorporated Isaac Nassi Vice President

Ridge Computer Company David Polger President

Harry Taxin Vice President, Marketing & Sales

Saber Technology Dennis Peck President

Don Rector

Sanders Associates Patty Craig Software Engineer

Ronald Derby Corporate Marketing Research Manager

Siemens Aktiengesellschaft-Dep E TDV Wolfram Howein Dipl. - Ing

Silicon Graphics Incorporated Craig S. Olson Product Manager

Jim Clark Chief Technical Officer

Vernon Anderson Chairman

Spectragraphics John Moreland Director, Marketing

Ray Barger Vice President, Marketing

Sun Microsystems Scott McNeally President

Syte Information Technologies Peter Shaw President

T. Rowe Price Michael Christodolou

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Teradyne Phil Odence Applications Engineer

Terak Corporation Tom Zender Director of Marketing

Wang Labs Dan Miley

Gary Neil Director, Industry Marketing Programs

Leigh Herndon

Wayne Kerr Ken Berquist Vice President

Ted Sisley Product Manager

Weitek Corporation Arthur Collmeyer President

Lynette DeNike Consultant

Dataquest incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dataquest

CAD/CAM Focus Conference

August 13 and 14, 1984

Burlington, Massachusetts

List of Attendees

AT&T Bell Labs Basant R. Chawla Department Head

Adage, Inc. Clarence A. Tilger Director of Market Planning

American Micro Systems Gary Miles CAD Planning Manager

Analog Design Tools, Inc. Richard N. Lanham Vice President, Marketing & Sales

Apollo Computer Anil Gadre

Apollo Computer David Nelson Senior Vice President

Apollo Coii5>uter Linda Sartes

Apollo Computer Lou Reynolds Product Group Manager

Apollo Computer Stephen Weinberger Director Market Development

Apollo Con^uter Win Burke Director of Product Marketing

Appliccm Francis Donlem Director of Engineering

Applicon Jeffery Simon Manager Systems Marketing

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Applicon Richard Miller Director of Marketing

Arthur Young and Company George Southwick Partner

Bank o£ America Al Schumer Assistant Vice President

Bank of America Dave Woonton Assistant Vice President

Boeing Aircraft Corporation Richard Hedlund Electronic CAD Manager

CAD/CAM Specialists, Inc. Dan Garms Vice President

CAD/CAM Specialists, Inc. Michael W. McDonald President

CAE Terminals Roger Budris Manager Advanced Planning

Cadmus Computer Systems Jeff Singer Vice President, Engineering

CalComp Craig Montgomery Product Line Manager

CalCon^ Mike Watterson Product Line Manager

CalComp Rich Peters Vice President, Business Development

Calma Company James Mannos Director, Systems Engineer

Calma Company Tom Teeters Director of Corporate Marketing

Calma Company Van Lewing Vice President, Electronics

Calma Company William Blasdell Vice President System Programs

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Century Capital Assoc. Arthur Vitolo

Ciprico, Incorporated Bruce Brunette Director, Marketing

Computervision Corporation Arnie Epstein Director of Systems Architecture

Conputervision Corporation Hank Kellogg Marketing Analyst

Computervision Corporation James Cook Director of Technical Research

Computervision Corporation Mark Krentzman

Computervision Corporation Matt Peterson Director, Systems Marketing

Computervision Corporation Michael Barry Director, Workstation Systems Dev,

Computervision Corporation Stephanie Marrus

Counterpoint Ccwiputer Pauline Alker President

Cybernex James B. Money Vice President, Marketing

Data General Corporation Donald McDougall Vice President & General Manager

Dataquest Incorporated Beth Tucker Industry Analyst

Dataquest Incorporated Bob Smith Marketing Manager

Dataquest Incorporated Brad Smith Associate Director

Dataquest Incorporated Chuck Kanupke Vice President TCIS East Coast

Dataquest Incorporated Dave Burdick Industry Analyst

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dataquest Incorporated Debra Jacob Secretary

Dataquest Incorporated Eileen Earth Research Associate

Dataquest Incorporated Frank Florence Marketing Manager

Dataquest Incorporated H. Edward White Associate Director OAS

Dataquest Incorporated Irene Earth Librarian

Dataquest Incorporated Jim Newcomb Director

Dataquest Incorporated Kelly Leininger Research Clerk

Dataquest Incorporated Pamela Shook Group Administrator/Reserach Assistant

Dataquest Incorporated Sharlene Burden Sales Support Rep.

Dataquest Incorporated Thomas Bredt Vice President and Director

Dataquest Incorporated Wendy Ledamun Research Associate

Diconix Ernest Trvax Manager Product Planning

Digital Equipment Corporation Barry Reynolds Finance Business Manager

Digital Equipment Corporation Chan Yan Chow Engineering Manager

Digital Equipment Corporation Cheryl Vedoe Product Manager

Digital Equipment Corporation Dorothy Rado Business Manager Marketing

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Digital Equipment Corporation John Clarke

Digital Equipment Corporation John Dimack Engineering Operations Manager

Digital Equipment Corporation Ken Bcabitz Planning Manager

Digital Equipment Corporation Libby Aston Engineering Supervisor

Digital Equipment Corporation Nick Johnson Technologist

Digital Equipment Corporation Norm Tripp Product Manager

Digital Equipment Corporation Peter Dauidoss Product Manager

Digital Equipment Corporation Prakash Bhalerao Engineering Group Manager

Digital Equipment Corporation Roslyn Morbay Product Marketing Manager

Digital Equipment Corporation Sri Sriran Marketing Specialist

Digital Equipment Corporation Susan Blount Senior Marketing Specialist

Digital Equipment Corporation Tim Yeaton Finance Manager

E.R.S.O. K. C. Chen Engineer

Eastman Kodak Company John C. Batterson Planning Specialist

Formative Technologies, Inc. Jose A. Villalobos Vice President, Marketing

Fujitsu America Inc. Katsuhide Hirai Manager, Applications

Dalaquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 GenRad Incorporated S. Russel Craig Director/ Corporate Marketing

General Computer Company Kevin Curran President

General Computer Company Pax Goodson Marketing Manager

General Electric Company James Whitten Manager Design Engineering

General Electric Company Marshall Kidd Manager Technical Analyst

General Electric Company Robert W. Baeder Stiaff Executive

General Electronic Company Dale W. Rowe General Manager 66TD

Gould Incorporated Dana A. Gauthier Manager Planning & Financial Services

Gould Incorporated Joseph Heaney Senior Vice President, Sales & Marketing

Hewlett-Packard Company Alan Benjamin Product Marketing Engineer

Hewlett-Packard Company Jim Casey Manager CAD/CAM Strategic Marketing

Hewlett-Packard Company Jim Lyons Product Marketing Manager

IBM A. C. Ewing Graphics Adminstrator

IBM Junius Brown Program Manager-Business Plans

IBM Richard Ma Senior Marketing Analyst

IBM Robert McGoldrick Manager Technology Development

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ICM, Incorporated Brian W. Kirschner Vice President, Research & Development

ICM, Incorporated Dr. L. L. Bacinka President

IDS Growth Spectrum Advisors Michael F. Guilford Senior Analyst

Impell Corporation Pat Strange Director

Intel Corporation Sudhir Bhagwan Assistant General Manager

Jennison Associates David T. Poiesz Vice President

Keystone Custodian Roland W. Gillis Vice President & Security Analyst

LSI Logic Corporation Mike Connell CAD Marketing Manager

LSI Logic Corporation Richard Perry Software Marketing Manager

Lexidata Corporation Forrest Gasik Manager, Forecasting

Lexidata Corporation Shelly Lowenthal Engineering Manager

Loomis, Sayles & Company Jaye E. Morency Securities Analyst

Matra Datavision John Keller Manufacturing Manager

McAuto Jeff Barnell Hardware Product Manager

Megatek Corporation Greg Elmassian Senior Engineer

Megatek Corporation Hiram French Vice President, Marketing

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Megatek Corporation Susan Harvey Corporate Planning Manager

Mentor Graphics Michael Bosworth Director of Marketing

Metheus Corporation Ken Tang Senior Accounts Manager

Mosaic Technology Arthur Campbell President

NCC Corporation Jack Mullins Liason from. NCR Corporation

NEC Electronics USA, Inc. Hotaek Rhim Manager Semi-Custom Design

NEC Electronics USA, Inc. Robert Scott Technical Staff Member

National Semiconductor N. T. Hsu Senior Engineering Manager

Nixdorf Lorraine Chouinard Product Line Manager

Oak Management Corporaticm Carla Haugen Associate

Olivetti Research & Development Charles Erickson Manager

Pacific Telesis Group Bill R. Frazer Director

Paragon Technology Philip Monego President

Parallax Graphics Jim A. Summers President

Perq Systems Richard Cox Director, Marketing Planning

Perq Systems Corporation Richard Rifenburgh Chairman of the Board/CEO

Dataquest incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Parl< Drive / San Jose, GA 95131 / (408) 971-9000 / Telex 171973 Prime Conputer Corporation Barry Wolman Director, Systems Development

Prime Computer Corporation Celia Bland Manager Workstation Marketing

Prime Computer Corporation Edward Vravlick Systems Concusltant

Prime Computer Corporation Maxin Smith Director Technical Strategies

Prime Computer Corporation Michael Payne Director, CAD/CAM Reserach & Dev.

Prime Computer Corporation Peter Mills Manager Workstation Programs

Prudential Bache Securities Laura Conigliaro Vice President

Prudential-Bache Securities Christine Chien

Prudential-Bache Securities Michael Weisberg Directx>r of Technology Research

Pyramid Technology William Gimple Vice President, Research and Dev.

Pyramid Technology Frank Madren Vice President, Marketing

RCA J.W. Scherer Manager Product Marketing

Racal-Redac Incorporated Ray Spencer

Racal-Redac Incorporated Richard Barrett Development Director

Raster Technologies Richard Fichera Director of Marketing

Ridge Computer Company David Folger President

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Ridge Computers Harry Taxin Vice President, Marketing & Sales

Saber Technology Dennis Peck President

Saber Technology Jim Richardson Vice President, Operations

Sanders Associates Ronald Derby Corporate Marketing Research Manager

Siemens Aktiengesellschaft-Dep E Wolfram Howein Dipl. - Ing

Silicon Graphics Incorporated Craig S. Olson Product Manager

Silicon Graphics Incorporated Jim Clark Chief Technical Officer

Silicon Graphics Incorporated Vernon Anderson Chairman

Smith, Barney, Harris, Upham & Co. Inc. David S. Lobel Divisional Vice President

Spectragraphics John Moreland Director, Marketing

Spectragraphics Ray Barger Vice President, Marketing

Sun Microsystems Scott McNeally President

Syte Information Technologies Peter Shaw President

T. Rowe Price Michael Christodolou

Terak Corporation Tcmi Zender Director of Marketing

Tower Capital Management Samuel F. McKay

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Valid Logic Frank Lynch Director of Product Marketing

Wang Labs Ann Palermo Analyst

Wang Labs Gary Neil Director, Industry Marketing Programs

Wayne Kerr Ken Becquist Vice President

Wayne Kerr Ted Sisley Product Manager

Weitek Corporation Lynette DeNike Consultant

Weitek Corporation Arthur Collmeyer President

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dataquest

CAD/CAM Focus Conference

August 13 and 14, 1984

BurlingtOTi, Massachusetts

List of Attendees

AT&T Bell Labs Basant R. Chawla

Adage, Inc. Clarence A. Tilger

Amecicem Micro Systems Gary Miles

Analog Design Tools, Inc. Richard N. Lanham

Apollo Computer Win Burke

Anil Gadre

David Nelson

Lou Reynolds

Linda Sartes

Stephen Weinberger

Applicon Francis Donlan

Richard Miller

Jeffery Simon

Arthur Young and Conipany George Southwick

Bank of America Al Schumer

Dave Woonton

Boeing Aircraft Corporation Richard Hedlund

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD/CAM Specialists, Inc. Dan Garms

Michael W. McDonald

CAE Terminals Roger Budris

Cadmus Computer Systems Jeff Singer

CalComp Craig Montgomery

Rich Peters

Mike Watterson

Calma Company William Blasdell

Van Lewing

James Mannos

Tom Teeters

Century Capital Assoc. Arthur Vitolo

Ciprico, Incorporated Bruce Brunette

Computervision Corporation Michael Barry

James Cook

Hank Kellogg

Mark Krentzman

Stephanie Marrus

Matt Peterson

Counterpoint Coitputer Pauline Alker

Cybetnex James B. Money

Data General Corporation Donald McDougall

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dataquest Incorporated Eileen Barth

Irene Barth

Thomas Bredt

Sharlene Burden

Dave Burdick

Frank Florence

Debra Jacob

Chuck Kanupke

Wendy Ledamun

Kelly Leininger

Jim Newcomb

Pamela Shook

Bob Smith

Brad Smith

Beth Tucker

H. Edward White

Diconix Ernest Trvax

Digital Equipment Corporation Libby Aston

Prakash Bhalerao

Susan Blount

Ken Brabitz

Chan Yan Chow

John Clarke

John Dimack

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Nick Johnson

Roslyn Morbay

Dorothy Rado

Barry Reynolds

Sri Sriran

Norm Tripp

Cheryl Vedoe

Tim Yeaton

E.R.S.O. K. C. Chen

Eastman Kodak Company John C. Batterson

Formative Technologies, Inc, Jose A. Villalobos

Fujitsu America Inc. Katsuhide Hirai

GenRad Incorporated S. Russel Craig

General Computer Company Kevin Curran

Pax Goodson

General Electric Con5>any Robert W. Baeder

Marshall Kidd

James Whitten

Dale W. Rowe

Gould Incorporated Dana A. Gauthier Joseph Heaney

Hewlett-Packard Company Alan Benjamin Jim Casey

Jim Lyons

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 IBM Junius Brown

A. C. Ewing

Richard Ma

Robert McGoldrick

ICM, Incorporated Dr. L. L. Barinka

Brian W. Kirschner

IDS Growth Spectrum Advisors Michael F. Guilford

Impell Corporation Pat Strange

Intel Corporation Sudhir Bhagwan

Jennison Associates David T. Poiesz

Keystone Custodian Roland W- Gillis

LSI Logic Corporation Mike Connell

Richard Perry

Lexidata Corporation Forrest Gasik

Shelly Lowenthal

Loomis, Sayles & Conpany Jaye E. Morency

Matra Datavision John Keller

McAuto Jeff Barnell

Megatek Corporation Greg Elmassian

Hiram French

Susan Harvey

Mentor Graphics Michael Bosworth

Metheus Corporation Ken Tang

Mosaic Technology Arthur Campbell

Dataquest Incorporated, A Subsidiary of A.C. Nielsen CoInpany / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 NBI, Incorporated Sky Stevenson

NCC Corporation Jack Mullins

NEC Electronics USA, Inc. Hotaek Rhim

Robert Scott

National Semiconductor N. T. Hsu

Ke Shih

Nixdorf Lorraine Chouinard

Oak Mcinagement Corporation Carla Haugen

Olivetti Research & Development Charles Erickson

PERQ Systems Corporation Richard Rifenburgh

Pacific Telesis Group Bill R. Frazer

Paragon Technology Philip Monego

Parallax Graphics Jim A. Summers

Perq Systems Richard Cox

Prime Computer Corporation Celia Bland

Peter Mills

Michael Payne

Maxin Smith

Edward Vravlick

Barry Wolman

Prudential Bache Securities Laura Conigliaro

Christine Chien

Michael Weisberg

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Pyramid Technology William Gimple

Frank Madren

RCA J.W. Scherer

Racal-Redac Incorporated Richard Barrett

Ray Spencer

Raster Technologies Richard Fichera

Ridge Computer Company David Folger

Harry Taxin

Saber Technology Dennis Peck

Jim Richardson

Sanders Associates Ronald Derby

Siemens Aktiengesellschaft-Dep E TDV Wolfram Howein

Silicon Graphics Incorporated Vernon Anderson

Jim Clark

Craig S. Olson

Smith, Barney, Harris, Upham & Co. Inc. David S. Lobel

Spectragraphics Ray Barger

Spectragraphics John Moreland

Sun Microsystems Scott McNeally

Syte Information Technologies Peter Shaw

T. Rowe Price Michael Christodolou

Tower Capital Management Samuel F. McKay

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Valid Logic Frank Lynch

Wang Labs Gary Neil

Ann Palermo

Weitek Corporation Lynette DeNike

Weitek Corporattion Arthur Collmeyer

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dataquest Dataquest

WORKSTATION INDUSTRY OVERVIEW

James R. Newcomb Director, Design and Manufacturing Automation Group Dataquest Incorporated

Mr. Newcomb is Director of DATAQUEST's Design and Manufacturing Automation Group. Before joining DATAQUEST, he was with Auto-trol Technology Corporation, where he held various executive positions, including Manager of Marketing. Prior to that, Mr. Newcomb was employed by Xerox Corporation, where he held positions as Manager of Systems IntegratiCMi, Manager of Advanced Manufacturing Engineering, and Divisional Program Planner. Earlier, he worked in an engineering capacity with Strippit-Hoidaille. His professional experience has provided him with an in-depth knowledge of manufacturing automation systems. Mr. Newcomb received a B.S.M.E. degree from Rochester Institute of Technology. He is a member of the American Society of Mechanical Engineering, SME, RI/SME, and CASA/SME.

Dataquest Incorporated WORKSTATION FOCUS CONFERENCE August 13 and 14, 1984 Burlington, Massachusetts

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Fdcusiyi^lr WORKSTATIONS

AUGUST 13^14,1384

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 TEOWGLOGY WFORMATION DIVISION DESIGN AND MANUFACTURING AUTOMATION GROUP

CAD/CAM NDUSTRY SERVICE

ROBOTICS INDUSTRY SERVICE

Source: DATAQUEST

- 2 - ® 1984 Dataquest Incorporated-Reproduction Prohibited

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 WORKSTATIONS DEFINITION- -SEGMENTATION

HOST DEPENDENT

STANDALONE

PERSONAL

Source: DATAQUEST

Dataquest Incorporated, A Subsidiary ot A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 WORKSTATIONS HOST DEPENDENT WORKSTATION

.^ ^1 HOST CENTRAL PROCESSING [1 UNIT |L^

• NON-RESIDENT CENTRAL PROCESSING UNIT • NO LOCAL OPERATING SYSTEM

Source: DATAQUEST

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 WORKSTATIONS STANDALONE WORKSTATION n • O/S IS RESIDENT IN PHYSICAL DEVICE • VIRTUAL-OPERATING SYSTEM • MULTI-TASKING • SUPPORTS NETWORKED COMMUNICATIONS • INTEGRATED GRAPHICS • NO SPECIAL POWER/ENVIRONMENT REQUIREMENTS

Source: DATAQUEST

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 WORKSTATIONS PERSONAL WORKSTATIONS

• LOCAL 8/16 BIT CENTRAL PROCESSING UNIT • NON-VIRTUAL MACHINE • SINGLE PROCESSING

Source: DATAQUEST

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 WORKSTATIONS

FUNCTIONALITY MULT!-TASKING

VIRTUAL OPERATING SYSTEM

GENERAL PURPOSE COMPUTING ! ]^

LOCAL OPERATING SYSTEM

NO SPEQAL ENVIRONMENT INTEGRATED GRAPHICS biD:::::::: COMMUNICATIONS SUPPORT STANDALONE PERSONAL HOST DEPENDENT WORKSTATION TYPE

Source: DATAQUEST

® 1984 Dataquest Incorporated-Reproduction Prohibited

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 FOCUS ON STANDALONE WORKSTATIONS

j^' ,vn f— f • 1 1 ^

^ ,-^. u y

Source: DATAQUEST

- 8 - 1984 Dataquest Incorporated-Reproduction Prohibited

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 STANDALONE WORKSTATIONS

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Coinpany /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 WORKSTATION EVOLUTION

FIRST GENERATION CORPORATE MAINFRAMES

SECOND GENERATION DEDICATED MINICOMPUTERS

THIRD GENERATION STANDALONE WORKSTATIONS

FOURTH GENERATION NETWORKED DISTRIBUTED PROCESSING

Source: DATAQUEST

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive I San Jose, CA 95131 / (408) 971-9000 / Telex 171973 FIRST GENERATION CORPORATE MAINFRAMES

• SINGLE APPLICATION ORIENTATION • NO INTEGRATION • LIMITED COMMUNICATIONS • UNPREDICTABLE PERFORMANCE • INCOMPATIBLE SW/HW

Source: DATAQUEST

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 SECOND GENERATION DEDICATED MINICOMPUTERS <»;;$KSSKS<;!SilK^;^SXS!Si!;!!SS8KiSSSKiSSUSKS^

• MULTIPLE DATA BASES • INCOMPATIBLE SW/HW • LIMITED COMMUNICATIONS • NON-STANDARD INTERFACES • FINITE INTEGRATION

Source: DATAQUEST

- 12 - ® 1984 Dataquest Incorporated-Reproduction Prohibited

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 WORKSTATIONS

"AWARENESS--CONCERNS""

• DIMINISHING PRODUCT LIFE CYCLES • INCREASING PRODUCT DESIGN CYCLES • LIMITED POOL OF ENGINEERING TALENT

Source: DATAQUEST

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ^jH)iV}j«jjSjS>>asiyjSJ?Syi8JJfryHHJS^^ A SOLUTION

IMPROVED ALLOCATION OF COMPUTER PROCESSING POWER

Source: DATAQUEST

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 «KK!«<4Si«»l«S":ii»«}n»^ THIRD GENERATION STANDALONE WORKSTATIONS

• PREDICTABLE LEVELS OF PERFORMANCE • IMPROVED PRICE/PERFORMANCE RATIOS • INTEGRATED HW/SW • UTILIZATION IN NORMAL ENVIRONMENTS • INITIAL STANDARDIZATION • AFFORDABLE GROWTH

Source: DATAQUEST

- 15 - ® 1984 Dataquest Incorporated-Reproduction Prohibited

Oataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 FOURTH GENERATION NETWORKED DISTRIBUTED PROCESSING

• HGH PERFORMANCE STANDALONE CAPABUTY • MAXIMUM RESOURCE SHARING • FASTER RESPONSE TIME • IMPROVED COMMUNICATION • TIGHTLY COUPLED INTEGRATION

Source: DATAQUEST

- 16 - ® 1984 Dataquest Incorporated-Reproduction Prohibited

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 HETEROGENEOUS SYSTEM COLLECTION

X r PERSOWtL WORKSTftTION (LOW COST) ^J^. FILE SERVERS MASS ST0RAI2 (TEDIUM COST) nftSS STORAGE I WCW PERFORMfWCE niWCOnPUTIR-BASED IJORKSTftTIONS TURNKEY SYSTEM OR ATTACHED PROCESSOR (HIGH COST)

GATEWnVI T

•^ // ^'LAN v/-- V ^ OTHER METUJOI«

~~Source: DATAQUEST~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 •.'•••.••.•.•.•,•..-.V. V.• •'••^:*^•.;";;^ ;;.V-^- Vf.*V:*: v"! Vi^'f.'

~~..... * •.'r: * •- ""T • ' • ' ;••'~~

~~.'i .. .;...... ,;....:... ^::, .. i ,.. ,: .>.:-.. . H . t .> " ' -:— -• •~~

~~••.'.~~

~~. WORKSTATIONm W ^^ 1 I • X V^ 1 ^^ I I %^ • MS *^ . ...» * •,-...•~~

~~:,• . i . i . . . ^v.. . ••••-••••.• ^ • ^: •-•.••.• ^ - -. 7.: • - •: • .• -.~~

~~'k 1" •;; SYSTEMS : ••. 4. , . •! i > >:••. *:••• AySPk i\nu ..... ^, ^,. ,.. .^^ [*} [ I ., . -i .4 >• . . COMPONENTS • % "• ••?~~

~~.». • . ••-• . j •*. -' ..••. !•:;• I.! . • • • • • - '^ •;;••; ;;••!:; ;v! '.i i !:;^;.;;:; \ :;C^-^ ^^:-v. :•. -•:: '• : ••::•::>^:i~~

~~• - ••• ••• • :••••~~

~~•' • • •:• i-\ ....~~

~~- 18 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 FOCUS ON WORKSTATION • ilTEGRATION S T A u~~

~~D '• INTEGRATED CIRCUITS A L 0 STORAGE N E COMMUNICATIONS W 0 GRAPHICS~~

~~SOFTWARE r s T A T 1 0 N S~~

~~- 19 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 S FOCUS ON WORKSTATIONS T A N D INTEGRATED CIRCUITS A L 0 N • MICROPROCESSORS E • ASICs W 0 • MEMORY R K . S T A T 1 0 N S

~~- 20 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

Dataqi.iest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ««SS?S?SW!?!*>?»!n«HHHK5yK^^ WORKSTATIONS 85gw:aBaw&e^gteaaa

~~CAPTIVE MERCHANT HP MOTOROLA DEC NATIONAL ATT/WE NCR INTEL Tl NEC ZILOG~~

~~Source: DATAQUEST.~~

~~- 21 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 vmiTOiiwiiBnaiBJSiyBBggsWjgns!^ MICROPROCESSORS--TODAY~~

• INTEGRATION OF EXISTING TECHNOLOGIES - MORE I/O FUNCTIONALITY • ADDITION OF SYSTEMS FUNCTIONALITY - PIPELINING - VIRTUAL MEMORY SUPPORT - MEMORY MANAGEMENT - CACHE MEMORY • MIGRATION TO 32 BIT MPUs

~~Source: DATAQUEST~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MICROPROCESSOR FORECAST 32 BIT MPUs CAGR = 74.8%~~

~~THOUSANDS OF UNPFS 2,000 i r / 1 1,500 1 t Jr 1 1 1.000 1 i 1 500 r —"j;^^ 1 f 1983 1984 1985 1986 1987 1988 1989 YEAR~~

~~Source: DATAQUEST~~

~~- 23 - ® 1984 Dataquest Incorporated-Reproduction Prohibited~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MICROPROCESSOR ASP FORECAST 32 BIT MPUs CAGR = -31.6? DOLLARS 500

~~• 400 1 1~~

~~1 1 1 ! 300 I 1 i i 200 r \ 1 :1 • J 100 ^^^^ftjQfc*. '~~

~~1982 1983 1984 1985 1986 1987 1988 1989 YEAR~~

~~Source: DATAQUEST~~

~~- 24 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~'ubsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MICROPROCESSOR REVENUE FORECAST 32 BIT MPUs CAGR = S1.75S~~

~~MILUONS OF DOLLARS~~

~~80 1 1 f j i~~

~~60 1 1 i1 4: ' \ 1 i t 40 T 1 20 ! ^0^\ •~~

~~1983 1984 1985 1986 1987 1988 1989 YEAR~~

~~Source: DATAQUEST~~

~~- 25 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MICROPROCESSORS--TRENDS~~

~~• UPWARD COMPATIBILITY WITH EXISTING WSTRUCTION SETS~~

~~• CHALLENGING THE PERFORMANCE OF MINICOMPUTERS~~

~~• HIGHER CLOCK SPEEDS - 10 MHz TO 25 MHz NOW - 50 MHz 3 TO 5 YEARS - 100 MHz COMING~~

~~• DEVICES/DIE INCREASING - 100K-400K NOW - 700K BY 1986 • 1 MILLION DEVICES SOON~~

~~• UTILIZATION OF COPROCESSORS~~

~~Source: DATAQUEST~~

~~- 26 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 S FOCUS ON WORKSTATIONS T A N D INTEGRATED CIRCUITS A L APPUCATION SPECIFIC 0 INTEGRATED CIRCUITS N E • CUSTOM W • GATE ARRAY 0 R • STANDARD CELL K ' S • PROGRAMMABLE ARRAY LOGIC T A T 1 0 N S

~~- 27 - ® 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 WHY ASICs ARE WINNING--TODAY~~

~~eSi9SiaWli^^~~

~~• POWER NEXPENSIVE MICROPROCESSORS • SELF-FULLFILLING PROPHECY • ASIC--FOCUSED START-UP COMPANIES • PROVISION OF POWERFUL DESIGN TOOLS • ACCEPTANCE OF CMOS VLSI • SOPHISTICATION OF USERS~~

~~Source: DATAQUEST~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive I San Jose, CA 95131 / (408)971-9000 / Telex 171973 ESTIMATED WORLDWIDE ASIC CONSUMPTION~~

~~BLUONS OF DOLLAFRS~~

~~1963 1984 1985 1966 1987 1988 1989 YEAR D PROGFMKt^ABLE LOGIC B STANDARD CELL 0 GATE ARRAYS M FULL CUSTOM~~

~~Source: DATAQUEST~~

~~- 29 - ® 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 IMPLEMENTATION ALTERNATIVES~~

~~COST~~

~~FULL ---i cusTon /' ^^ N 1/,.^ ^ STtWDrtflCELL*O \ ^-^,,j4 „,...'• V ^^ STRXTURO / r^ GATE \^ -; t-OCIC y ARRAVS ***j*"^* ^—*•—' '^^ f *" " S* ** N / \ —^ PROGRftPirWBLE ^^-^ \\ Loac ,•'~~

~~„. A^ . ..y~~

~~ETANI>^RD DAYS WEEKS MONTHS WODOCTS YEARS~~

~~30 ® 1984 Dataquest Incorporated-Reproduction Prohibited~~

Dataquest Incorporated, A Subsidiary ot A.C. Nielsen Company /1290 Bidder Park Drive/ San Jose, CA 95131 / (408) 971-9000 /Telex 171973 + FOCUS ON WORKSTATIONS S T A N D INTEGRATED CIRCUITS A L 0 N • MEMORY E W 0 R K S T A T 1 0 N S

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MEMORY FORECAST--DYNAMIC RAMS AVERAGE SELLING PRICE~~

~~DOLLARS~~

~~150 \ 100 V . 1 ^256K 50 - 64K 1 1^ A ^ ^ •—~~

~~1980 1981 1982 1983 1984 1985 1986 1987 1988 YEAR~~

~~Source: DATAQUEST~~

~~- 32 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ffigg'SggSJHHJSJBgJWWgyjJiSHS'^^ MEMORY REQUIREMENTS~~

~~g}«jijijjij)gjjijS9!jjgy;8JHgjjJiJJi^^~~

~~1,024 X 1.024 RESOLUTION~~

~~NO. OF NO. OF MEMORY NUMBER OF CHIPS COLORS BIT PLANES CBYTES 16K 64K 256K 1M~~

~~1 1 128K 64 16 4 1 16 4 512K 256 64 16 4 256 8 1MB 512 128 32 8 4,096 12 1.5MB 768 192 48 12 16M 24 4MB 1.536 364 96 24~~

~~Source: DATAQUEST~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 FOCUS ON WORKSTATIONS S T A nN STORAGE D AL 0 N E W 0 R K • S T A T 1 0 N S

~~- 34 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 STORAGE FORECAST COST PER MB OF ON-LINE STORAGE DOLLARS 60 i 1 i 1 \ i 50 i t 1 1 ' ^^^^1^^^ fc^ \5 1/4" (30-100 MB) ALL DJSK ^W ! % i 40 1^^ <

~~30 * * ^^^^^^^^^^^^^ ^^^9i ^^^ 20 T- ! " 1 j i 10 1 1 i i 1 1 1 i i 0 * 1 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 YEAR~~

~~Source: DATAQUEST~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 TRENDS IN COMPUTER STORAGE~~

~~• LOWER COST PER MB • HIGHER AREAL DENSITIES • SMALLER FORM FACTORS • PERFORMANCE ENHANCEMENTS • PERSONAL STORAGE • EXPLOSION IN PRODUCT AVAILABILITY~~

~~Source: DATAQUEST~~

~~- 36 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 S FOCUS ON WORKSTATIONS T A N D COMMUNICATIONS A L 0 N E~~

~~W 0 R K S T A T 1 0 N S~~

~~- 37 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 COMMUNICATIONS FORECAST COST PER NETWORK CONNECTION THOUSANDS OF DOLLARS

~~STANDALONE WORKSTATION~~

~~1983 1984 1985 1986 1987 1988 YEAR~~

~~Source: DATAQUEST~~

~~- 38 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 COMMUNICATIONS FORECAST~~

~~BILLIONS OF DOLLARS 5~~

~~DATA PBX CAGR 37~~

~~1984 1985 1986 1988 YEAR~~

~~Source: DATAQUEST~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ^JHgiW^SgJSft'jyjtJjJJH^HHJjjJiSJgJaggJjJH^^^ COMMUNICATIONS- -ISSUES

~~• PROPRIETARY VS. STANDARD NETWORKS • HOW DOES UNIX FIT • OPERATING SYSTEM NETWORKS • PRICE/PERFORMANCE RATIOS • ROLE OF PBXs AND DATA PBXs~~

~~Source: DATAQUEST~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 S FOCUS ON WORKSTATIONS T A nM GRAPHICS D AL 0 N E W 0 R K . S T A T 1 N S0

~~- 41 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ^JSJSSBggJjgaJHKJBSSBBHWjJSJJBgJ^^ GRAPHICS EVOLUTION~~

~~• STORAGE DEVICES/HOST DEPENDENCY • RASTER/HOST DEPENDENCY • RASTER WITH LOCAL GRAPHICS PROCESSING~~

~~Source: DATAQUEST~~

~~Dataquest Incorporated, A Subsidiary ot A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ggj<;WJHSJS'i8^'KgJJg{?iSK~~

• UTILIZATION OF LOCAL VLSI • HIGH SPEED MATRIX ENGINES • ADVANCED SHADING AND RENDERING CIRCUITRY • LOCAL GEOMETRY PROCESSING • HIGHER RESOLUTION DISPLAY TECHNOLOGIES • EVER DECREASING COSTS

~~Source: DATAQUEST~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 S FOCUS ON WORKSTATIONS T A nM SOFTWARE D A L 0 N E W t 0 R K S T A T 1 0 N S

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 SOFTWARE--THE PAST~~

~~• S/W AND O/S HARDWARE DEPENDENCY • LIMITED PROGRAMMING TOOLS • LONG DEVELOPMENT TIMES • LIMITED PORTABILITY/LIMITED LEVERAGE~~

~~Source: DATAQUEST~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 SOFTWARE~~

~~• COMPUTER ADED SOFTWARE ENGINEERNG • UNIX AND "UNIX LIKE" O/S • STANDARDIZATION • MIGRATION OF SW INTO FIRMWARE HARDWARE~~

~~Source: DATAQUEST~~

~~- 46 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 *,'.*»',f"l'."^'^',\'. •,',*,',"^*T'-*-'-' r''^'H*-*^4-' •".F'.,'.',*^'.**' :::::::::;V:;

~~•• :;*':::::• THANK YOU1;;;;:;;;;:; ;~~

~~•i ::i;VLSl:i:: • - ••• • • • •;~~

~~T. _ , ^ » »~~

~~' • ' ' ^ ' = >.-,.:.; H.H > ... ;...^ ;.* ..^ ... V -, >,:^v,. y .;.,. .::, ^...^,~~

~~- 49 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Oataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Rldder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 WORKSTATIONS~~

~~ISSSiSSWSflllSli&i^llMlSI'SXIia'MMIiSVSS^^ SYNOPSIS~~

• DESIGN OF CLOSELY COUPLED SYSTEMS (HIGHER LEVELS OF SYSTEMS INTEGRATION) • EVOLUTION OF COMPLEX MULTI-PROCESSOR DESIGNS • FUSION OF GRAPHICS AND IMAGING TECHNOLOGIES • ACCEPTANCE OF STANDARDS (DEVELOPMENT OF OPTIMAL SOLUTIONS FOR SPECIFIC APPLICATIONS • PUSHING PERFORMANCE TOWARDS HUMAN PERCEPTION

~~Source: OATAQUEST~~

Oataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 «H!!ig»»jjj8«9!jSjgyj»'^{g»8S»^38iHS^^ FUTURES in^ssstisss^

• A SINGLE LEVEL. NETWORK-WIDE DEMAND PAGED ADDRESS SPACE • PROVISION OF A UNIFORM INTERFACE BETWEEN ALL SYSTEMS AND RESOURCES • CONFIGURATION ARCHITECTURE THAT SUPPORTS A WIDE RANGE OF PERFORMANCE LEVELS • TECHNOLOGICAL INDEPENDENCE

~~Source: DATAQUEST~~

~~- 51 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 USEFUL AFFORDABLE COMPUTING POWER~~

~~-52 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 STANDALONE WORKSTATIONS C l^^^e^^^^~~

~~1984 MARKETS DESIGN AUTOMATION~~

~~MECHANICAL APPLICATIONS 37% ELECTRONIC APPLICATIONS 13% AEC APPUCATIONS 10% MAPPING APPLICATIONS 3% OTHER APPLICATIONS 2%~~

~~CAP~~

~~OTHER Al CASE~~

~~53 ® 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.G. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MARKET GREDIBILITY~~

~~- 54 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 (::>aDataquest Dataquest~~

~~ARCHITECTURES, GRAPHICS, AND PERFORMANCE~~

~~David Burdick Industry Analyst Dataquest Incorporated~~

Mr. Burdick is an Industry Analyst in DATAQUEST's CAD/CAM Industry Service. Before joining DATAQUEST, he held System Engineering positions with Auto-trol Technology Corporation, where he was the key technical consultant for turnkey CAD/CAM systems. Prior to that, he worked as a Design Engineer for International Harvester and was involved in the design and implementation of large-scale CAE/CIM systems. Mr. Burdick has in-depth knowledge of image processing, graphics interfaces, workstation design, and AEC application areas, including piping, structural, and HVAC design. Mr. Burdick received a B.S.M.E. degree fron Bradley University in Peoria, Illinois.

~~Dataquest Incorporated WORKSTATION FOCUS CONFERENCE August 13 and 14, 1984 Burlington, Massachusetts~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD/CAM INDUSTRY SERVICE DATABASE SEGMENTATION~~

~~Source: DATAQUEST~~

~~- 3 - O 1984 Datiaguest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CADICAM INDUSTRY SERVICE DATABASE SEGMENTATION~~

•

~~Source: DATAQUEST~~

~~AEC « ARCHITECTURAL. ENGINEERING AND CONSTRUCTION~~

~~PC - PRINTED CIRCUIT BOARD~~

~~IC « INTEGRATED CIRCUIT~~

~~EDA - ELECTRONIC DESIGN AUTOMATION~~

~~- 4 - C) 1984 Dataguest Incorporated-Reproduction Prohibited~~

~~Oataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CADICAM INDUSTRY SERVICE DATABASE SEGMENTATION~~

~~€ Source: DATAQUEST~~

~~- 5 - C) 1984 Datacpiest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CADICAM INDUSTRY SERVICE DATABASE SEGMENTATION~~

~~Source: OATAQUEST m~~

~~- 6 - Q 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Oataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MARKET SIZiNG~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD/CAM TOTAL REVENUE~~

~~BILLIONS OF DOLLARS 15~~

~~CAGR = 40^~~

~~1883 1884 1885 1886 1887 1888 YEAR~~

~~•J Source: DATAQUEST~~

~~- 8 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD/CAM REVENUE ^^ ^'^^ BY WORKSTATION TYPE~~

~~MLUONS OF DOLLARS 7,000~~

~~1985 1986 1987 1988 YEAR~~

~~Source: DATAQUEST~~

~~- 9 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD/CAM WORKSTATIONS PERCENT OF REVENUE~~

~~1983 1988~~

~~m PERSONAL M STANDALONE D HOST DEPENDENT~~

~~Source: DATAQUEST~~

~~- 10 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD/CAM UNITS BY WORKSTATION TYPE~~

~~THOUSANDS OF UNPFS 70 60 HOST DEPENDENT 50 CAGR362 40 STANDALONE 30 CAGR 74% 20 PERSONAL 10 --CAQR 62%"~~

~~1983 1984 1985 1986 1987 1988 YEAR~~

~~Source: DATAQUEST~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 AVERAGE CAD/CAM WORKSTATION PRICE HARDWARE AND SOFTWARE~~

~~THOUSANDS OF DOLLARS 150 if) •~~

~~• HOST DEPENDENT CAGR -6^ 100 „ _ • ^ I !~~

~~i ^l*''"^'"*™^^ :~~

~~50 t STANDALONE CAGR{-n~~

~~1983 1984 1985 1986 1987 1988 YEAR~~

~~Source: DATAQUEST~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 AVERAGE CAD/CAM WORKSTATION PRICE HARDWARE ONLY~~

~~THOUSAhDS OF DOLLARS~~

~~60 -^ 1 j 50 *^^^miTWlP|TItir*VT'^r 1 HOST! DEPENDENT:-'•^"'' ^''''^'^-^ ^^^TTr^BCEnrCni^^j^^^^ "^Vt'iUl^^Jgf^^ .... CAGR -5? -^'—'- ^ • ,....^-„ . 40~~

~~30 iii;jiiv'r"""""~~

~~'*•• 20 CAGR -2% I 10~~

~~1 [ A 1983 1984 1985 1986 1987 1988 YEAR~~

~~Source: DATAQUEST~~

~~- 13 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 AVERAGE WORKSTATION PRICE/PERFORMANCE~~

~~THOUSANDS OF DOLLARS~~

~~19S3 1984 19SE mh 1987 199? VEWi~~

~~BASE SYSTEM KEY INFLUENCES • O.5-1.0 MIPS • 256K RAMS • MONOCHROMATIC • HIGHER PERFORMANCE CPUs • 90MB DISK • INCREASING COMPETITION • 1MB MEMORY • ETHERNET • UNIX~~

~~Source: DATAQUEST~~

~~- 14 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 THE TECHNOLOGICAL CHALLENGES WORKSTATlbNS FACE~~

~~IN /~~

~~- 19 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 SYSTEM FUNCTIONS~~

• VARIOUS DATA TYPES • COMPUTE INTENSIVE FUNCTIONS - MATRIX MANIPULATION - FLOATING POINT CALCULATIONS - LARGE VIRTUAL ADDRESS SPACE REQUIREMENTS • SOPHISTICATED GRAPHICS - HIGH SPEED RASTER OPERATIONS - COMPLEX GEOMETRY FORMS - SHADING AND RENDERING

~~Source: DATAQUEST~~

~~- 20 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ENGINEERS CREATE MbbiFY; AND COMMUNICATE IDEAS THROUGH A LANGUAGE CALLED *^GiEOMETRy"

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 NUMERICAL CONTROL ENGINEERING DRAWINGS MACHINING~~

~~ENGINEERING ANALYSIS TECHNICAL PUBLICATIONS~~

~~Source: DATAQUEST~~

~~- 22 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 THE GRAPHIC PROCESS--PAST~~

~~SOFTWARE HARDWARE~~

~~GEOMETRY CREATION~~

~~DISPLAY LIST~~

~~TRANSFORMATION~~

~~SCALING/CLIPPING~~

~~SHADING~~

~~^ DISPLAY~~

~~Source: DATAQUEST~~

~~- 23 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Rldder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 THE GRAPHIC PROCESS--TODAY~~

~~SOFTWARE HARDWARE~~

~~APPLICATIONS GEOMETRY~~

~~TRANSFORMATION~~

~~SHADING~~

~~SCALING/CLIPPING~~

~~DISPLAY~~

~~Source: DATAQUEST~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 THE CAD/CAM OEM ENVIRONMENT PAST~~

~~Computers Oc^rittris Woritnatiors SysXams Awicatimf CoiMunioti^nj R&D $~~

~~PRESENT~~

~~Source: DATAQUEST~~

~~- 25 - @ 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 GROSS MARGINS PAST~~

~~INTERNALLY DEVELOPED SOLUTIONS~~

~~PRESENT~~

~~EXTERNALLY DEVELOPED SOLUTIONS~~

~~Source: DATAQUEST~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 DEGREES OF INTEGRATION~~

~~CHIP LEVEL PRODUCTS~~

~~BOARD LEVEL PRODUCTS~~

~~WORKSTATION LEVEL PRODUCTS~~

~~Source: DATAQUEST~~

~~- 27 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 TIME TO MARKET VS. DEGREE OF INTEGRATION~~

~~HIGH~~

~~DEGREE OF INTEGRATION~~

~~LOW~~

~~TIME TO MARKET~~

~~Source: DATAQUEST~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD/CAM SOFTWARE DEVELOPMENT MIX~~

~~NOW FUTURE~~

~~D NEW APPLICATIONS M MAINTENANCE -^ GRAPHICS @ SYSTEM SUPPORT~~

~~Source: DATAQUEST~~

~~- 29 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 THE NEW DEVELOPMENT ENVIRONMENT~~

~~JHSSSRSS'~~

• LOCAL VLSI FOR ALL GRAPHIC FUNCTIONS • INCREASED PORTABILITY--REDUCED MACHINE DEPENDENCE • VENDORS USING WORKSTATIONS FOR DEVELOPMENT- -C.A.S.E. • MORE EFFICIENT DATA STRUCTURES • SMARTER COMPILING AND DEBUGGING TOOLS

~~Source: DATAQUEST~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 DATAQUEST~~

~~- 31 - 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 .CT^^T^ Dataquest Dataquest~~

~~WORKSTATIONS, THE TRANSFER OF TECHNOLOGIES~~

~~Dr. David L. Nelson Senior Vice President Apollo Computer Incorporated~~

Dr. Nelson is a founder of Apollo Ccanputer Incorporated and has served as Vice President of Research and Development since the incorporation of the company in February 1980. From March 1977 to January 1980, Dr. Nelson served as Director of Research at Prime Computer, Inc. Prior to that, he was Manager of Research and Development at Digital Equipment Corporation. Dr. Nelson received a B.S. degree from the University of Wisconsin and a Ph.D. degree from the University of Maryland.

~~Dataquest Incorporated WORKSTATION FOCUS CONFERENCE August 13 and 14, 1984 Burlington, Massachusetts~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o CaD n c ID~~

~~O o "O o~~

~~> CO c CT~~

~~> O iapollo o o 3~~

~~M ID O a a (D "D lU PERSPECTIVES ON THE COMPUTER INDUSTRY: o CA EMERGENCE OF WORKSTATIONS CO O>~~

~~CD O O O -( a> ao D~~

~~0^ > (7> ^ Q ^o z LU § 3 CL zd O O o UJ~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o SJ. w n c (D~~

~~•o o~~

~~W c CT (A a~~

~~> O iOpoio o o 3~~

~~-D EU BEEN CONSISTENT AND PREDICTABLE FOR THE PAST CO fll 25 YEARS . . . BUT.. .. o >~~

~~(a •-J tj o o •3 o > lapollo~~

~~> b o o 3 ... ALL TRENDS EVENTUALLY T3 01 BREAK DOWN, CAUSING a INTERESTING QUALITATIVE a. a> tj !U CHANGES.~~

~~05 Dl 3 o >~~

~~lO --J CO o u n c a~~

CD Q. ^Qp OIK w c CT 01 a TECHNOLOGY IMPROVEMENT S' RATES > (1964-1990) o IMPROVEMENT RATE!. YEAR KBITS/CHIP KSlMiP $(CPU-I-MEM) MBYTE/MIP MEMORY PROCESSOR SYSTEr o o — 3 1964 0 $500 $1500 1.0 100% 30% •o 01 • • • • • • • • • • • • • • • • N3 (O O 1974 1 $40 $136 1.2 100% 30% — a •. 1976 4 $24 $52 1.4 100% 30% 61% 1978 14 $14 $24 1.7 90% 30% 50% 1980 42 $8 $12 2.0 70% 30% 38% 1982 107 $5 $6.8 2.5 60% 30% 34% 1984 273 $2.9 $3.8 3.0 60% 30% 34% 1986 700 $1.7 $2.1 3.6 60% 30% 33% CO 01 1988 1792 $1.0 $1.2 4.3 60% 30% 32% 1990 4600 $.60 $.70 5.1 60% 30% 32% o >

~~X3 O~~

~~> w c (T~~

~~> o lapollo o TECHNOLOGY IMPROVEMENT o 3 RATES •D (U~~

~~N) (O O CONSTANTS VARIABLES d Q. (D COST PER POUND ILECTRONIC DENSITY POWER PER CUBIC FT BITS/CIP (70-100%/YR) COST PER WAH LOGIC (40%/YR) COST PER CUBIC FT DISK STORAGE (30%/YR)~~

~~U 3 AGGREGATE SYSTEM 30-40%fYR o (1.35)'5 10 > FACTOR OF TEN IMPROVEMENT EVERY SEVEN YEARS~~

~~to CO .0 OH o> ^ s (/> LU V> UJ < < sa Ik (/} S Z < o 0 CO o> 2 1~~

~~" I/I tt < 0 U4 . >- — 0* Z UJ < > LU~~

~~o o^~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o CO o»~~

~~< g o X rs O~~

~~o «o~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o a~~

~~> en c cr lapoio (A ci. 0)'~~

~~> DEMAND PAGED VIRTUAL MEMORY b~~

~~o $1M£6 o 3 •o 01 IBM 370~~

~~$100K a! a PRIME 400 (t> VAX 780 Tl 0) $10K o APOLLO DN300 CO~~

~~1960 1970 1980 1990 >O~~

~~O CD oto o o CO o~~

~~r^ ^ o^~~

~~o o iu o s.~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive/ San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o W c; UJ ea o trt z Z UJ o z 1— DC < 1— < w>~~

~~o 0 u. ::«; •iA o o 2 o *" «> «>~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o lu~~

~~> 0) c CT~~

~~> b iopolb~~

~~o o 3 X) 01 3 IMPLICATIONS OF THE END OF TIMESHARING 3)~~

-Q * OPPORTUNITY FOR QUALITATIVE IMPROVEMENTS a IN THE USER INTERFACE (CONCURRENCY, WINDOWS, GRAPHICS. ETC.) in 01 • PREDICTABLE LEVEL OF PERFORMANCE/USER I • REQUIREMENT FOR LAN AND A DISTRIBUTED SYSTEM J ARCHITECTURE TO ALLOW EASY SHARING o

~~(D a > CO lopoio I PRODUCTIVITY PRODUCTIVITY > • •»• OF OPERATIONS OF PROFESSIONALS~~

~~ERA OF DEDICATED o • »»• o 3 TIIVIESHARING COMPUTING •a a ro (O o NEW GENERATION a a DISTRIBUTED (D -1 "0 WORKSTATIONS 01~~

~~w u~~

~~o CO o(O o~~

~~<£-^> CJ •oa~~

~~CO c (c^r Q! B)'~~

~~> b iOpolo~~

~~o o 3 DEDICATED COMPUTING~~

~~I MAJOR ARCHITECTURAL T3 a PROBLEM:~~

~~O) 01 • SERVICE PROTOCOLS • NETWORK VIRTUAL MEMORY o > • REMOTE PROCEDURE CALL~~

~~CoO o o w c CJ (A a Ol'~~

~~> O lapoio~~

~~o o 3 APPLICATION ENVIRONMENT~~

~~ro CO 5o bT d. a I (D NATIVE OPERATING SYSTEM~~

~~£/) 01 PRIMITIVE~~

~~>O DISTRIBUTED SYSTEMS (APOLLO) UNIX-BASED WORKSTATIONS DESKTOP COMPUTERS (IBM PC) PORTABLE COMPUTERS (TANDY MOD 100)~~

~~o o o~~

~~B> a c~~

~~05 c cr CO Q! iapollo ai"~~

> PREDICTED RAM/DISK MEMORY CAPACITY O RAM MEM/DISK MEM (MEGABYTES) o SYSTEM PRICE 1983 1985 1987 1989 o 3 $120K 4/200 9/400 20/800 45/1600 $ 60K 2/100 4/200 10/400 22/800 $ 30K 1/50 2.2/100 5/200 11/400 $ 15K .5/25 1.1/50 2.5/100 5.7/200 Q. (D $ 7.5K .25/12 .5/25 1.2/50 2.8/100

~~^ INCREASING FUNCTIONALITY, CONSTANT COST~~

~~CO (U CONSTANT FUNCTIONALITY, DECREASING COST~~

~~O ASSUMES: 507o IMPROVEMENT RATE FOR RAM > 41% IMPROVEMENT RATE FOR DISKS~~

~~o(O o o o Dl~~

~~T3 O~~

~~> V> C u lapollo~~

~~> LOCAL AREA NETWORKS~~

~~o o 3 ARTIFICIAL DEMISE OF p INTELLIGENCE TIMESHARING~~

~~ro to o PC PHENOMENON a CO T3 GRAPHICS PROFESSIONAL PRODUCTIVITY~~

~~ADVANCED USER o ENVIRONMENTS TECHNOLOGY >~~

~~U3 O O O~~

~~(O w •d o~~

~~> CO c rapollo a-~~

~~> O FUTURE INFLUENCES~~

~~o o • CONTINUED PROGRESSION OF COMPUTING 3 CLOSER TO THE USER (FROM CENTER TO OFFICE TO BRIEFCASE) • EASE OF USE, "SEAMLESS" INTEGRATION~~

~~T) • GROWING INFLUENCE OF Al IN CAD/CAM 01 • LANS: FAST, LARGE. SWITCHED, AND GATEWAYED. W SJ~~

~~O> D~~

~~o o •3 o~~

~~> w CJ~~

~~> b lapoio~~

~~o o 3 x> SI 3 IS3 CONOLUSIQN (O O a. a. (D THE COMPUTER INDUSTRY IS WITNESSING A MAJOR TRANSFORMATION IN THE STYLE AND USE OF COMPUTERS .. . MORE~~

~~w SIGNIFICANT THAN TIMESHARING MORE 0) 3 THAN 15 YEARS AGO. o >~~

~~.c^ o 00~~

~~(C o o o~~

ID u->J » o*

~~Q. CO fa LoU o < LU CD~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ::^i)::::aDataC|Uest The Role of Geometry in the Workstation Marketplace

~~James H. Clark~~

~~Silicon Graphics, Inc. Mountain View, Ca.~~

~~Introduction~~

The workstation era in computing has evolved from the availability of 32-bit microprocessor CPU's, "portable" operating systems, programming language developments, declining memory costs, low-cost graphics, and communications networking technology- More recent developments have enabled custom VLSI technology to inexpensively increase graphics speeds by factors of several hundred or more. In this presentation, each of these developments will be discussed and applications that require geometric manipulation in their graphics systems will be highlighted. In the process, the historical roots of Bit-mapped graphics will be presented and contrasted with those of Geometric graphics. Particular emphasis will be placed on the applications that use each capability.

~~Computing Costs vs. Human Costs: 1974-1984~~

In the past ten years, the price/performance for a typical computer system has decreased from $1.5M/MIP in 1974 to about $40K/MIP in 1984, corresponding to roughly a 43% per year compounded decrease, as shown in Figure 1. During the same period, the burdened salary of a typical engineer has risen from about SSSk/year to roughly $75K/Year. In the early 1980's, for the first time the computer user was more expensiv» than the computer.

These trends indicated a need for both a professional personal computer and an efficient interface to it. Graphics has become this modern interface because it is such an essential part of engineering work. Thus the "workstation", as it is discussed here, is minimally defined as a combination of some form of raster graphics system with a general-purpose computing environment in one physical package.

~~- 1 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 I~~

~~o-^ Bl THOUSANDS OF DOLLARS~~

~~(cA tJ 506 (1) d' COST OF COMPUTER IN $/MIP ai' 456 > VAX 780 o 406~~

(D 3 350 o (O o 3 c •o 306 lU 3 250 — X^VAX 780 i trt 200 I < (A 150 w I ^-, P ^S^^P 9000 T) n> 100 — BURDENED SALARY 68020 -^ w en 01 OF ENGINEER O 3 50 3 o > 0 1 1 1 1 1 1 1 1 1 1 1 •74 '75 '76 '77 '78 '79 '80 '81 '82 '83 '84

~~.b O 00~~

~~(£> O O O~~

•f -:*

~~(O CJ The Modern Workstation Environment~~

The modern professional workstation provides a general-purpose computer that may be used alone or connected in a network with a wide variety of computing and support resources. A typical environment might consist of workstations, mainframe computers, terminals, hard copy media, database systems, and so forth as shown in Figure 2. Both high-speed computing and high-speed graphics are an integral part of this professional computing environment.

~~What distinguishes one workstation environment from another? In addition to cost, some pertinent issues are addressed by the following questions;~~

1. Applications Environment - Is the workstation general-purpose, allowing applications software for many applications on one machine? What applications are available on the workstation? How fast does it process an application?

2. Computing Environment - How are the workstations interconnected or networked? What are the Operating System, available languages and applications development tools? In the workstation's environment, can one do everything that is possible in a time-sharing system, such as share data transparently and work at any physical workstation?

3. Graphics Environment - Is the interface high-level, application (user-space) Oriented and conveniently coupled with the computing capability? Is there support for 3D wire-frame and shaded pictures. How fast is it? Is it color or B&W, high-resolution? Is there multi-window support?

~~General-purpose Computing~~

~~The computing capability of today's workstation is enabled by numerous technological developments, including~~

~~1. Declining costs and increased speeds of microprocessor CPU's, semiconductor memory and secondary storage media.~~

~~2. Improved operating systems, compiler techniques, languages and applications development tools.~~

~~3. Communications networking hardware and standard network software protocols for communication.~~

~~- 3 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 D W W a c a~~

~~SILICON GRAPHICS, INC. ^ (/> c cr u ci. THE GEOMETRY NETWORK S' > la a 3 3-D GEOMETRY GRAr>HIC ARTS Ut (D PROCi^SSING MAINFRAME y^^D ANIMATION to o c o 3 •1 •o « 01 ro~~

~~NJ U> O zr (D SIMULATION Ci j^~ (D a O ITWIIIW 3 CB CD z (D en #-• QJ 'ra I I 9 3 9 o o -^ 9 > MECHANICAL DESIGN AND ANALYSIS~~

oCO ELECTRONIC o o DESIGN d g (D mmamiammsa^^^^ General-purpose Hardware Developments The 32-bit microprocessor available today has the computing speed of super minicomputers of less than ten years ago. These microprocessors provide the foundation of most workstations' general-purpose computing capabilitv. VLSI technology has moved the CPU from sets of discrete parts to custom integrated circuitry. Concurrent with microprocessor developments, Winchester disk technology is providing remarkable secondary storage capacities, while improved and more dense semiconductor memories continue to lead the revolution in computing systems. These trends are expected to continue to yield more and more powerful general-purpose systems for very low cost.

UNIX: The Portable Operating System. » If UNIX has not already become the de facto standard operating system for workstations and most minicomputers, it certainly appears to be on its way. It is the single most widely used operating system in universities — thus virtually every technical undergraduate in the country is familiar with it. Moreover, there are more systems programmers for UNIX than for any other operating systems because Computer Science students have been able to access the source programs for UNIX — UNIX was the first widely available "open" operating system. Developed in the early 1970's at Bell Labs, UNIX provides a rich and usable environment for applications development. The major question about UNIX in most people's minds is which version — System V and successors distributed by Bell Labs, or 4.2 BSD implemented by the University of California at Berkeley.

The two systems have slightly different sets of utility programs, and each has features lb . not in the other. The fundamental differences between them as they are distributed, and the effects on applications, are outlined below.

1. Demand Paging - With demand paging, physical memory may be less than the logical memory required by an application. Many engineering applications require a large virtual program and data space, and demand paging. In an engineering environment, this is the biggest shortcoming of System V as it is distributed by Beil.(BerkeleY UNIX was developed primarily because of this shortcoming.)

2. File Structure - At Berkeley, it was felt that a different file system organization would provide faster file system performan than Bell UNIX. The mechanism used to speed it up is more effective only when the disk is 75%

~~- 5 -~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 utilized. When less than 25% of the disk is free, experience has shown that the system is not as fast as the standard UNIX file system.

3. Network Protocols - System V has no standard support for network communications and interprocess communications. 4.2 BSD incorporates a communications protocol defined for the ARPANET by the Department of Defense. (The development of 4.2BSD was funded by the DOD.) These modules are easily ported to any UNIX version, so it is of no particular advantage to any system.

The main point about any operating system feature is its effect on the application or its development. Demand Paging is essential to any engineering workstation and should be totally transparent to the application. The value of the reoraginzed file system is often debated, at least as implemented. Networking is essential today as well — the shortcomings of 4.2BSD networking are treated in the next section.

The ideal engineering UNIX system is one that incorporates demand paging and utilities of both System V and 4.2 BSD with expanded networking features to enable a distributed system. Bell Labs will continue to support, develop and standardize UNIX, whereas UC Berkeley is not in the UNIX business. Thus an integration of the useful 4.2 features into Bell-distributed UNIX will probably dominate the UNIX workstation market eventually — it leaves the applications development and support environment least affected.

~~Ethernet: The Open Network Architecture. Any distributed workstation environment must support a means for sharing data and resources on a network. Two basic technologies are in predominate:~~

"ifc. 1. Token Ring - This communications technology requires that all resources on the network be in a closed loop. Each node is allowed to send information to another in the ring after a token passes it — thus, messages are synchronized and require the active participation of all nodes on the network. The network must be disabled to install or remove a node, and it is therefore more susceptible to failures of any node.

2. Ethernet - The Ethernet is not a closed loop. An Ethernet module communicating with another first "listens" to the network to see if it is in use. If not, it sends its file or message to the recipient. A "collision" occurs if two different nodes simultaneously sense that the network is quiescent and begin transmitting at the same time. The message is garbled, so each must "back off" and re-transmit its message after a randomly chosen interval, which decreases the likelihood that they will again simuttaneously transmit.

~~- 6 -~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Part^ Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Messages are asynchronous. Nodes are passive unless they are using the Ethernet, so they may be removed and inserted at will.

The most significant difference between these two technologies is that the Ethernet is an "open" architecture that has been adopted as a standard by Xerox, Intel, Hewlett Packard, Digital Equipment Corporation, Silicon Graphics, SUN Microsystems, and many others. All commonly used token ring approaches are "closed." No standard token ring approach has been widely adopted — each company that employs it has its own closed system.

Network Protocols. Both Ethernet and token ring require a protocol, or "language", for message transmission. There are two published standards being used in most Ethernet installations, XNS and IP/TCP.

XNS (Xerox Network Software) is a simple local area network communications protocol defined at Xerox PARC, where the Ethernet was invented, it has the advantage of being tailored for local communications, and thus has no excessive baggage associated with each message. IP/TCP was defined for the ARPANET and imposed on the Berkeley 4.2 developers by ARPA. IP/TCP has the disadvantage of having all baggage necessary to route messages to any point on the ARPANET, which is not a local area network, but is nationwide.

Distributed System An open network architecture requires facilities other than explicit file transfers between nodes and remote logins to other nodes and foreign hosts. It should allow any system user to refer to and use any file on the system implicitly, without explicit file copy. More generally, it should allow a virtual "system call" or shell command — where the command is actually executed should be transparent to the user. This encompasses the distributed file system notion and more.

~~- 7 -~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Graphics: The Principal Computer Interface Vision is the most used human sense, so quite naturally graphics has become our most important form of computer interface. As with computer systems, however, graphics systems vary in their suitability for certain applications. The goal of the engineering or technical workstation should be to provide a graphics system suitable for technical applications. The graphics system should be as close to the application as possible.

Two basic approaches to graphics are used in workstations, bit-mapped and geometric. Bit-mapped graphics was developed for page imaging applications and has operations tailored for image manipulation of single bit images. It does not extend conveniently to color. Geometric graphics was developed as a tool to assist in the design and analysis qf engineering problems. Geometric operations are tailored for the manipulation and display Of geometric objects in the frame of reference of the designer It assumes that the image is a rendering, or artifact, of the geometric design process, not the object of it.

Bit-mapped Graphics Bit-mapped graphics was first developed at Xerox PARC in 1974 for the Xerox Alto. This was the first professional workstation. Because of the nature of Xerox's business, it was designed primarily for page layout and imaging applications — thus it used a bit-mapped raster display. The bit-mapped display is used in these applications because the image of a page to be produced on a laser printer is constructed from a large two-dimensional array of binary digits.

In the Alto, one microprogrammed processor did everything — it was the graphics processor, disk controller, ethernet controller and central processor all in one. High-levei^ language compilers were written for the Alto, and most programmers used them, but when something had to run very fast, custom microprograms were implemented for it.

A very important one of these custom microprograms was the "bit-bit", which means Bit Block Transfer. It was developed because of the relatively large amount of time required to draw, redraw and manipulate a page image on the Alto. With the bit-bit in microcode, one could more quickly modify a bit-image. Because construction of the bit image was relatively time consuming and generalized, bit-bit operations allow the image to be "un"- altered with the "exclusive or", and the image was infrequently regenerated.

~~- 8 -~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 The focus of bit-bit or bit-mapped graphics systems is therefore in creating, manipulating and storing a bit-image, as in page layout systems. They are specialized for such applications and do not use specialized, higher-level processors to assist in manipulations of geometric objects, such as most engineering applications require. Their primarv benefit is that they typically utilize a single processor for everything, thus reducing their costs. The Apple Macintosh is the best realization of a bit-mapped graphics workstation.

Geometric Graphics Systems Geometric graphics originated with Sketchpad, which is a graphics system implemented on the TX-2 computer at MIT in 1964 by Ivan Sutherland. As an engineer and scientisf, Sutherland was interested in providing a graphical interface that allowed interactive, computer-assisted design of geometric engineering objects. In Sketchpad, objects were designed and manipulated in terms of the application's geometric coordinate system, or as we commonly say now, "user coordinates", "user space", or "world coordinates." The image created by a geometric graphics system is simply a rendering of the geometric objects being designed and manipulated — it is not an important entity in isolation, since it can easily be regenerated from the geometric description.

Fundamental geometric notions in use today were defined in Sketchpad. Because the geometric objects are described in one coordinate system (the user coordinates of the application) while the renderings of these objects are described in another (the image or screen coordinate system of the display), the notions of "windows" (user-space), "viewports" (screen-space), and a means for "mapping" from one to the other were defined. Likewise, because a window on the user's drawing space might include only part of the whole drawing, "clipping" of information had to be performed before the mapping. Finally, because the user's application space is geometric, objects might be arbitrarily translated, rotated or scaled ("transformed") in the space and multiple "instances" of Objects might be made before they are passed through the window, undergo clipping and SO forth. Although Sketchpad was primarily a 2D system, it defined these concepts in a general way that was applicable to 30 design as well.

~~Design as envisioned by Sutherland was interactive and in real-time. Since the clipping and mapping functions defined in Sketchpad were very compute intensive, Sutherland~~

~~- 9 -~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 concentrated on hardware to perform the functions as fast as possible. The "clipping- divider" designed by Bob Sproull with Ivan Sutherland at Harvard University in 1968 was the first hardware implementation of a system to implement the concepts defined in Sketchpad. This device used SSI .semiconductor technology and by itself require four 19 inch racks. In 1969, Sutherland joined the faculty at the University of Utah and with David Evans founded E&S Computer Corporation.

Working as a student of Sutherland's in 1972, the author designed the first completely three dimensional interactive surface design system. This system consisted of two general-purpose computers (a PDP-10 and an LDS-2 processor), a matrix multiplier, a Clipper and perspective divider and an analog vector generator. The hardware was designed for the project, and it used more, modern MSI technology. The systems output was directed to a dual-CRT head-mounted display designed by Sutherland. It used as input a mechanical, 3D "wand." The total system, excluding the PDP-10, required four nineteen inch racks of equipment and if produced and sold would have been priced at about $250,000. The geometric computing part of the graphics system was valued at about $70,000.

Starting in 1979, the author and coworkers at Stanford University began a research project with the goal of designing and implementing the first VLSI chip dedicated to these geometric functions. This chip, called the Geometry Engine, is shown in Figure 3. Twelve copies of the Geometry Engine chip arranged in processing pipeline shown in Figure 4 perform more than ten million floating point or integer operations per second.

The Geometry Engine is much more than a fast floating-point ALU. In addition t» incorporating all of the processing necessary to do each of the geometric operations discussed previously, the Geometry Engine also performs perspective division, and rather than processing just vectors, it processes solid, color, and shaded objects. In addition, it generates parametric cubic curves and rational bicubic surfaces, which are important in the design of automobiles and aircraft. It is a customized floating-point array processor for graphics. Because it is implemented in custom silicon, the Geometry Engine is easily replicated and costs very little.

~~Floating-point chip sets currently available provide alternatives to the Geometry Engine, but they do not have floating-point divide — they provide only the basic~~

~~- 10 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 PAD* D u W~~

~~D Ht~~

~~I/O t~~

~~P A O I~~

~~Figure 3: Geometry Engine Floor Plan~~

~~•Hifw CnQmsB •~~

~~' . ' OtpmtK^^imt~~

Figure 4: Geometry System add/multiply/subtract math operations. Divides require other chips and support logic, and each divide requires substantlallv longer than the muttiply, since it is a table lookup followed by at least three multiplies to obtain the requisite precision. Divides are essential for both clipping and 3D perspective geometric graphics. Moreover, these math Chips alone incorporate none of the algorithms for processing graphical data or manipulating solid objects with shading. Thus, in order to accomplish the Geometry

- 11 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Engine's set of functions, substantially more hardware would be required in conjunction with these chip sets. The result is both more costly and slower.

User-space or Screen-space From the above historical account, we see that the primary difference between Bit mapped graphics and Geometric graphics is that Geometric graphics systems provide primitive operations that are relative to the coordinate system of the geometric application, while bit-mapped graphics systems focuse on primitive operations that are relative to the screen coordinate system of the bit-image application. Typical geometric graphics primitives are rotate, translate, scale, line, polygon, surface and so forth, all in the coordinate system of the geometric application. Typical bit-mapped graphi&sj primitives are bit-bit, rectangle copy and so forth, all relative to the coordinate systerh of the raster hardware.

With the exception of the IRIS workstation by Silicon Graphics Inc., all workstations on the market are bit-mapped graphics workstations, and most of them are also black and white rather than color. This is because 'bit-bit" and such operations are binary Operations, meaning they deal with operands that are rectangular arrays of pixels that are a single bit each. Even those that are color treat the bit-planes that stack up to form a set of color planes in a bit-mapped manner. Those that incorporate Geometric graphics do SO in software, with differences in speed by factors of several hundred.

Importance of Graphics Speed Why is speed so important in a graphics system, be it geometric or bit-mapped? Some applications, such as simulation, absolutely require it. Others simply may need it tcT improve productivity, since workstations are interactive systesm, which implies that the user is waiting for a result to appear on the graphic display. If a graphics system cannot respond interactively, the user must do the task differently.

Some arguments occasionally support doing the job in a different way — to use a more "batch" oriented approach. Sometimes this is the case — interactivity does not replace thought. Batch-processing arguments were commonplace in the early days of timesharing, but "interactive engineering" is "productive engineering" in most of my experience, and if the graphics system can be produced for the same cost but much more

~~- 12 -~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 speed, the user is more productive. As Figure 1 points out, user time is now more valuable than computer time. With a slow graphics system user time is wasted, not computer time.

Why Design Our Own VLSI Circuits? Fast geometric graphics makes no sense without special-purpose hardware to do the computations required. However, specialized hardware always adds extra cost. In the general-purpose computing area, microprocessors have ushered in the era of low-cost computing. In the geometric graphics domain, the Geometry Engine has ushered in a new era of low-cost geometric graphics.

More generally, however, the modern system designer must be able to design her/his own integrated circuits, because technology has provided the ability to put very complex subsystems on a single integrated circuit. Carver Mead has'used the. following argument: "In the limiting case, technology will allow us to put an entire system on a Single integrated circuit. We will therefore either design that system ourselves or buy it from someone else." Although it is difficult to conceive of a complete system on only one IC, the argument provides an interesting perspective. The essential point is that as systems get larger it becomes increasingly difficult for the semiconductor manufacturer to second guess what the system designer would like for building blocks. Moreover, the building blocks require small scale circuits to "glue" them together.

Gate arrays and standard cells only partially solve the problem. Gate arrays do not currently allow very high densities. Perhaps worse, they do not currently provide a means for integrating RAM and ROM with logic. Standard cells, on the other hand, while allowing higher densities, impose a certain architectural style of thinking on the designer, since they are primarily used to replace the MSI building blocks that designers have been using. Each of these techniques allows faster implementation than full custom design, however.

Perhaps the biggest advantage of full-custom integrated circuits is in the architectural freedom they provide. This freedom can be exploited to employ parallelism in ways that are rare or nonexistent in other integrated circuit techniques. The Geometry Engine is one of the better known examples, but other highly parallel architectures dedicated to various special problems are certain to be in the future. Silicon Graphics Inc. will ensure

~~- 13 -~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 that the workstation marketplace benefits from highly parallel architectures for graphics and computing applications. System design now implies integrated circuit design The VLSI era is the era of the system designer.

Conclusions and Summary From the foregoing remarks, the technical workstation should be a balanced combination of general-purpose computing and either bit-mapped or geometric graphics, depending upon the application.

For the computing part, demand paged UNIX is the preferred operating system, and it should provide both System V and 4.2 BDS utilities. For compatibility with the wides{^ variety of other systems, Ethernet provides a more open network architecture than a token passing scheme, and both IP/TCP and XNS communications protocols are desirable. The communications file sharing environment should be as transparent as possible, allowing the "virtual system call" if possible. The ability to login to "foreign" hosts as an intelligent graphics or text terminal and transfer files are also important.

For the graphics part, pure page imaging applications are perhaps better served with Bit-mapped graphics systems. Most technical and engineering applications are better served with Geometric graphics systems. The geometric graphics system, being much higher level, requires extensive software to make the graphics easily used by the application — drawing surfaces, curves and color, geometric objects in 3D user space should be as simple as drawing a line in screen space or modifying a pixel. Since technical work requires documentation, the system should provide means to produce it, allowing a mixture of text and graphics. The graphics system should also provide ^ multiple window environment that allows the user to multiplex between several tasks.

Speed in both computing and graphics are important, but slow graphical display tends to make the system non-interactive in engineering design and analysis. With a slow graphics system user time is waisted, not computer time. Silicon Graphics Inc. is dedicated to seeing that both the computing and graphics speed needed by engineering applications are treated on an equal footing.

~~- 14 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dataquest Dataquest~~

~~THE ROLE OF PROPRIETARY ARCHITECTURES IN WORKSTATION DESIGN~~

~~David Folger President/Chief Executive Officer Ridge Ccanputers~~

Mr. Folger is a founder of Ridge Computers and has served as President, Chief Executive Officer, and Director since the company was incorporated in 1980. Prior to joining Ridge, he worked at IBM Corporation, at Zilog, Inc., and Exxon Enterprises. He received a B.S. degree frcxn Harvey Mudd College, and an M.S. degree in Electrical Engineering from the University of California at Berkeley.

~~Dataquest Incorporated WORKSTATION FOCUS CONFERENCE August 13 and 14, 1984 Burlington, Massachusetts~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 1~~

~~CD a. > CO~~

a. 0)' >5 o^ THE ROLE OF PROPRIETARY ARCHITECTURES > o 2. 0) IN WORKSTATION DESIGN Z3< ro AUGUST 13 & 14, 1984

~~05 0)~~

~~O (1) CD o > ,,^RIDCE I~~

~~H (D~~

~~U D 0) ST c (D~~

1 RIDGE PERSONAL MAINFRAMES (D a. > w • Mainframe computation performance for individual c CT 0) 5! users B)' >< o^ > > Efficient handling of large programs and databases p z (D' > High performance, interactive graphics o o > Software compatibility: • UNIX* System V (with virtual memoryl Io High performance network capabilities

~~CO (U > Office environment compatibility 3 > High reliability o > UMX li a tmd*nur1( or Ba L«bonb>rtM IWUMI to a tndmurfc of Xldak, Inc. ^^j.~~

~~o.b . 3~~

~~CO -vl~~

~~CoO o o~~

~~(D CD o Eo n c (D in o O -1 •oo RIDGE NETWORK SYSTEM > ARCHITECTURE O)~~

~~9. 0)' o_ > p Computation z S' Server~~

o Local Area o 3 Network x> 0) Databaae s Server XI OJ d Q. I Single User ± CD o o T3 Ridge WS 0) Ridge O Color 32 I Color W B> C O (0 (0 [rod Single User O> Engineering Cluster Workstation rM^mQE I

~~~j CO ~uJ i o o -1 •o o oa~~

~~0) c INSTRUCTION ARCHITECTURE CJ w a 0) >5 o^ > > RISC - Reduced Instruction Set Computer b • One instruction per 125 nanosecond clock~~

o o • No limits on addressability 3 •o 0) • Register-Oriented ro 5 I 3) • 16 General purpose registers a *• 6 I • 16 Special registers for kernel use -1 •D B> • Load/Store memory reference format

~~(n 3 • Clean instruction abort~~

~~o > • Compare and branch instructlorts • Branch prediction I i^RIDCE o(O o H (D o B ST £1~~

~~3 O O •3 o RIDGE 32 CPU INTERNAL STRUCTURE > w Instruction c CT a Fetch Unit Q! B)' •<—\ o —«. > b~~

~~o o 3 •o 0) 3~~

~~l>0 (O o I 3) d Ln Q. I (D T3 D> o~~

~~05 0) 3~~

~~o > (O Ol~~

~~.(^ i^RIDCE o ro (O --J I~~

~~-^l —A CO -^J CO D 0)~~

~~V) c cr ROS OVERVIEW CA Q! B>' •—<1 o^ Ridge UNIX > UNUC UtiUtles^ o UNIX Implementation~~

~~<5_ System (0 (D 3 Calls o o 3 •a 01 3~~

~~to CO I o cr\ ol I Q. (D TJ 0) 5r O~~

~~Iy) 0>~~

~~o en ? O > Ol UNIX File Structure~~

~~rJ^RlDGE~~

H (D f 3 O O -^ •o o (D a > RIDGE SOFTWARE ENVIRONMENT w • UNIX System V with virtual memory I • Berkeley UNIX 4.2 enhancements ° 4 G-byte code space and 4 G-byte data space o • Linear addressing with 32-blt addresses o 3 •o B> • Shared code among multiple processes; private data >< fi space for each process s I Q! I I • Paged virtual memory TJ lU • Message-based Interprocess communications: • Asynchronous and network transparent W a> 3 « Multiple Incoming message queues • Links to other processes' message queues o > > Short messages 132-bytesl - Long messages for data and file access |4K-byte pagesi

~~CO -J CO o n c a in~~

•a o RIDGE 32 HARDWARE cW 5 o^ > Up to 8M byta main memory b 60M, 142M, 445M byta Winchester discs «> o o 3 8" floppy disc •o lU RS-232 and printer/plotter ports r>o oCO I a 0° Options lup to ninel: % I —1 •D fiJ " High resolution monochrome and color graphics • Ethernet

~~CO IB • Magnetic tape system 3 • Parallel Interface o > i^WDCE o 3~~

~~U3 ~J CJ o I 3 O o •3 o IPC UNIQUE FEATURES~~

~~> w c a u Short messages in registers a 5' >3 o Long messages copy page pointers > b Static linlcs minimize overhead 0>~~

~~C- O 0> a >O (O en i^nDGE §~~

~~o(O o o~~

~~CO -uJ o u n» c (D 5(A" o •o3~~

~~(D a > INSTRUCTION FORMATS CO~~

~~Q. S" •5 o^ Register > p 0 7 8 11 12 15 z §! 0) (P o o 3 •a D> Short Displacement~~

~~CO 15 o 0 7 t 3^ a. O a. a I OPCODE R1 R2 Displacement T3 B> 1 -^ 1 Long Displacement (n 0) 15 3 0 7 t *^ o > OPCODE R1 R2 Displacement to OI 1 ^ 1~~

~~•s i!^RIDGE~~

~~2. (D~~

~~(O u o »to .cO (S u 5" o o —t •oo~~

~~(D > CO c cr RIDGE 32 CPU (0 a~~

~~o^ INSTRUCTION PIPELINING > b~~

~~to (O o I I o~~

~~w D> c_ O u> ? O > (O tf^uctlon Ol~~

~~i^RIDGE oCO o o~~

~~CO ->l CO o B n c (D (5A" o o-» •oo~~

> O) c RIDGE 32 VIRTUAL ADDRESS TRANSLATION CT CO a S' 16 32 •-<1 o^ > segment :i offset p z 2. MAIN MEMORY u> O o 3 T3 VIRTUAL TO U Translation •< REAL TRANSLATION I r mapping (O ' TABLE o table i K (D I T) MEMORY CONTROLLER CO CD 3 C_ o (0 oCD > to (1>

~~o .po CO -.1 {Located In main memoryl oto i^RIDGE~~

~~to -•I CJ o » B .a c CD (o0 O •o~~

~~(D a MONOCHROME GRAPHICS w c SYSTEM IT U 9: •»)3• o > b Rfdge Graphics Main o Central Display 3 Memory •o 0) Processor Controller :3~~

~~NJ (O I O~~

~~3) Ul i E °- I CPU Memory Bus I/O Bus (D ' —1 TJ a> Page Out~~

~~CO 0) 3 Page In o > ri^RIDCE .1~~

~~oto o~~

~~H (D_ (D~~

~~U o lU r+ E c (3D O o -^ •o o~~

~~> w~~

Q. D> RIDGE 32 INTERNAL LAYOUT •3 o_ > o Backplan* 8" Floppy (D 13 Piovldtt I/O BUS 1.2 Mbyte o O3 Fetch •o 0> Extcuto 3 Memory confaoler 60 Or 142|lKiyle 8 Card cage 14" Disc ad (D I -1 TJ 0> jr Air plenum Uaock O W/duei squirrel cage blower uCO 1 Mt>yte memory array c_ O Universal Mertece (A n> Disc contrder o > CPU Private (£> Ol Memory BUS

~~.b. O 19 Expanskml 3 (O -J 8 o o i^RIDGE~~

to u I o o •3 o a> p. SOFTWARE VISIBLE MODEL OF PROCESSOR > w c CT cn ol B>' Instruction types: •-<^ OPCODE Rl R2 Displacement O > Load/store b Register-register (D CD General (D 3 o o Purpose 3 (No user visible status •o 01 Registers or condition codes! I. Ul Program Counter I o 32 CO 0> ^1 o M (D O> (O Code/ Memory CJI Data Address i^RIDGE I

~~CO ->I U -.iii-i-i ' ?v :)aDataquest Dataquest~~

~~SYSTEMS ARCHITECTURE—UTILIZING STANDARDS~~

~~Scott McNeally President Sun Microsystems, Incorporated~~

Mr. McNeally is President of Sun Microsystems, Incorporated, which he cofounded with three colleagues in 1982. Previously, he was Director of Operations at Onyx Systems, a microcomputer systems manufacturer. Prior to that, he was a member of the corporate manufacturing staff at FMC, and held various operations and sales positions at Rockwell International. Mr. McNeally received a degree in Economics from Harvard University, and an M.B.A. degree from Stianford University.

~~Dataquest Incorporated WORKSTATION FOCUS CONFERENCE August 13 and 14, 1984 Burlington, Massachusetts~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 SYSTEMS ARCHITECTURE UTILIZING STANDARDS~~

~~SCOTT MCNEALY PRESIDENT SUN miCROSYSTEmS, INC.~~

~~JULY 1984~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Parl< Drive / San Jose, CA 95131 / (408) 971-900) / Telex 171973 Architecture of a Scientific Computing Environment~~

Abstract Scientific computing In this country needs a combination of technologies to accomplish research and development goals In the next few years. This article proposes an architecture for the scientific computing environment using microprocessor-based personal workstations accessing back-end computing resources. Systems built by multiple vendors will be connected using local network technology and standard network protocols. This paper describes the architecture of this new environment and the role UNIX will play In Its emergence and growth.

Introduction To provide a quality scientific computing environment- wa need to characterize the computing needs of scientists and engineers. We need maximally producth/e computing environments to offset the shortage of tralned engineering and scientific personnel. We need productive engineering environments to compensate for the decreased lifetimes of products. We need an architecture which allows us to Incorporate new technology easily and with minimal Impact on users of the system. Professionals need a responsh/e tnteracth/e system, the ability to run large computations and. In a laboratory environment, the ability to collect and analyze experimental data. Inexpensive microprocessor-based workstations with bitmap displays, leveraging the low cost of memories and virtual memory microprocessors such as the Motorola 68010 and 6S020, provide cost-efFecth/e Interactive response. Local area networks can tie such workstations to back-end machines: large supenmlnlcomputers, malnframes and supercomputers. Peripherals such as disks, tapes, communications Interfaces and printers can likewise be shared by accessing them over the local area networic. Data collected In a laboratory by a suitable embedded microprocessor can be moved to a personal workstation for graphical Inspection, stored In a network-accessible data base, and processed on back end rnachlnes.

No single vendor can produce the component technologies to make this system cost-effecth/ely. It Is not possible for a single organization to simultaneously track advances in languages, user interfaces, data bases, graphics, specialized high-performance symbolic and numeric architectures and expert systems, let alone the many application areas for these component: technologies. Multi-vendor systems will be required and must depend on standards to Interconnect the components.

Architectural principles for the new environment We now list seven architectural principles we see as key to the construction of a new multi-vendor open system environment. Wa seek a conservath/e synthesis o( proven kieas so that we can have a high degree of confidence In the viability of the system architecture. All the principles and approaches proposed here have been proven in systems which are In use today; we have avokjed Ideas whose long-term value Is unproven.

~~1 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Architecture of a Scientific Computing Environment~~

1. Take a network service approach We can Identify two ma|or approaches to connecting a network of personal machines. The first, which we call the network operating system or closed system approach. Is typified by the Apollo DOMAIN system [DOMAIN] or the Locus system defined at UCLflk [Locus]. These systems represent an attempt to extend the monolithic operating systems wtJch were present on malnframes and superminicomputers to the local network environment. They tend to be inflexible, with a strong advantage to components of the system provkfed by the primary vendor over "foreign" peripherals. The other approach is a loosety-coupled and open system design whiclx gives network services primacy over the network operating system. An Important system which first took this approach was the Pliot system [Pilot]. This model of a network system has also been incorporated In the Beri

The 4.2BSD system contains sufficient facilities in the system inte'rface that an open network system can be built without changing the systerii calls. Open systems dertved from 4.2ESD and based on the network sen/Ice approach are being constructed at Sun Microsystems and other vendors. 4.2BSD provides a portable and slandard operr-systems base for the new scientific computing environment.

2. Share resources while minimizing competition Our measure of system perfonnance sees quality In both predictable lov^-latency response and in high throughput. These are often contradictory goals. Dedicating peripherals gh/es low-latency response and high throughput, but Is not cost-effectlve. In centralizing services we place predictability and responsh/eness of the servers, and thereby the quality of the system, at risk. A central file server or compute server will be practical only If It does not become overioaded and thereby defeat predictable workstation response. Sharing access to devices on a local network Is essential to cost-etfeoth/e computing. Disk drives, tape drives, printers and other peripherals are subject to economies of scale. Centralizing file servers provides a slgnlilcant cost and malntenance advantage over systems where each user has only a local disk. For example, the larger capacity disk drives are both cheaper per megabyte of storage and also faster than the small, Inexpensh/e disk drives which one can affonJ to provide with each workstation. Even tf the costs were the same, conskierations for shared data access and the difficulty of determining the amount of disk storage required by each user make it difficult to effectively provide each user with local disk storage. Architectures for file server nodes to maximize throughput are an area of cun^ent study. Effective use of cache memories at server nodes shouki eventually result in server-based systems with performance equalling or surpassing that of systems which provide smaller and slower peripherals to each user. Nodes providing compute service can be scheduled in a Jot}-shop or batch mode fashion, rather than timeshared. to Increase predictability, and minimize competition, at the expense of some flexibility.

3. Provide a uniform base level of services The Simplest form of connections between machines is to allow file transfer, remote login and remote command execution. Provkllng these services to all machines in an

~~-2 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Architecture of a Scientific Computing Environment~~

environment with uniform syntax shouU be a first priority. This level of service Is the basis Of the success of networks such as the ARPANET. Having this uniform level of services is far more Important than any of the services described below and should always be provided for all possible machines. Small, single task personal computers such as the IBM/PC or the Macintosh will have client telnet and FTP implementations so that they can access any file on the network, back themselves up to any other machine, and act as terrnlnals to larger hosts. Larger hosts can exchange files in a symmetric manner, and a terrnlnal attached to any of the larger hosts can serve to log In to any other hosL The DARPA standard TCP/IP/UDP (stream and datagram) transmission protocols. FTP (file transfer) and Telnet (remote login) protocols provide a widely accepted and Implemented standard for this base level connection. Ethernet provides a high-bandwklth and Inexpensh/e, vendor-Independent hardware interconnect.

4. Use standard operating system interfaces In the microprocessor market there are several Important operating systems. In the 8-blt worid we have Apple/DOS and CP/M. In the 16-blt world MS/DOS, and In the 32-blt world UNIX and the new fvlaclntosh operating system. Each of these defines an applications program interface for which large numbers of applications exist. Providing users access to one or more of these system interfaces can bring along a large number of applications at low cost. UNIX Is the most Important operating system standard, because It runs on a wide range Of machines. UNIX meets two needs: the need for a standard applications Interface, and the need for a systems building block for constructing a open network system. The last version of UNIX to come from the research group which originated UNIX. Version 7. defines the basic applications Interface for all widely used versions of UNIX. Applications programming features added In System V are not nearly as essential as the Version 7 facilities. We believe that the slightly extended Version 7 Interface, as codified In the vendor-Independent lusrtgroup standard. Is the best basis for an applications program standard. Further evolution of the system facilities from this base (with the exception of networking facilities) Is, In general, unnecessary and undesirable. 4.2BSD Is the only version of UNIX which defines a standard and stable yet extensible interface to networking facilities. Experience to date with the system suggests no major Obstacles which need to be rectified. It will be several years before an altemative standard Interface for a networked UNIX system can be developed. This suggests that adoption of the 4.2BSD network interface by the Industry would be a positive step. Since an Implementation of the Interface Is readily and freely avallable. It has already been adopted by a number of vendors.

5. Implement processes, monitors and remote procedure call We believe that the extension of current systems to provide multiple processes per address space and monitors for synchronization shoukj be the highest priority area of extension to 4.2BSD UNIX, because complicated Interacth/e programs and networic services are most naturally expressed using concurrency. Process/monitor technology as implemented In [fviesa] and [Modula-2] has been proven to work quite welt to meet these needs. Operating systems such as UNIX do not provide support for these language facilities today. These faclltles will be needed to be provkled to support more modem languages, including ADA. The facilities can be added as libraiy routines for use with okjer languages such as C. Extensions to C and UNIX to achieve this have been designed and implemented at Sun; we Intended to publish them after further review, bnplementatlon, and use [Nugget].

We prefer to use these tightly coupled processes together with remote procedure calls [RPC] to bulkl our distributed systems. This provkjes a clear distinction between local and remote programs. It also allows substantial hardware simplifications: with coupled

~~-3 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Architecture of a Scientific Computing Environment~~

processes sharing complete address spaces, they avokj the need to share indh/kjual memory pages, avoiding based pointers, and allowing simplified hardware architectures as we can use data caches which do not need reference to translated addresses In the critical timing paths.

6. Have a uniform file system interface Most operating systems now provide hierarchical file systems similar to the one provided In the original UNIX [UNIX]. Providing a network sen/Ice which makes all flies In the network environment available transparently (s an enormous advantage. Operating system Independence In the protocol fs very desirable because, for example, we would like to be able to access the file system both from IBM PC's, Macintosh'es, and more powerful workstations such as the Sun. and use the protocol to access flies stored on other systems such as VAX/VMS and IBM malnframes (that is. under VM/CMS). A networked file system based on these principles Is being constructed at Sun Microsystems.

A dangerous trap to be avokled here Is Indefinite extension of the protocol for the network file system. We believe that the level of functionality of the 4.2BSD UNIX file system Is adequate for the basic network file system, and that additional functionality such as that described in [Glfford] can be added transparently by different implementors of networ1( file systems, without affecting the client code.

7. Access data base services using rpc protocols We suggest provkJIng data base access through applications-level network sen/ices, rather than as part of the operating system facilities. This allows a range of Implementations of data base services based on applications needs, and can yield modular and well-specified systems. The distributed version of MIcrolNGRES, as currently avaltable in a network of Sun Workstations, Is an example of this style. Both high-performance and high-avallablllty systems can be constructed atop an efficient remote-procedure-call mechanism in a distributed environment. This architectural approach also allows construction of highly avallable or reliable data storage systsms [Glfford] [ISIS].

Conclusions The most Important goal In creating a scientific computing environment Is ensuring predictable response time. In a professional environment you often need access to more cycles than your workstation can supply, and these cycles can be cost-effecth/ely supplied by back-end compute servers. A loosely coupled system permits a cost-effective mix of woricstations and server machines to be used. To bulkl components for the next generation scientific computing system we must start by taking a network service approach. The most suitable base operating system to use is 4.2BSD, as Its networking facilities can be used to write network applications and protocols without changing the programming Interface of the system. Back-end computational and server nodes can be placed on the network and scheduled to provide predictable response time. 4.2BSD comes with remote login and file transfer services which can be extended to other machines in the networ1( because the standard TCP/IP protocols on which they are based are avallable for a wkle variety of machines. The 4.2BSO system provkles a standard applications Interface, derived from UNIX Version 7. which meets the needs of most applications. You can look to the emerging vendor-Independent lusr/group standard for a definition of the common set of facilities to be provided. There is little to be galned by additional evolution of the programming interface beyond the facilities of 4.2BSD: most further evolution can take place using appllcatlor>-level protocols.

~~-4 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company 11290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Architecture of a Scientific Computing Environment~~

To support high-perfomiance network services, we win need to extend the C language and program execution environment with support for lightweight processes and monitors for synchronization. Remote procedure call can be used to connect processes on possibly different machines, and is a natural extension of current programming styles. The UNIX file system abstraction, as defined In Version 7 UNIX, and made less UNIX-speclfIc In 4.2BSD. Is a suitable basis for a network file system facility. We can avoid the temptation to make continual changes to the file system abstraction by dealing with data base services such as locking, transactions and flne-gralned protection using applications-level protocols based on remote procedure call rather than building such functionality Into the basic network file system. A cost-effective and flexible system can be constructed, using the principles summarized above, without Inventing new hardware or software technology, using 4.2BSD UNIX as a base. Forthcoming advances In microcomputer technology, cost reductions in memories, and high-performance supercomputers can then be easily integrated Into our working environments.

~~-5-~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Architecture of a Scientific Computing Environment~~

References DOMAIN Nelson. DavW L. and Leach Paul J. The Evolution of the Apollo Domain. Hawaii International Conference on Systems Science. 1984. Describes a popular network operating system. GIfford GIfford, O.K. information Storage in a Decentralized Computer System. Xerox Computer Science Laboratory CSL-61-8. June 1931, revised March 1982. 'Describes how to build storage system services for use on a local network with Interesting properties (relocatabllity. security,, reconflgurablllty. etc.) IDM Br1ttor>-Lee, Inc. IDM Product Description. A description of a high-performance relational database server. ISIS Bimian. K.P.. Skeen. D.. Abbadi. A. El. Dietrich. W.C. and T. Raeuchie. fsis: An Environment for Constructing Fault-tolerant distributed systems. Comeil Unh/ersity Department of Computer Science. Technical Report TR 63-552, May 1983. A description of network services providing fault-tolerant operations. Locus Popek. G. et al. Locus: A Network Transparent. High Reliability Distributed System. Proceedings of the Eighth Symposium on Operating Systems Principles. December 1981, pp. 169-177. Another network operating system, which hides all details of the local network and Includes transaction processing capabilities. Mesa Lampson. B. W. and RedelI. D. D. Experience with Processes and Monitors In Mesa. Communicatons of the ACM. 23, No. 2 (February 1980). pp. 105-117. The seminal paper on a process/monitor based runtime system. Modula-2 WIrth. Niklaus. Programming In Modula-2. Springer-Veriag. 1982. Modula uses monitors for synchronization, and can build processes from a coroutine mechanism. Nugget Joy. W. Nugget: Parallel Processing Runtime Enhancements to C. Sun Woridng Document. May 1984. Describes a set of library routines for C programs which provide lightweight processes, monitors, condition variables and an exception handling mechanism, providing facilities much tike Modula-2 or Mesa to C programmers, Pilot Radell. D.D.. Dalai. Y.K. Horsley. T.R., Layer. H.C.. Lynch. W.C. McJones. P.R.. Murray. H.Q.. and S.C. Purcell. Pilot: An Operating System for a Persona) Computer. Communications of the ACM. 23. No. 2 (February 1980). pp. 81-92. An operating system for a personal computer which emphasizes a network service approach.

RPC Nelson. B.J. Remote Procedure Call. Xerox Computer Science Laboratory CSL-61-9. May 1981. The seminal thesis on remote procedure call. Standards Pratt. V, Standards and Performance Issues In the Workstation Market. EEE Computer Graphics and Applications, 4, No. 4 (April 1984), pp. 71-76. Emphasizes that standards emerge not only from committees, but also by common usage and In the marketplace after natural selection. UneedS Dallalre, Q. American Universities Need Greater Access to .Supercomputers. Communications of the ACM, 27. No. 4 (May 1984). ppp. 292-299.

~~- 8 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Architecture of a Scientific Computing Environment~~

UNIX D.M. Ritchie and K. Thompson. The UNIX Time-Sharing System. Communications of the ACM. 17, No. 8 (July 1974). pp. 365-375. WandS W.N. Joy. Riding the Learning Curve: The Sun Workstation Architecture. Proceedings of the Seventeenth Annual Hawaii International Conference on System Sciences. 1984, pp. 480-484. A paper from which the current article Is derived, which proposes that the way to build cost-effective scientific computing environments In the 80's Is with workstations and supercomputers.

~~-7 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dataquest Dataquest~~

~~THE UTILIZATION OF ARTIFICIAL INTELLIGENCE IN DEVELOPING USER INTERFACES AND NEXT-GENERATION WORKSTATIONS~~

~~Richard Rifenburgh Chairman of the Board—Chief Executive Officer PERQ Systems Corporation~~

Mr. Rifenburgh, in his role as chairman of the board and chief executive officer, has developed PERQ Systems Corporation into a leading high-performance graphics systems company. He brought with him over 25 years of experience in the computer industry when he joined the company in 1983. Mr. Rifenburgh was formerly president and CEO of Mohawk Data Sciences Corporation. Since 1968, Mr. Rifenburgh has also founded or served as a director for a number of public and private companies including Rand Capital Corporation, a venture capital concern, and Sage Systems, an electronic office products manufacturer. Because of his expertise in the financial as well as the technical side of the computer industry, the Computer and Communications Industry Association elected Mr. Rifenburgh its chairman in 1984.

~~Dataquest Incorporated WORKSTATION FOCUS CONFERENCE August 13 and 14, 1984 Burlington, Massachusetts~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o fi>~~

~~o «• >• • •3 .Hti tt ' ' o Ju,;/ t. •• - Ji f.', !). W B " • %• .^' 4 CI w Itf b * * C> ^ I'i i' tt ii &v 4. •~~

~~CO ^it5';«*i-iI(j»' I a > p THE UTILIZATION OF ARTIFICIAL INTELLIGENCE IN DEVELOPING USER INTERFACES~~

~~IV> I~~

~~I3) - AND NEXT GENERATION WORKSTATIONS S I a « -, TJ fi>~~

~~W lU 13 oC u o >~~

~~.c>. o CO oCO o o~~

~~CO --I CO 2 O P~~

~~o c/5 f^ C i 3 .2 tj S a -- « S * [jj -S •= "£ P~~

•^

~~O CJ~~

~~- • ""i. • ' s SHV • I •• ^ < -if J • > a X a V, u ^ • • •• • « 9 3 < ^ a » • « a 9 a .^-.SJJ • a -= J«J4«SS' *'.5»'i -.^f^igr "•-•.?«•-~~

~~- 2 -~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o S. BT n c a a> o «.» . •a u ^ >• o ,•.•_.. , £ i- -.i i' •...«••>•

~~•fc i.; J.. ;. ii 3i ,'j . 1 • > '4.-.% .M- :-"••• : '.' w ^•!tlfc:k,.:i.';-4P' 5- '•ttrtuaf' EVOLUTION OF I PROCESSING > O~~

~~o o 3 Mechanical calculators •D~~

~~Numerical processing TJ (u TT O Symbolic processing CO~~

~~O u CD o >~~

~~4^ o 00 oCO o o~~

~~CO -.1 CO o w~~

~~jcQ a (fl~~

~~13 n tt ^ ' 8 , xi«-. .^ % ji -„ .u *. O l» U V' i! :». V S bi«»«.l.u..>. ASPECTS OF ffi *Itt .*•»..§: I'l;, . . .~~

> ^UImil;».-.,,ji*' CO •vfctofc'' ARTIFICIAL INTELLIGENCE I a V Problem solving > o ^ Logical reasoning ? vv.i Language understanding vi. Programming i>o CO o v^. Learning Q. Q. (D -0 Expertise t:it Robotics and vision CO

~~o Systenms and languages u> 0) o >~~

~~IL CD~~

~~CO o 0)~~

~~o -J* - o .bat al • • •O A^l. MKl • • • ^ ltij!a:£wcx». g^ «<]te>ti£C. ^alj4Utl!lttr V) ^'^atif IT PROBLEM SOLVING (A~~

~~B)' •3 o > b 'ii^ Search techniques~~

~~o o 3 B> J i Exhaustive searches 3 VJ» to ' o -n Wl a Heuristic searches~~

~~w Backgammon, chess u 3~~

~~o >~~

~~o OS o(O o o~~

~~--4 (O~~

~~CO D fi>~~

o -• • * O .KiIKk ' ' •3 JliiSfeiM u > • ° tmHimbu • - E liiJ(iti<«>«> M • • * • g^ tIi<.<»i-ii a^ « > - > ^Kftaiauiir (/) "^vaHv*' LOGICAL REASONING a. B)' K Theorem proofs (D 0 o O3 •o a> 3 Data reduction a la I CD

~~•D 0) Q Database query~~

~~w tu 3~~

~~o >~~

~~(D O O O~~

D u s n c CD CO . •• » — .•« « ' g ' -3 liiiOiCkw'' k - O tt*S£Hl> M < < * - LANGUAGE » tiUiltliSifr MW« . • ^ ''^llts^^iiiJ' > ^•«aiti>»*'' w UNDERSTANDING c ' b # Natural language interfaces o o 3 •o B>

~~M I 5 ^ Speech understanding a ci I CD TJ 0> # Language translation CO 0) o CO ai~~

~~to -J~~

~~oCO o o~~

~~-I CD~~

~~CO ->I CO c s o V) 2~~

~~-a #~~

~~• • 9i. * S ; iC. • 1 !1 jJ » ^ ». i 4 3 U < ,'. ij • « ', M ^ ; -J j . . J - „ J ^ -.-2 y ^ ^ P =T St .1. c:^' '«.J! *••; * «:*«;*» *•* IJi^'~~

~~- 8 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 u o C G Ji <0 O 2 "a u •IV O 9 %^am s« ^ a> CA Ld CL, C~~

~~J -•^ 1 -.f~~

~~• • 3(v~~

~~• '•'•-'• z% • <*::•; i.^^i • • « ,• s .» •' ^ J * a -« .: ^ -.( '• j: y .11 3 * -•-. ."i .-.jj~~

~~"«•-•: \ '• ^ " * Y, Ss»'~~

~~- 9 -~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 C/9 a> V) ^ i 15 H 4> "T5 Oi UJ «> -J. eu M 4) X •• If <^ 2

~~5 •:!~~

~~'3. • ' .a. . ' . ^-.SS • > t i •• i •• n • • • «.1» 1 - .^ a • • < ji < -i >; .-C'J ^ a J 3 s j( .i iOJ •'^"5.3* ««^ •'MiB ji**"~~

- 10 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 2 -*

~~OiHirt c/5 c tma4 CA O > o • a« o c bfi o fe O o 2 0^ < "^ c • •< ^1^^ u (•^m C/3 e/) u *s 9 • la^ > o -o •O jC p o CO »' J --'*i O • .;? fiQ • * ^^ O~~

* ^i* ; .2 • ' -^ • J Si

* • " * -t -i !>, a ^.[^ m « « 4 3 -j 3 .. :/5jr • « 3 * B s^ (t-ai - •'aa'.'i'Siisar "•-jiS./"'

~~- 11 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 I~~

8 *»" •5 .«•••• o AAm»»'• s •iitaHi'X' • - SYSTEMS CD MRttl»»M« • ' • ^ >£<««» HH* • - I 'MIVH^'-J' CT AND LANGUAGES CO a 0)' •3 o > p Tools and ideas z SL 01 - Time-sharing CD o3 o 3 •o - List processing S) 3 i;; ' - Interactive debugging g M

~~Q. Q.~~

~~O - Lisp~~

~~W - Prolog D>~~

~~C_ O (fl - Planner CD o >~~

~~O 00~~

~~oCO o~~

~~iL (D~~

~~CO o 0)~~

(D (n o O [• •• - •5 o /...«.-• to .,;. ^ ••. Ma tf * -.. (D '^ ..1 B 1^ ff J. • fe -"^ £<<.' '• *« ;lfV M > CO %|g^,,. ARTIFICIAL INTELLIGENCE I SYSTEMS TODAY a > O >< ^ I ^ PROSPECTOR - SRI International

~~TJ ^ DIPMETER ADVISOR - Schlumberg^r 0) 3r O ^ DRILLING ADVISOR - TeknowIedge/Elf-Aquitaine W D> 3 o > (O Ol o.b. 00~~

~~I oCO o o 5. (D X —k~~

~~CO o 0)~~

~~o o <• • • •g ,BilU« • • j((tk>« » •• H'SSiuti a a > WU IS 1)1(1 •!•<•• •«i«!!uaiaiii.)!» ' -~~

~~CO c cr (fl «Sb«?' INTERNIST/CADUCEUS S' •3 o > b~~

~~(D 80% of all internal medicine o O 3 •D a 3 500 diseases~~

~~Q. ' CD 3500 manifestations of diseases T> B> -1 ?r O Preparing to undergo clinical trials CO~~

~~O 0> (D o >~~

~~o 00~~

~~(O o o o~~

~~•Nl~~

CO u t/» en 0 CO & k E "3 (A "7fl3 o (« g CO 4) bJ >> CO 2 ^^ .fc*0 «n IM •IM 4) «•« a >- 0. o be t/i ^ o • c•< ^•M a^n • fftf H "rt k u O O •& 1 U} ^ 1 O. "3 Od 4) od jo2 o ino 1 > tt: a > 01^ < O bJ O < H 0£ Z U u zo 0u« H O C/3 ZJ • 1 .. o . . 3^ • . 1 i::3v u S: • » » ^ 5 g * • J • a « iiia CL *.

~~- 15 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o ST .a c CD (0~~

~~o . • •~~

^ JHaaIi>> ^ «a«Ba>.- Q. BaMaaa* • • • - tiMBBBB**- PERQ WORKSTATION FAMILY ^ îiAaaaaar g> ÎBlllHaiir a "'•••IIW CO PERQ Al {LN-:^SOO) o! tu' > b !L 0) • C0nîii»& Lisp language CD 3 o o 3 •a (U • Supernuni perforntance

~~IV)~~

~~33 <^ #Low cost - less than $40,000 Q. ' CD~~

~~T3 B) • Based on ACCENT~~

~~W B> 3 «Operates with LINQ~~

~~o >~~

~~o(£> o o~~

CO o B C a CO ^ . * - 3 ^a • •• o ia«> • • .3 Jia«lia> • o atiilSiMINa • . § tttti&OCiaa " • • PERQ AI g^ «fe<%stiiliBa • a - IH^tiuw auii' > MfeLitekidwr CIO ^«iaiii« I.N-;i.50() CONFIGURATION I a > o 9 16K Writable Control Store a>KV a 3 o I * 2 Mbyte memory

~~10 (£> 1.^ « 43 Mbyte 5 1/4" disk~~

~~O « Landscape display~~

~~CO S> 3 c • Ethernet interface O tn (D >O (£> 01~~

~~O CO~~

~~to O O O -I CD o 0)~~

~~E ACCENT OPERATING SYSTEM~~

~~> w "•«iMW*>' USER a. a >—<1~~

~~> o WINDOW~~

~~(D MANAOER 0) (D 3 o o 3 •a su ACCENT j QNIX LISP ENVIRONMENT i ENVIRONMENT ENVIRONMENT -* I ro S(O ^|_i a ' ACCENT (D -D D> ] KERNEL 1.~~

~~:Z::J.:.._ d cn u ' WORKSTATION | o i i 0) CD O>~~

~~(D ~J CO o 5J n c (D CA~~

3 O O "§ <••••• s iHf"*"' o SSSSS^' a. .„••••••, • ACCENT OPERATING SYSTEM > ••!••••• CO LISP ENVIRONMENT I "••••' 5' >5 o^ > £E. bCO CD • Full Gomman Lisp implei^etitalion o O 3 •D (1> 3 • Compiler and interpreter lO

~~-n VO Q. CI I CD~~

~~TJ D) • Screen editor based on EMACS and extensible in LISP~~

~~CO a> c- o CA 0) o >~~

~~oCO o o~~

CO •«j CO f o o T) am > - ° Mmm* • E «««••-• S «»•••••- - •!•«•••••^, • ACCENT OPERATING SYSTEM > miMUMfelM^iiiii"mtmmtf'ifiv *r « a.CO >< QNIX ENVIRONMENT o^ > o 2. • Complete UNIX System V environment, (/> CD O licensed from AT&T o 3 •D 0> • Provides System V system calls >< i;5 '

~~3 O °^ I • Over 150 utilities (D -^ T3 0> • Use of the network is transparent~~

~~W D> • Standard UNIX shell C- O 0) 0) o >~~

~~o 00~~

~~I o(O o IE. CD~~

(D -J u D 0) i -^ j« • • o <•••• • S iHtmmm" S> «HKliii>* CQO- miliKaa* * • m;s;*•*•••<••' r ACCENT OPERATING SYSTEM a. oi' NATIVE ACCENT ENVIRONMENT > b !L • Client-server paradigm CO (V O o 3 •o 0) e Consistent user interface •< lO (O o • Screen oriented editor 31 E Q. CD Tl 01 • Over 100 utilities ?r O

~~CO Q> e Integrated on-line HELP facility~~

~~I o(O o~~

~~CO o 0>~~

o3 o •D - Ja « - ^ mmif ' Si Aitiiaa* • S fl>aK«aa- MMlHai*** • • • > taSMIIBaa.- g? . ^IIIINiaiiaiia^BHmiiiiHr ' ACCENT OPERATING SYSTEM (A "vvaHv*' a n)' NATIVE ACCENT ENVIRONMENT > p 2 <0 (0 • Client'Server paradigm (D 3 o o 3 •o • Consistent user interface a> 3 l>0

~~03 B) « Integrated on "line HELP facility 3 o >~~

~~o CO~~

~~(D oO o f 3 a o o .laitaa • • u AwivM* > •~~

~~a OF ARTIFICIAL INTELLIGENCE > p NATURAL LANGLAi..L ll^: I'LRI'ACLS~~

~~•^ Help systems~~

~~r ii» Database queries~~

~~Cfl 0> 3 ^ Operating system interfaces~~

~~o > to tn~~

~~I ! H (P (1) X o 0> J3 C (P Ul~~

3 .<• • ^o .•««• _o JMaaa-. O KliiKOK^- 5 KiKll>li M • • • *

~~> •«IL"H; CAD APPLICATIONS CO c cr (A 5" OF ARTIFICIAL INTELLIGENCE~~

~~Ul (P 3 o _^ I v^ Expert system lU 3 >< 1^ ^ Productivity aid a> I T3 01 5- O ^ Quality improvement w 0> 3 C- O CA n> O> CO c^~~

~~<3 CO~~

CO CO o B B n c (D o sm » • -3 iSaa' - o «•!••• - • u fiailaa* • - >(D ttMHMIllla • • w ^iiBSKs? CAD APPLICATIONS c '^tHdIiO' I OF ARTIFICIAL INTELLIGENCE oS INTEGRATLI) DAIALASI: iViA \Ai:a!^ii:M

~~(/> (D o O3 •o 01 m Application oriented 3~~

~~to ' i K (D I .^ Constraint nianagement -1 13 W~~

~~w >*• Application based data links. a> 3~~

~~o >~~

~~(O o o o~~

~~(O ~j u a r-^ a> c «(D~~

~~•oa~~

(/I acKaMMa**

~~mSas•laa^' ? CAD APPLICATIONS I OF ARTIFICIAL INTELLIGENCE O > O (D o COiNTINCENCV iMANACtliViLM o •ao 3~~

~~CO O 4> Consistency checking NJ~~

~~# Application specific~~

~~^ " Automatic reconciliation o > ID~~

~~CD -vt CJ I 3 . - • g 4a* • ^ 4«aiB>«' S «li««Ba«- ^ •<*««•«•*» a > HiiEsIs; CAD APPLICATIONS cr (A a 5' OF ARTIFICIAL INTELLIGENCE~~

~~> o ON-LINE HA!\DBOOKS~~

~~2. (A (D 3 o # Engineering handbook data o 3 •a lu 3~~

~~CO I o # Application relative OJtO S-J a~~

~~w 3 # Ease of use~~

~~o > (O Ol~~

~~.b. O 00~~

~~(O O O O~~

~~CO u o w B n c (D cn •o O «• • • S iSSt*. • g- JMiBMi a > ' Q. lii&MiiJtM K • > Uatiiia 11II « • CO a. "mBp-'••••' CAD APPLICATIONS BJ' >< O OF ARTIFICIAL INTELLIGENCE > O~~

~~o CO COVERNiVILNTAL CODE CHECKING (D o o 3 •D lU 3 <^ Design verification~~

~~CO I o 3 to Q.' CO <# Design reconunendations Q. <5 I T) 0) # Multiple regulations~~

~~CO 0)~~

~~o u> # Safety considerations CD O>~~

~~o 00~~

~~oCO o o~~

CO CO o » ST J3 c (D (0 O O •3 O B> mm ' ' ffi ..Waia* • P- . !{»>«•> >- cn UiiiU«ii«i<« c CT W o! '«g::!j? CAD APPLICATIONS o>' -«aatI*' >2 o > OF ARTIFICIAL INTELLIGENCE O I CONSTKAlNT LAi\(;tAGrS o o 3 •o 01

~~^ Engineering aid 5 10 Q! ^ > CD TJ S) O Equation solving w (U c_ O (fl ffl >O~~

~~.c 3o CD ~J oCO o~~

CO -«i u if to U CJ u ^ "c "^ E z •> .•''J 0fi C/3 bJ r .1 u o J> > Z c J c o « fliS o hJ ^ E- 7L c < Z O « E u & 0 a ZJ la 1 "flJ < Q. > ^ "O •To3 a. CL C< U o HMM 0 i M U CC H E & < es < 0 2:: k b. LJ o a. - . »^ H # ^ • * • aHi O •^ • * laaaiI ^^^ • < a iifitm •*•* • • • aaan-sia • • **ummimi • • «9«<«Ma' r •^US^HUiii^

~~- 30 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 as~~

~~O ••: e H , .; 10 os! • 2^ £ CU ^ ' I 'S < 0 E o < X »• • 07 o~~

~~bJ W.. en > c Q C0 • fli s 4irf •"'-. « en < 4i« cu en ^^" U • tmMi • •• «> O~~

~~a • av • •• • ««l • * 'ssitt • a Ha Tfnm • < • m aattmUM • • ••••«i&Mr '-••asissaa' '•aasaajHT~~

~~- 31 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o 0)~~

Q. •* ii iSi:• « " •• - ' l..A.'c a* i» » i • > t;::iV V t lA ^ m - - • 1 .:. (lb t< » «l HI ' WCT ^tilfc'KjiittK (». « b (D J /-> r r-v i'si i.:, ,:• J', i" " t' i i i '•<.* i\ l' 3 ^ 1982 lOK - lOOK LIPS 8 ' a NJ Q. (D I T3 0) 5- & 1992 lOOM - lOOOM LIPS O

~~C_ o U> _(D O> (O Ol o.b. CO~~

~~(O 8 o~~

~~u I3 O o •S o ••• • • 11) <•••• • «tlllM|B« • i•••••• > • • • CO c CT W 5: '*«!«£» CAD - AI IN THE FUTURE 0>' •5 o^ > FUTURE CAD SCENARIO o~~

o o m Design by specification 3 •D B) ^ Progressive clarification of re^uiretnents OJ CO o Ul # User design of aesthetics I # Machine generated implementations i # Automatic constraint reconcihation a w a # Extremely high interactivity C- O a> ca O >

~~o 00~~

~~<£> --4 W THE UTILIZATION OF ARTIFICIAL INTELLIGENCE~~

~~IN DEVELOPING USER INTEKFACE5~~

~~ANIi NEXT GENERATION WORKSTATIONS~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Part< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 COMPUTER EVOLUTION~~

~~A HAINFRAHES~~

~~A MINICOMPUTERS~~

~~A WORKSTATIONS~~

The driving force in Artificial Inteili-3ence has been the availability of sufficient computing power to allow the iwplenientation of interesting applications at a viable cost. Artificial Intelligence applications inherently require vast amounts of computational capability and data storage. The evolution from mainframes to minicomputers to workstations has brought the cost of computing per user into a range where economic application of artific:-3i intelligence techniques is feasible.

~~- 2 " Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 E'JOLUTION OF PROCESSING~~

~~A HECHANICAL CALCULATOKS~~

* NUMERICAL PROCESSING

~~A 3YHP0LIC PEOCES&ll^G~~

The very first computers were nathino «ore than mechanical calculators. As the first computers evolveci they were in the main directed toward repiacino the mechanical calculator. Therefore, computers were first proqraramed so that the internal states represented numbers. Numerical processinq has been the basis of most business and scientific processing to date. However, there is no fundamental reason why the internal states of a compuie' need represent only numbers the internal states of a comouter cculd just a= easily be representative of any symbols that we desirf-. It is this e;;tensicri of the computer to the more generalised form of symbolic processing which is the fundamental difference between traditional computing and artificial intelligence.

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ASPECTS OF ARTIFICIAL INTELLIQEMCE~~

* PROBLEh SOLVIM:-

~~A LOGICAL rEASD.^ING * LANGUAGE U^PEPEI ANTING~~

*. PROGP.AHMING

* LEAPNING

~~k EXPERTISE~~

* RGPOTICS AND 'J IS ION.

k SYSTEMS AND LANGUAGE^ In this talk I will not even try tc tackle the probleiK of enactly what artifici-sl intelligence is, since that is the subject of considerable dec-ate and is best left to those who concern themselves with such things. It is cfte^ said that if it works, it isn't artificial intelligence anyir;cre, Chesi prograiBs and spelling checkers were once the subject of much AI attention today they are so commonplace that they are not general ]_v considered tc be AI. Instead I will address the areas in which work is being done by AI researchers to give an overview of the breadth of wnat AI is and tc give some food for thought as to how it wight be applied to the CAD prociem.

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 PROBLEH SOLVING~~

* SEARCH TECHNIQUES

~~A EXHAUSTIVE SEAFCHES~~

~~f HEURIEIIC SEAFCHES~~

~~f PACKQAnhON, CHESS~~

The area that most people associate with AI is the Qeneralised area of probieiii solving, particularly as it reiatei- tc games. In this class of problem you are ge'":erally evaluatirq a vast number of potential soiutionE tc a problem tc find the best one. Traditional programs approach this problem by attempting tc evaluate every possible potential solution the Ec-caile: exhaustive search technique. While this technique is finf fcr sifficie problems, it fails apart wnen addressing problems like a best cness wove, where there are literally billions of potential soiutiori to evaluate each time.

To rescive this problem, AI introduces the idea of a heuristic search. A heuristic search is done on the basis of some rules of thumb to narrow down drastically tne number of potential solutions which must be evaluated. This is in fact exactly what a human being does when he considers a chess move. h\ narrowing down the number of solutions that iiust be evaluated, the program may net make the best possible move, but depending on the heuristics (rules of thumb) employed, it can make a good, in fact usually a very good, nave. The process of interviewing e';perts in a given field to determine a ?ood set of heuristics for a given problem is referred to as knowledge engineering.

~~- o •- Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 LOGICAL REASONING~~

~~A THEOREH PROOFS~~

* DATA REDUCTION

~~)k DATABAS: QUEft'r~~

The ne>tt area of interest to AI researchers is the area of loQical reasioninq, These programs deal with a set of data (or facts) and attempt to deduce fron; the facts, lOQical conclusions. One aspect of this is the proof of variOLH niathematical proofs, given the facts associated with iriathematics. Another more practical application is in the reduction cf large amounts of data giver: facts about the logical relationships between the data. In an even more practical sense, logical reasoning is applicable to the day-tc-day problem of database queries.

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 951317 (408) 971-9000 / Telex 171973 LANGUAGE UNDERSTANDING~~

~~A NATURAL LANGUAGE INTERFACES~~

* SPEECH UNDERSTANDING

~~f LANGUAGE TPANSLATIOIv~~

One area w^ich has received much publicity lately is the whole area of lariQuage understanding. The natural lanQuage interf-3ce is the most widely knowr; application of languaoe understanding, although speech understanding and language translation ewploy the same principles. An English natural language- interface is in fact a language translator from English to a machine cowniand language. One of the first applications of natural language interfaces was to the problem of database query to assist managers in accessing the often tonipiev 'JBia structures contained in many crmtdon business applications.

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 PROGSAMHING~~

~~A AUTOMATIC PROGRAMhING~~

~~^ LEARh^ING PROGRAMS~~

Automatic pro3ranin;iri9 is the dreani of wany, but to AI researchers it ii a solvable problem being investigated at many research laboratories across tre country. Systems exist today which ask relative simple questions about a program specification and generate optiiaised. error-free code for a narrow range of e>;pertise. The nain advantage of these systems come in the area of program saintenance, where small change: to tne specification can be ma-^e to fi>; ihe elusive bug, and a new set of error-free code can be automatical!; generated.

Another related area is programs which beccm? smarter as they run. Ihe=e programs change themselves over time in response to the user and its application. For example, a program say learn that you often mis-type a particular command in the same way, and adiust its syntax to cospensate for vour errors.

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 LEARNING~~

* EXAMPLE

~~A PERFORMANCE •.~~

~~k INSTRULTIIN~~

This leads to the area of learniriQ systems in general. Unfortunately, this has been one of the least productive areas of AI research. Basic-ally, learninq systems learn by example, by actually performing a task and observing the result, or by the user instructing the rrogran- as to how to perform a particular task.

~~_ 9 _ Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 EXPERTISE~~

* EXPEFT SYSTEH5

~~A KNOyiEDGE DATABASES~~

Clearly the most touted ares of artifici-al intelliQence is the area of expert systems. In simple terms an expert systes; uses rules (heuristics) in conjunction with a data set in order to reach conclusions about a particuisr problem. The rules are generally derived by knowledge engineers working with established experts in a psrticular area to develop a set of heuristics that possess much of the reasoning power of the expert himself. While the systeni-s generally cannot be any better than the expert himself, its at-ility to procesi larger amounts of d-ata with iiuch higher consistency often produces results that are, practically speal-ing, beyord the capability of the expert upor' which it la &ased.

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ROBOTICS AND VISIOhJ~~

~~A INIiUtUIAL ROBOTS~~

* OBJECT RECOGNITIOf^

~~A SUF^'EILLANCE~~

In terms of practical applications, robotics represent the largest market '.••; date for applied AI systen-=. While the robot inarkct in the US has beer, relatively slow to devSiCp, the Japanese have been far more agressive. '^s a resijlt many coaipaniet in tne US have active prograos in place tc, close th? robot Qac.

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 SYSTEMS AND LANGUAGES~~

* TOOLS AND IDEAS

~~- TIME-SHA.RING - LIST PPOCESSING - INTEi~~

~~k SPECIALIZED PPOGPArthING LANGUAGES~~

~~- LISP - PP'JLOG - PLANNER~~

One are.2 that AI people point with pride is their contrib-jtion to tr;? deveiopment of tools and ideas. Time-sharing, list processinq, and interactive debijQqing are all concepts originally conceived by groups of AI researchers. There has also been considerable work on specialized prograrnining languages which more appropriately address the needs of the AI programmer, and which are beginning to gain acceptance outside the immediate AI community.

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ARTIFICIAL INTELLIGENCE SYSTEHS TODAY~~

~~A INTERNIST/CAIiUCEUS - University of Pittsburgh~~

~~A PROSPECTOIt - SftI International~~

~~A IiIPhETEP ADv'ISOP - Sclunsberser~~

~~A DRILLING AD'^ISQP - leknowledqe/Elf-Aquitaine~~

Often we refer to AI as if it were somethinq in the future. Actually it here today, and in daily use in a number of very practical applications. Some of the Most noteworthy are indicated on this slide.

~~- 13 - Dataquest Incorporated, A Subsidiary of A.G. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 INTERNISI/CABUCEUS~~

~~A 807. OF ALL INTERNAL MEDICINE~~

~~A 500 DISEASES~~

~~A 3500 HANIFESTATIONS OF DISEASES~~

~~A PPEPAFING TO UNDERGO FORMAL CLINICAL TRIALS~~

The Internist system was developed at the University of Pittsbursh as an e;;per" system in the area of intern.3l medicine. It already has incorporated 807. of the available knowledge about internal mecicine. It knows about 500 diseases and 3500 manifestations of these diseases. It is preparin? to undergo foriiil clinical trials. The future is now. This system routinely performs analyse; as well as »ost internists, and better than doctors from other sperialities.

~~- 14 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 OTHER EXPERT SYSTEMS~~

~~A PEOSPECTOP - GEOLOGICAL EXPERT SYSTEM~~

~~A DIPHETER ADVISOR - OIL WELL LOG ANALYSIS~~

* IiRILLING ADVISOP - DRILLING PROBLEM DIAGNOSIS

The Prospector system developed at SRI is an expert in geology and mineralogy. It has found mineral deposits valued in the hundreds of millions of dollars. In one case it found a deposit under a wining company waste site.

The Dipmeter advisor developed by Schlurwerger does oil well log analysit. Schlumberger each year issues credits of about $40M due to errors made in well log analysis. The automation of this tedious process save SchluiBber^ET several million dollars each year.

The Drilling Advisor, developed by Teknowledge for the French oil company. Elf A'quitine analyses drilling problems and provides recommended courses of action. Considering that the operating cost of an oil rig typically costs over $100,000 per day, and the abandonment of an oil well due to probleiIis typically costs several million dollars, it is easy to see why Elf A'quitine expects the system to return its full development cost on its first successfiji application.

~~The point here is that expert systems with impressive returns on investment are already in use. If you are wondering when you should begin to investigate this technology, the time is now.~~

~~A representative of Elf A'quitine once commented, '(get quote from 'Fifth Generation about Qold nuq^ets)~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 PERQ/LINQ SYSTEM DESCRIPTION~~

~~(INSERT OVEPyiEU OF PERG/ACCEKI hULTIPLE ENVIRONMENT SYSTEh AT THIS POINT;~~

~~- 16 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD APPLICATIONS OF ARTIFICIAL INTELLIGENCE~~

~~NATURAL LANGUAGE INTERFACES~~

~~^ HELP SYSIEHS~~

~~jt DATABASE QUERIES~~

~~^ OPERATING SYETEn INTERFACES~~

~~Now I would like to discuss some possible application areas for AI in CAD systems of various typet.~~

CAD systems by their very nature are geometric, not textual. Users generally point at things as opposed to typing te:;t. As a result, natural language interfaces are not as applicable to CAD systems as they are to some of the Bore traditional applications. They are extremely useful, however, in the implementation of help systems, and as a tool for doing queries en databases. They are also valuable when applied to the operating system interface, since many users of application prograois are not generally familiar with the synta- of operating systeri; commands. A natural language interface frees the user from having to remember whether the first or second file in an append command of the operating system is the appendee or appendor.

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD AFPLICAIIONS OF ARTIFICIAL INTELLIGENCE~~

~~SPECIFICATION URITIMG~~

~~A EXPERT SY5TEH~~

* PRODUCTIVITY AID

* QUALITY IHJFOVEMENT

The entire area of specification writing is the focus of some interest anv:!r.9 CAD companies. Many users are as concerned about the specifications as thes' are about the ultimate iniplenientations. The profitability in many industries, such as -srchitecture and construction, are as determined as much t\ tht specification, proposal and pricing as they are by the actual implementation.

This area lends itself well to the concept of expert systems. Specifications can be more thorough and consistent, and better advantage can be nade of past experience by incorporating it into the Knowledge base.

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen CoITipany / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD APPLICATIONS OF ARTIFICIAL INTELLIGENCE~~

~~INTEGRATED DATABASE MANAGEMENT~~

~~>t APPLICATION OFIENTED~~

~~A CONSTPAINI HANAGEhENT~~

* APPLICATION EA5ED DATA LINKE

The control of a database in a CAD system is in B-any ways far more sophisticated than traditional databases. For one thing, the data tends to be interlinked in far tiore complicated ways. For e:;anipie. an enqinecr chanqinq one snsall part in a system may effect any nuniter of other tartt in obscure ways. It could affect the tensile strength cf a bridge bean;, cr the systems perforieance on a particular FCC or UL test, or it could affect circuit loadins so that other parts would have to be changed.

Databases in engineering applications have literally hundreds of constraints between various items in the database. These application-based data links all must be addressed each time any part in the system is changed. The con>ple;:ity of this task is enormous, as atanv of us well know.

~~- 1 9 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD APPLICATION'S OF APTIFICIAL INTELLIGENCE~~

~~CONTINGENCV hANAGEHENT~~

~~A CONSISTENCY CHECKING~~

~~k APPLICATION SPECIFIC~~

~~k AUTOHATIC PECONCILIATIOK~~

This entire area of contin9ency wanaQement is one of the most lucrative areas to which to apply AI techniques. For example, when the design of a Dipe systen; in a building is coBipleted, an AI program could check all the pipe and joint size; ancf types to insure th.5t they are consistent with each other and with the specification of the liquid or gas flow in the pipe. If inconsistencies were noted, they could be identified to the designer who could then request the .prograw to automatically resolve the differences, adjusting pipe and joint sises and strengths to be consistent with the overall specifications.

~~Timings in logic circuits could be handled in a similar wanner.~~

~~- 20 " Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAIt APPLICATIONS OF AP.TIFICIAL INTELLIGENCE~~

~~ON-LINE HANHBOOKE~~

~~A ENGINEEPING HANrBOOK DATA~~

~~k APPLICATION RELATIVE~~

~~9r CONTEXT OFIENTED~~

~~k EASE OF USE~~

One of AI s contribution to the word processing world was the EDeliir? checker. In such the same vein the opportunity exists to bring the engineering handbooks on line to the designer in the saB'e way that tht dictionary was brought on-line to the writer. By makinq this reference itaterial relative to the application and responsive to the conte;;t m which the designer is working will provide an easy to use aid that could speed up ti^e design process for nany of the Bore tedious checks which must be routinpiy nade by designers.

-• i.:. — Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD APPLICATIONS OF ARTIFICIAL INTELLIGENCE

~~GOyERNiiENTAL CODE CHECKING~~

~~k DESIGN VERIFICATION~~

~~k DESIGN RECOHHENDrtllONS~~

~~k HULTIPLE REGULATIONS~~

~~k SAFETY CONSIDERATIONS~~

One of tht more interestin5 and lucrative areas for the applicatior cf eitperi systems is in the area qovernwental codes and requlations. The incredible number of Qovernmental reqijlations which apply to even the siaplest piece of hardware staggers the imagination. In this era of aultinational trade, virtually every engineer nust daily deal with not only the federal, state, ari: city regulations, but also the requlations of England, Gernany, Japan.... Add on to this, the whole subject of safety and product liability.... It is difficult to imagine that the present situation can continue without some e:-;pert assistance.

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD APPLICATIONS OF ARTIFICIAL INTELLIGENCE~~

~~CONSTRAINT LANGUAGES~~

~~A ENGIHEEFINfj All'~~

~~A EQUATION SOLVING~~

One of the more specialised areai of AI is the concept of constraint languagei Constraint languages allow a user to supply any of the variables to a set of equations and solve the missing variables. This relatively siispie aid greatly simplifies the use of equations in marIy traditional engineering applications.

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD APFLICATIO^iS OF APTIFICIAL INTELLIGENCE~~

~~INTERACTIVE APPLICATION DEVELOPHEKT~~

~~A OBJECT ORIENTED PROCFAMnlNG~~

~~A PROijRAMhING DEVELOPnENT EN\'IRONMENTS~~

One of the wajor contributions of the AI coniniunity over the years has been their contributions to the idea of interactive proQramrring. The developmer;-; environments now in daily use by knowled9e engineers provide an interactive development environment generally considered to be superior to any of tht traditional development environments. In the past these techniques were corI^idered too large and cuiBbersoree to be of practical interest, byt as tn'? compute power per user has gone continually upward with the popularity of workstations, the Lisp development environment is receiving increasing attention and justifiably so.

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD APPLICATION EXAMPLE~~

~~EOR«DRAW DRAFTING SYSTEM~~

*; EXPERT SYSTEM

~~A DESCRIPTIVE GEOMETRY~~

~~A ARCHITECTURE~~

~~k PROCESS AND INSTRUMENTATION DESIGN~~

An e;:3ii!ple of expert system techniques applied to the user interface of 2 computer aided design package is offered by the Formdraw package marKeted by Formtek cf Pittsburgh. Data entry is provided in the traditional manripr by what appears on the surface to be a sinple 2D-drafting system. In fact, the system is an expert system on descriptive geometry. That is to say that it understands the various principles of descriptive geoiuetry. For example, it understands what an intersection is, and if you tell it to make two line;; intersect, it will do so. If you tell it to make to parallel lines intersect it will advise you that parallel lines cannot intersect.

In addition. Formdraw can also be an expert system on architecture in their architecture system. So that it understands the concept of 'walls', and understands the unique qualities of the intersection of walls. It also understands that 'windows' go in walls. It also understands that walls do not intersect at windows, so if you asked the system to 'intersect' two walls, and the intersection occurred at a window, the system would advise you that walls do not nor«ally intersect at a window. If you attempted to delete a wall, it would Know that those things in the walls, such as windcws, must also be deleted.

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD - AI IN THE FUTURE~~

~~JAPANESE FIFTH GENERATION~~

~~k 1982 — — 10 - lOOK LIPS~~

* 1992 —— 100 - lOOOh LIPS

I could not give a taU on AI without mentioning the Japanese Fifth Generation. The Japanese have announced a plan, referred to as the 'Fifth Generation', to develop computer systems by 1992 capable of fron; 100 million to 1 billion LIPS (Logical Inferences Per Second). This represents an increase of 10,000 times, four orders of magnitude, from the present technology of less than lOOK LIPS. Imagine, if you will, a system capable of 10,000 times the reasoning power of our present systems. This means that a problen- which takes three hours to solve today could be done in one second in 1992 or a years worth of computing could be accomplished in one half a day.

The Japanese feel that if they could develop this technology ahead of the rest of the world, they could establish superiority, not only in this field, but ir: many other fields which have as their basis engineering solutions to problems. They believe that these KIPS (Knowledge Information Processings Systems) would provide them with superior design and manufacturing capability which would manifest itself in superior, lower cost, and higher quality products.

~~- 26 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD - AI IN THE FUTURE~~

~~FUTURE CAD SCENARIO~~

~~A DESIGN BY SPECIFICATION~~

~~A PROGRESSIVE CLARIFICATION OF REQUIREMENTS~~

~~A USER DESIGN OF AESTHETICS~~

* MACHINE GENERATED IHPLErtENTATIONS

* AUTOMATIC CONSTRAINT RECONCILIATION

* EXTREMELY HIGH INTERACTIVITY

~~Consider if you will what desi3n nay be like with the availability of thi; incredible conputins capability.~~

Desiqn would be acre of a process of refining a specification, then the worl<.inq out of a detailed inpleaentation. Instead of workins on details, a desisner would respond to a series of inquiries about the requirements of a design. For exasple, the designer of a hospital would answer questions as to the nuaber of beds, cliaate, would there be a aaternity ward, etc.

The user would specify factors of aesthetics, such as the nuiiber of flocrs — which would be autonatically laid out by the system. The systea ai^ht respond and tell you that the aasiiaua nuaber of floors allowed by the adjacency requirements and the underlying ground conditions was four.

~~If the user specified a aore efficient window design, the systea would automatically adjust the sise of the heating and air conditioning plant.~~

In circuit design, a siailar process would tal'e place. The CAE systea would ask what kind of circuit you wanted to design... a controller perhaps; What Kind of device? A disk. And on and on, until the controller was fully specified. Then the designer would decide, with the aid of the systea, whether the controller should be a printed circuit board, a gate array, or a custon VLSI chip. The decision aade, the systea would iapleaent the detaile: design.

~~Is this far fetched? Well I broke ay teeth on a '???•. It seeas entirely realistic to ae.~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 (IDaDataquest Dataquest~~

~~COMMUNICATIONS INTEGRATION~~

~~Thomas H. Bredt Vice President Director, Telecommunications Industry Service Dataquest Incorporated~~

Dr. Bredt is a Vice President of DATAQtJEST and Director of the Telecommunications Industry Service. His major areas of responsibility include system-to-system communication, terminal-to-system connections, local area networks, public data networks, and protocol standards. Before joining DATAQUEST, Dr. Bredt was at Hewlett-Packard for eight years where he participated in the development of computer systems, manufacturing applications, and data communication products. His most recent position was Engineering Manager for HP's Information Networks Division, and manager of the company-wide Data Communications Strategy. Prior to his experience at HP, Dr. Bredt worked at Bell Laboratories and was a member of the Electrical Engineering faculty at Stanford University for seven years. Dr. Bredt's undergraduate degree is in Science Engineering from the University of Michigan. He has a Master's Degree in Electrical Engineering frcm New York University, and a Ph.D. in Computer Science from Stanford.

~~Dataquest Incorporated WORKSTATION FOCUS CONFERENCE August 13 and 14, 1984 Burlington, Massachusetts~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CbMMUNlGAtldKS INTECSRATiON~~

~~THOMAS H. BREDT VICE PRESIDENT DIRECTOR TELECOMMUNICATIONS INDUSTRY SERVICE~~

~~- 1 - ® 1984 Dataquest Incorporated-Reproduction Prohibited~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ESTIMATED U.S. REVENUE FOR VOICE/DATA WORKSTATIONS (IF-SOLD, END-USER) (MLUONS OF DOLLARS) $1,628 CAGR 1984-1988 60.72 PC CAGR 117.52

~~SYNCHRONOUS CAGR 149.4)i~~

~~ASYNCHRONOUS CAGR 30.4^ 1983 1984 1988~~

~~SOL«CE C-rT^C'i.'E.'T~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ESTIMATED U.S. SHIPMENTS OF VOICE/DATA WORKSTATIONS (THOUSANDS OF UNITS)

~~CAGR 1984-1988 1,070 83.4% y 385 PC CAGR 195.2^~~

~~353 SYNCHRONOUS 94.5 / CAGR 181.8?> 30.0 «n -1 5 - _5-6— 332 ASYNCHRONOUS 1.0 63.9 CAGR 41.0^ 57 5 1983 1984 1988~~

~~JOiJ^CE WTWIC-T~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 SB9aaBmmimammmiaamBmeeKmmmmmimmmmatawmmmKmKswgBsaaKammmmarmeaamim:aieassaK» VOICE/DATA WORKSTATION

~~1988 AVERAGE SELLING PRICE~~

~~ASYhJCHRONOUS S 895~~

~~SYNCHRONOUS S 1.425~~

~~PERSONAL COMPUTER $2,150~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 I»n VOICE/DATA WORKSTATION MARKET TRENDS oaKM~~

• EARLY TERMINAL PRODUCTS WILL LEAD TO PRODUCT FAMILIES • PERSONAL COMPUTER WILL HAVE 36 PERCENT OF UNIT SHIPMENTS BY 1988 • INTEGRATED TELEPHONES WITH SOFTWARE SUPPORT FOR CALENDAR AND TIME MANAGEMENT WILL BE OPTION FOR EVERY TERMINAL AND PERSONAL COMPUTER PRODUCT BY 1988 • NEW MARKET ENTRANTS AND TERMINAL VENDORS WILL COMPETE FOR POSITION AS INDEPENDENT SUPPLIERS OF VOICE/DATA WORKSTATIONS • SHAKEOUT WILL OCCUR AS VENDORS SEEK CORRECT BLEND OF PRODUCT DESIGN. PRICE. AND DISTRIBUTION STRATEGIES • PROPRIETARY DESIGNS FROM PBX MANUFACTURERS WILL COMPETE WITH NONPROPRIETARY DESIGNS FROM WORKSTATION AND COMPUTER MANUFACTURERS

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ESTIMATED US. SHIPMENTS OF VOICE/DATA WORKSTATIONS~~

~~(THOUSANDS OF UNITS} -~~

~~MK COMPANY 1883 MARKET SHARE~~

~~1 NORTHERN TELECOM 15.0 50.02 2 TYMSHARE 9.9 33.02 3 GTE 2.4 8.02 4 MTTEL 15 5,02 5 DAVOX 1.0 3.32 OTHERS 2 .72~~

~~TOTAL 30.0 100.02~~

~~- 6 - ® 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~RANK COMPANY 1983 (SM} MARKET SHARE~~

~~1 NORT^CRN TELECOM $195 47.7Z 2 TYMSHARE e.9 21.82 3 MfTEL 6.0 14.72 4 GTE 3.6 6.82 5 DAVOX 2.3 5.62 OTHERS JB 1.42~~

~~TOTAL S40.9 100.02~~

~~SOURCE C*T(»OuE.^~~

~~- 7 - ® 1984 Dataquest Incorporated-Reproduction Prohibited~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MKOaKH BtSBOHBC Ese smaecnmRsitsRer CABLE-BASED LOCAL AREA NETWORKS

~~APPLICATION AREAS VEhDORS~~

~~COMPUTER TO COMPUTER FETWORK SYSTEMS~~

~~TERMINAL TO COMPUTER BRDGE COMMUNICATIONS. WTERLAN. SYTEK. UNGERMANN-BASS~~

~~PERSONAL COMPUTER 3COM. CORVUS. NESTAR RESOURCE SHARING~~

~~OFFICE WORKSTATION NETWORKS DATAPOINT. WANG, XEROX~~

~~FACTORY NETWORKS CONCORD DATA SYSTEMS~~

~~SPECIAL NETWORKS CE.G- CAD) APOLLO~~

~~- 8 - ® 1984 Dataques-t Incorporated-'Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 «H MARKETING CHANNELS FOR LOCAL AREA NETWORKS~~

~~LAN VENDOR~~

~~SYSTEM WTEGRATOi R JEll (AODSHW AND/OR SW VALUE]~~

~~END-USER~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ESTIMATED U.S. MARKET FOR LOCAL AREA NETWORKS (MLUONS OF DOLLARS)

~~CAQR 81,031 46.5X /~~

~~• / • $224 / S144 ^~~

~~1983 1984 1988~~

~~SOURCE WTA<:IUE«:T~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ESTIMATED LOCAL AREA NETWORK AVERAGE RS-232 -C CONNECTION COST {HARDWARE EQUIPMENT ONLY. US. DOiXARS) $525

~~S445 CAGR 1984-1988 -14.82;~~

~~S235~~

~~1983 1384 1988~~

~~StlftCE WTAi-niEJT~~

~~- 11 - ® 1984 Dataquest Incorporated-Reproduction Prohibited~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 BssaoMBS^iBBEtBcainEBRUMjiBBJU^BBnaaBnaaajBiaEniRsawnnnnMH LOCAL AREA NETWORK MARKET TRENDS

• STEADY PROGRESS ON STANDARDS FEEDING STRONG GROWTH • IEEE 802.3 AND IEEE 802.5 WILL BE DOMINANT OFFICE NETWORK STANDARDS • BROADBAND NETWORKS WILL BE USED IN FACTORIES AND AS BACKBONE INFORMATION HIGHWAYS IN LARGE BUILDINGS • PERSONAL COMPUTER MANUFACTURERS MOVING QUICKLY TO INTEGRATE RESOURCE SHARING NETWORKS INTO PRODUCT LINES • VLSI CHIP DELAYS SLOW REALIZATION OF MARKET POTENTIAL • DATA BANDWIDTH FOR COMPUTER-COMPUTER FILE TRANSFER. IMAGE PROCESSING. AND GROUP IV FACSIMILE WILL PROVIDE WINDOW OF OPPORTUNITY • MAkE-VERSUS-BUY IS A BOARDROOM DECISION FOR OEMS. SYSTEM INTEGRATORS. AND COMPUTER MANUFACTURERS

~~- 12 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company 11290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 WOtHPOt ESTIMATED U.S. REVENUE FOR LOCAL AREA NETWORKS~~

~~(HARDWARE EQUIPMENT ONLY)~~

~~^NK COMPANY 1963 ($M} MARKET SHARE~~

~~1 NETWORK SYSTEMS $37.0 25.75? 2 UNGERMANN-BASS 17.5 12.2% 3 SYTEK 16.9 11.7X 4 DATAPOINT 15.1 10.5^ 5 3COM 9.3 6.55?. 6 CORVUS 6.9 S2Z 7 MTERLAN 6.7 Alt OTHERS 32.6 22.5?~~

~~TOTAL SI 44.0 100.0%~~

~~SOI.TCE C-ftTiV>JE:T~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 DATA PBX DEFINITIONS~~

• A DKSTAL PBX THAT ALLOWS TERMINALS TO SWITCH AND CONTEND FOR COMPUTER RESOURCES • SWrrCHMG ENABLES TERMff^AL USER TO ACCESS DIFFERENT APPLICATIONS PROGRAMS ON DIFFERENT COMPUTERS • CONTENTION ALLOWS MORE TERMINALS TO ATTEMPT TO ACCESS FEWER COMPUTER PORTS • DATA SWrrCHING/CONTENTION ONLY; NO VOICE swrrcHWG AS WITH VOICE/DATA PBX

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ESTIMATED U,S. REVENUE FOR DATA PBX MARKET (IF-SOLD, END-USER) (MLUONS OF DOLLARS) S422 CAGR / 1384-1988/ 37.2Z /

~~S119 /~~

~~S77 /~~

~~1983 1984 1988~~

~~SOtJUCE MTWii.tn~~

~~- 15 - ® 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ESTIMATED U.S. SYSTEM SHIPMENTS FOR DATA PBX MARKET~~

~~4.850~~

~~CAGR / 1984-1988/ 33.U~~

~~1.575 1.025 y'\~~

~~1983 1984 1988~~

~~SOURCE CwTAOLitn~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ESTIMATED U.S. LWE SHIPMENTS FOR DATA PBX MARKET (THOUSANDS OF UNES) 2.300 CAGR A 1884-1988 / 33.2 Z /

~~730 /~~

~~400 /~~

~~1983 1984 1988~~

~~SOtmCE DHTnCCJEfT~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 BCCB^Bi^nvBBSBBa^wHnKsnaaIeHnwHBiKaaBHiRaaHiBaHaBi DATA PBX TRENDS

tuaj-±ju..iju,.>i w!iiIiwIji.'inw»^w»weBaBa^aM—M•^•wcaMfir.uu.i]j«jj»AAm..iMU'.'gig'ig • DISTRBUTED ARCHfTECTURES TO NSTALL NODES WhCRE TERMINALS ARE LOCATED • T1 (1.544 MBPS) LINKS BETWEEN NODES FOR HGH-SPEED TRANSMISSION • USE OF VOICE/DATA MULTIPLEXERS FOR SIMULTANEOUS VaCE AND DATA TRANSMISSION OVER EXISTING N-PLANT TELEPHONE WIRING • PRICE-COMPETmVE ALTERNATIVE FOR LOCAL AREA I^TWORKING • MARKET GROWTH OVER THE PERIOD RESULTING FROM INCREASING MINICOMPLI'ER AND ASYNCHRONOUS TERMINAL POPULATIONS

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ESTIMATED U^. REVENUE FOR THE DATA PBX MARKET (IF-SOLD, END-USER)

~~RANK COMPANY 1983 ($M) MARKET SHARE~~

~~1 MK:OM $33.0 42.9^ 2 DEVELCON 15.0 19.5% 3 GANDALF 14.0 18.2% 4 WFOTRON 6.6 8.6 r 5 CODEX (MOTOROLA) 4.0 12% 6 M/A COM LNKABU 3.0 3.9% ALL OTHERS 1.4 1.7Z~~

~~TOTAL $77.0 100.0%~~

~~SOt«CE t~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 »nR9nHBSMH9B7 VOICE/DATA PBX MARKET iVMiva~~

ESSENTIALLY NONBLOCKING: • 100 PERCENT OF AVAILABLE DATA PORTS AT 19.2 KBPS ASYNCHRONOUS • 100 PERCENT OF AVAILABLE DATA PORTS AT 56 KBPS SYNCHRONOUS • 10 CCS VOICE TRAFFIC TO EACH VOICE STATION

~~• 30 CCS VOICE TRAFFIC FOR EACH ANALOG TRUNK~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ToammmKm^tKTmimîmmâ^m^^mâaimmmmBamKmamaKtaaaKai PRIVATE BRANCH EXCHANGES

• VIRTUALLY ALL OF THE PBX SYSTEMS INTRODUCED IN THE LAST 18 MONTHS ARE ESSENTIALLY NONBLOCKING • DATAQUEST PREDICTS THAT BY 1988, 85 PERCENT OF ALL PBX SYSTEMS SOLD IN THE UNITED STATES WILL BE DIGITAL. INTEGRATED VOICE AND DATA SYSTEMS OF ESSENTIALLY NONBLOCKING ARCHITECTURE (BASED ON DOLLAR VOLUME}

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ^n^bannt A NEW PBX MARKET SEGMENT~~

THE ULTRALOW-COST PBX • ANALOG VOICE ONLY • FIXED SIZE--NOT EXPANDABLE ABOVE 40 LINES • SUPPORTS 2500-TYPE TELEPHONE SETS ONLY • SIMPLE FEATURES--NO BELLS AND WHISTLES • IF-SOLD. END-USER VALUE OF S210 PER LINE (1984) CAGR: 1984-1988 s 131.8Z 1984: S5M 1988: S145M

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ESTIMATED U.S. REVENUE FOR THE PBX MARKET (F-SOLD, END-USER) (MILLIONS OF DOLLARS} CAGR S4.064 1984-1388 ^ 7.8%J>^ S3.009 2.248 >400 LINES S2.620 / CAGR 10.U 1.530 1.366 955 100-400 LINES 840 CAGR 3.32 837 <100 LINES 634 716 CAGR 3.i:r 617 5 145 ULTRALOW-COST CAGR 131.8i? 1983 1984 1988 fa*c£ w^»mx.:

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ESTIMAtED U.S. SYSTEM SHIPMENTS FOR THE PBX MARKET (THOUSANDS OF SYSTEMS)

62.0 >400 LINES CAGR > 5S CAGR 8.82 11984-198 8 / 6.9 100-400 LINES 28.92 / CAGR .AX 23.4 <100 LINES CAGR 5.5X 29.7 97 9 £l.£ ^ *A 1ft ULTRALOW-COST 6.8 48.9 CAGR 122.4^ 6.6 18.9 18.6 ?0 1983 1984 1988

~~SOciRCE D*iT«nLt;T~~

~~- 24 - @ 1984 Dataquest Incorporated-Reproduction Prohibited~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ESTIMATED U.S. LINE SHIPMENTS FOR THE PBX MARKET (THOUSANDS OF UNES) CA6R 1984-1988 ^ 5,439 9.3Z^^^ 3.616 >400 LINES 2,294 CAGR 7.0% 3.567 y 1.748 1,552 100-400 LINES 1.150 :CAGR 1.0% 1.105 <100 LINES 1.094 1.164 CAGR 5.5% 939 ULTRALOW-COST 921 831 CAGR 142 6% 2d 1983 1984 1988 •'

~~SOURCE CWTrtPLtr~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ESTIMATED U.S. PBX PRICE PER LINE (F-SOLD, END-USER)~~

~~• (DOLLARS PER UNE]1 CAGR SEGME^f^ 1983 1984 1988 1984-1988~~

~~> 400 UNES 8880 8875 8980 2.9 ^~~

~~100-400 urcs S765 8760 8830 2.22~~

~~<100 UNES S670 8675 8615 -2.3%~~

~~ULTRALOW-COST N/A 8210 8175 •4.52~~

~~TOTAL MARKET $791 8788 8747 -1.3?~~

~~^- SOW:CE WTrtC«JErr~~

~~- 26 - ® 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen CoInpany /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 PBX INDUSTRY TRENDS~~

• INTENSE PRICE COMPETITION TO GAIN/KEEP MARKET SHARE • AT&T DIVESTITURE WILL AMPLIFY NEED TO FOCUS ON DISTRIBUTION • ABILITY TO MANAGE SOFTWARE DEVELOPMENT WILL BE KEY TO FUTURE SUCCESS • DATA TERMINAL CONNECTIONS WILL BECOME MORE COST-EFFECTIVE WITH AVAILABILITY OF COMPUTER-PBX INTERFACES • PBX MANUFACTURERS WILL EMPHASIZE INTEGRATED DESKTOP WORKSTATIONS TO MOVE INTO OFFICE AUTOMATION MARKET • COMPUTER-COMPUTER COMMUNICATION AND IMAGE PROCESSING WILL LEAD TO NEW PBX ARCHITECTURES THAT MOVE INFORMATION IN PACKET FORM

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 PRIVATE BRANCH EXCHANGES~~

T3 EXPLOrr DIGPTAL SWITCHING, EACH PBX MANUFACTURER HAS INTRODUCED A DESKTOP RS-232 SERIAL DATA COMMUNICATIONS ADAPTOR — NORTHERN TELECOM ADDON DATA MODULE — ROLM DATA TERMINAL INTERFACE — INTECOM DATA OPTION BOARD

~~VARIOUS MODELS OF THESE ADAPTORS CONNECT TO ASYNCHRONOUS DEVICES AT SPEEDS UP TO 19,200 BITS PER SECOND AND SYNCHRONOUS DEVICES AT SPEEDS UP TO 64.000 BITS PER SECOND~~

~~PRICES ARE TYPICALLY $300 FOR ASYNCHRONOUS ADAPTORS AND $800 FOR SYNCHRONOUS ADAPTORS~~

~~- 28 - Q 1984 Dataquest Incorporated-Reproductlon Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 PRIVATE BRANCH EXCHANGES~~

PBX VENDORS NOW VIEW THEIR NEXT OPPORTUNITY IS TO CAPTURE THE DESKTOP OF THE OFRCE WORKER — TELEPHONE — SPEAKERPHONE — NAME/ADDRESSTTELEPHONE NUMBER RLE — PERSONAL CALENDAR — REMINDER SUPS — CLOCK — DISPLAY TERMINAL

~~- 29 - @ 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 PRIVATE BRANCH EXCHANGES~~

~~SOME PBX MANUFACTURERS WILL DEVELOP THEIR OWN DISPLAY TELEPHONE TERMINALS — ROLM CYPRESS — NORTHERN TELECOM DISPLAYPHONE~~

~~SOME PBX MANUFACTURERS WILL SELECT AN OEM PRODUCT — ZAISAN ES.1~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 PRIVATE BRANCH EXCHANGES~~

~~THE NEXT FRONTIER IS EFFICIENT CONNECTION OF MULTIPLE TERMINALS VIA THE PBX TO A HOST COMPUTER~~

~~TODAY, EACH PORT ON A HOST DATA PROCESSING OR OFFICE AUTOMATION COMPUTER MUST HAVE A CORRESPONDING UNE AND ADAPTOR ON THE PBX~~

~~GATEWAYS TO IBM HOSTS HAVE BEEN INTRODUCED BY ROLM AND NORTHERN TELECOM~~

~~- 31 - @ 1984 Dataguest Incorporated-Reproduction Prohibited~~

~~Oataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 PRIVATE BRANCH EXCHANGES~~

~~TERMINAL HOST GATEWAY COMPUTER~~

~~TERMINAL~~

~~- 32 - @ 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 PRIVATE BRANCH EXCHANGES~~

~~TWO METHODS OF CONNECHNG MULTIPLE TERMINALS TO HOSTS VIA PBXs HAVE BEEN PROPOSED~~

COMPUTER-PBX-INTERFACE (CPI) — 1544 MILLION BPTS PER SECOND — 24 FULL DUPLEX CONNECTIONS — EACH CONNECTION O-SSW) BPS — CONTROL IS EMBEDDED IN EACH CONNECTION — PROMOTED BY DIGPTAL EQUIPMENT CORPORATION AND BY NORTHERN TELECOM

~~- 33 - @ 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 PRIVATE BRANCH EXCHANGES~~

DIGFTAL MUUIPLEX INTERFACE PMI) — ANNOUNCED NOVEMBER 2nd — 1544 MILUON BITS PER SECOND — UP TO 155 FULL DUPLEX CONNECHONS — EACH CONNECTION USES ONLY BANDWIDTH REQUIRED — CONTROL IS ALLOCATED TO ONE 64,000 BIT PER SECOND CONNECTION — PROMOTED BY AT8J, HEWLETT-PACKARD. AND BY WANG

~~- 34 - @ 1984 Dataquest Incorporated-Reproduction Prohibited~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ffir«K5S«5SHnS?g??g!gK!KSSaXgSSBRgS5!S!asaS!S5R^ THE BATTLE FOR THE WIRING CLOSET

~~1983 DATA LINES SHIPPED (THOUSANDS)~~

~~DATA PBX 400,000 LOCAL AREA NETWORK 72,500 VOICE/DATA PBX 35,700~~

~~TOTAL 508.200~~

~~Source: DATAQUEST~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 THE VITAL LINK IS COMMUNICATIONS ^letMataatmnmitaamm^~~

~~• MTEGRATED VOICE AND DATA WORKSTATIONS~~

~~• CABLE BASED LOCAL AREA NETWORKS~~

~~• DATA PBXs~~

~~« VOCE AND DATA PBXs~~

~~- 36 - @ 1984 Dataquest Incorporated-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973~~

~~Dataquest~~

~~NETWORK CONFIGURATIONS~~

~~Peter J. Shaw President and Chief Executive Officer Syte Information Technology~~

Mr. Shaw is founder, president, and CEO of Syte Information Technology. Mr. Shaw has held engineering, marketing, and sales positions with several other companies, including Perkin-Elmer, and was named president of Megatek Corporation in 1981. Mr. Shaw is a director of two San Diego based companies, Telesoft Corporation and Systech Corporation, and a former director of Data Systems Corporation. He is founding director and a member of the National Computer Graphics Association, and served as chairman of the Vendor Advisory Committee for the SIGGRAPH special interest group of the Association for Computing Machinery. Mr. Shaw received his bachelor's degree in engineering from the City College of New York and his M.S.A. degree from the University of Connecticut.

~~Dataquest Incorporated WORKSTATION FOCUS CONFERENCE August 13 and 14, 1984 Burlington, Massachusetts~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 DATAQUEST CONFERENCE~~

~~AUGUST 13-1^. 198^1~~

~~"WORKSTATIONS"~~

~~THE EVOLUTION OF AN INDUSTRY~~

~~NETWORK CONFIGURATIONS~~

~~PETER J. SHAW., PRESIDENT~~

~~SYTE INFORMATION TECHNOLOGY INCORPORATED~~

~~SAN DIEGO,, CALIFORNIA~~

~~- J. - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 a CONRGURING A NETWORK~~

~~•o o~~

~~(D a v> c CT Consider CA o!~~

»•

~~> b • Computational Power o o 3 • Utility - Application •a u~~

~~3) to d <^ I (D ' -1 TJ D> • Rexibility~~

~~m • Expansion 01 C_ o (A (D o >~~

~~4^ O (S~~

~~syte ^ August 198A D~~

~~•oa~~

~~Wc a w Q! D>' •< > b~~

~~3 o 3 •o a>~~

~~U3 i o ($) 33 u> d (D T3 0)~~

~~W (tt~~

~~>O CRT Personal Stpennini o Computer 00~~

~~w Workstation: A Definition I 2. > o • Single User o o 3 • Single Processor •o lU 3 lO S ' • Single Function a S- ' (D •D Dl • 1 - 4 Megabytes, RAM~~

~~W • Mid and hb^ Resolution Graphics 0> 3 o > • 50 Megabyte Disk Space/User~~

~~o(O o o~~

~~(O ~4 u ^ytB ^ August 1984 I ENGJNEERING/SCIEmiFIC NETWORK o o •o o~~

~~> w a. S' A Definition: -3 > o~~

~~o A CX)MBINATIQN Of (^imputing Periphei^l^ o 3 •o fii And Communications Elements I Ul Q. a. CD Working Together TJ 0) it To Serve Multiple Application D) 3 o Requirements >~~

~~o 00 oCO o o~~

~~(O •-4 CO Syte August 198A o 0) NETWORK FUNCTIONALITY~~

~~•a o~~

> w Rexible, Expandable: I a • Multiple Levels of Computing > o o ^ Multiple Operating System Support o 3 •o 01 3 • Wide Range Of Multi-Vendor Peripherals N3 g ' 5 (T> Q. Q- I n> ' • A Standard, Accessible Network Medium -1 13 01 • ConrYmnication Between Networks

~~W Di C- o (A _~~

~~•-4 (O ^yte ^ August 1984 I CONnGURATION ALTERNATIVES~~

~~T) O~~

~~(D a > w I a > o A SINGLE SOLUnON o o 3 •o 01 3 SATISFIES NO ONE r^ 5"~~

~~CO o (A (D O >~~

~~.0. o CD~~

~~SytB August 1984 D NETWORKS o o •D O~~

~~<0 a > CO I~~

~~> b A Series of o o 3 •o 0) 3 Single User Workstations M o ' 5 00 Q. <^ I to ' •V May Not Be Enough~~

~~w 0) o >~~

~~oCO o o~~

~~-Nl~~

~~CO *yte ^ August 1984 #- o WORKSTATION s~~

~~M c Benefits: «CT a. sa -5 o^ > p z (0 en * Predictable Response (D o o 3 T3 U * High Degree of Interaction - Good User/Machine Interface 5 VO * Psychology of Ownership~~

~~o >~~

~~4^ O 00~~

~~SytB August 1984 o SUPERMINICOMPUTBR s~~

~~w c cr w When: g! BJ'~~

~~> b • A Very Large Job o o 3 - The Job Is Too Big lO I «o ' o i° (D I -a B> • A Very Small Job~~

~~O) D) 3 - The Jobs Are Too Small o >~~

~~4^ O 00 oCO o o~~

~~(O ^yte August 1984 mo J2 c (D 3 NETWORKED SUPERMINI: O o •S o~~

CD Q. A Modem Architecture CO c tr CO 9! Specia.1 Purpose o>' •5 Ccmputing Accelerators o^ > p z (D' \ o o 3 •a s> Supermini BaselIne (O Tightly Coupled, O 1-J Alpihajujmer 1 c 5 M Bit-HBipped Graphics. g I Terminal SuffXHTyt 1 MIP. 1 Megabyte RAM (D Mid And High Resolution, X u Monochrome And Color.

~~CO 0) Expandable In 3 X Peripherals Gr&ptilcs / Multiples Of Baseline Termin&i. o > Support Mass Storage ICX) Megabytes Disk, t Minimum Tape~~

~~(O -J 03 August 1984 o (u~~

~~o NETWORKS o —1 •oo~~

~~• Workstations o~~

~~(D • Multi-User, Powerful, Computers 3 o o 3 •D - With Graphics D>~~

~~!•' 5 to d - Without Graphics (D - Variable Perfomnance~~

~~CO ID ^ Special Purpose Acoel^ratdrs o > • Clusters ^ Mass Storage and Peripherals oCO o o • Convnunications~~

~~^ytt August 1984 o lu CONCEPT: o o •3 o~~

~~> 0) I Netwoiked Computers ° > o o o 3 •o D> 3 vs. lO I - ' T3 SI A Computer Across The Network~~

~~0> o (A CD o > to~~

~~ot^ 00~~

~~2. n>~~

~~sytff August 1984 o <>> ELEMENTS OF AN ADVANCED NETWORK I m c •<—I > p~~

~~O • Software Dependent o 3 • Resource Sharing N> I to a. Q- I (D ' • 10 - 15 MBits/Sec TO • Workstation and Supermini w a~~

~~O >~~

~~o 00~~

~~CD O O O~~

~~H (D~~

~~^yte August 1984 o NEirWORKED CONRGURATION o o~~

~~single User Workstation~~

~~Special Purpose Accelerator Comrputlng Besource~~

~~[^Gateway~~

~~Server August 198A A NETWORK CONFIGURATION^ OR WHAT TYPE OF EQUIPMENT IS~~

~~CONNECTED TO A NETWORK,, WILL VARY AS WIDELY AS THE APPLICATIONS~~

~~ATTEMPTED AND THE BUDGET CONSTRAINS. To TRY TO DEFINE THE~~

~~OPTIMUM NETWORK WITHOUT UNDERSTANDING THE PARTICULAR ENVIRONMENT~~

~~(PHYSICAL,, TECHNICAL^ SOCIAL OR PSYCHOLOGICAL^ AND FINANCIAL)~~

~~IS AN EXERCISE DOOMED TO FAILURE. THE NETWORK MUST REFLECT~~

~~THE INDIVIDUAL GOALS AND CONSTRAINTS OF THE ORGANIZATION IT -~~

~~ATTEMPTS TO SERVE. AND AS THAT ORGANIZATION AND ITS NEEDS~~

~~CHANGE,, SO MUST THE NETWORK.~~

~~BUT,, RATHER THAN DEFINE THE NETWORK NOW,, LET'S FIRST~~

~~TALK ABOUT SOMETHING MUCH MORE BASIC SOMETHING TD LIKE TO~~

~~CALL AN "ELEMENT"; AND,, SINCE WE ARE NOW TALKING ABOUT COMPUTERS,,~~

~~LET'S CREATE A "COMPUTING ELEMENT".~~

~~PEOPLE BUY COMPUTERS AND SOFTWARE AND GRAPHICS AND^ YES,,~~

~~NETWORKS TOO^ FOR ONE SIMPLE REASON: THEY HAVE A PROBLEM~~

~~TO SOLVE OR INFORMATION TO OBTAIN OR A PRODUCT THEY WOULD~~

~~LIKE TO CREATE. THEY HAVE A COMPUTING APPLICATION.~~

- 16 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 l=lidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 IN ORDER TO REACH A SOLUTION CORRECTLY AND EFFICIENTLY.,

~~ALL APPLICATIONS REQUIRE AT LEAST ONE STRONG COMPUTING ELEMENT.~~

~~AND BY COMBINING THIS COMPUTING ELEMENT WITH OTHER ELEMENTS.,~~

~~BE THEY COMPUTING^ OR GRAPHICS^ OR SOFTWARE., OR NETWORK^ WE~~

~~CAN CREATE THE "ENVIRONMENT" IN WHICH THE USER CAN SATISFY~~

~~HIS REQUIREMENTS. I THINK THAT THIS IS WHERE MANY SYSTEM~~

~~DESIGNERS TAKE THE WRONG TURN. THEY LOOK AT THE FOREST., BUT~~

~~SOMEHOW MISS THE TREES. THE ROOTS ARE NEGLECTED^ THE TRUNK~~

~~IS NOT SECURE., BUT QUICKLY THEY ADD THE BRANCHES., CONCENTRATING~~

~~FIRST NOT ON THE COMPUTING ELEMENT., BUT RATHER ON GRAPHICS~~

~~RESOLUTION AND NUMBER OF COLORS; NETWORK MEDIUM INSTEAD OF~~

~~FUNCTIONALITY AND EXPANDABILITYJ THE GENERIC NAME OF THE DEVICE~~

~~(WORKSTATION^ FOR EXAMPLE) RATHER THAN WHAT IS REQUIRED OF~~

~~THE DEVICE; AND ISOLATED SPECIFICATIONS RATHER THAN SYSTEM~~

~~THROUGHPUT.~~

- 17 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 IT ALL STARTS WITH THE BASICS^ THE COMPUTING ELEMENT.

~~DEVELOP OR ACQUIRE A POWERFUL^ FLEXIBLE^ EXPANDABLE COMPUTING~~

~~ELEMENT CAPABLE OF COMMUNICATION WITH THE USER AND/OR THE~~

~~NETWORK OR THE GLOBAL ENVIRONMENT AND USE IT AS THE BASIC~~

~~ELEMENT OF THE NETWORK. BuiLD THE FOUNDATION^ GROW THE ROOTS.~~

~~IT MATTERS NOT WHAT STYLE THE HOUSE IS OR WHAT NAME THE TREE~~

~~GOES BY^ BECAUSE WITHOUT THE UNDERLYING STRUCTURE., THE HOUSE~~

~~WILL FALL^ THE TREE WILL NOT GROW> AND THE NETWORK WILL NOT~~

~~BE EFFECTIVE.~~

~~ONCE THE COMPUTING ELEMENTS HAVE BEEN DEFINED^ THEY MUST~~

~~BE COMBINED WITH ADDITIONAL HARDWARE AND SOFTWARE TO CREATE~~

~~A NETWORK., DIFFERENT ELEMENTS WORKING AND COMMUNICATING TOGETHER~~

~~TO SOLVE COMPUTING APPLICATIONS.~~

~~NOW., HAVING LAID THE GROUNDWORK^ I WILL CONCENTRATE ON~~

~~DEFINING AND TYING DIFFERENT ELEMENTS OF A NETWORK TOGETHER.,~~

~~AND^ FOR THE PURPOSE OF THIS SESSION., CONCENTRATE ON ELEMENTS~~

~~IMPORTANT IN ENGINEERING AND SCIENTIFIC COMPUTING APPLICATIONS.~~

- 18 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 THE VARIOUS ELEMENTS OF A NETWORK NEED TO BE AS DIVERSE

~~AS THE TYPES OF APPLICATIONS RUN BY USERS ON THE NETWORK. IF~~

~~WE WERE TO EXAMINE A TYPICAL ENGINEERING OR R & D DEPARTMENT~~

~~OF A TYPICAL CORPORATION., WE WOULD SEE THE DEMAND FOR A WIDE~~

~~RANGE OF DEVICES ON A NETWORK.~~

~~COMPUTING RESOURCES ON A NETWORK CAN RANGE FROM A SIMPLE~~

~~VDT CONNECTED VIA LOW SPEED RS 232 COMMUNICATIONS TO A TIME-SHARING~~

~~MICRO TO A SUPER-MINI DEDICATED TO A SINGLE USER. GRAPHICS~~

~~MAY OR MAY NOT BE REQUIRED. AND EVEN WHEN GRAPHICS IS NEEDED^~~

~~THE NUMBER OF COLORS^ RESOLUTION^ AND FUNCTIONABILITY MAY~~

~~VARY SIGNIFICANTLY.~~

~~IN CONFIGURING A NETWORK THAT WILL MEET THE NEEDS OF~~

~~A PARTICULAR ORGANIZATION,, A CAREFUL STUDY MUST BE COMPLETED~~

~~TO DETERMINE THE TYPE OF WORK TO BE DONE^ THE DIFFERENT AVAILABLE~~

~~HARDWARE AND SOFTWARE TO DO THE WORK,, AND THE AMOUNT OF MONEY~~

~~AVAILABLE TO IMPLEMENT THE SYSTEM.~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 IN DESIGNING A NETWORK^ CARE MUST BE TAKEN SO THAT FUTURE

~~NEEDS CAN EASILY BE ACCOMMODATED. THIS MEANS NOT ONLY ADDITIONAL~~

~~LIKE DEVICES^ BUT ALSO NEW COMPUTATIONAL ELEMENTS THAT WILL~~

~~TAKE ADVANTAGE OF NEW TECHNOLOGIES AND NEW APPLICATIONS^ AND~~

~~WHICH WILL MEET THE FINANCIAL ABILITIES OF THE USERS IN THE~~

~~ORGANIZATION.~~

~~A NETWORK MUST BE FUNCTIONAL^ FLEXIBLE^ AND ECOMONICAL.~~

~~A TYPICAL GRAPH OF THE TYPE OF MACHINES THAT MAY BE REQUIRED~~

~~BY NETWORK USERS MIGHT LOOK LIKE THIS:~~

~~- 20 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 00 ON H~~

~~i .Li~~

g t - 21 - V Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 FOR TODAY'S PURPOSES^ LET US CONCENTRATE ON WORKSTATIONS

~~AND SUPERMINIS AND THEIR ROLE ON A NETWORK.~~

~~FIRST^ LET'S DEFINE WHAT WE MEAN BY WORKSTATION AND.,~~

~~FURTHERMORE^ LET'S TAILOR OUR DEFINITION TO ENGINEERING AND~~

~~SCIENTIFIC ENVIRONMENTS. THE WORD WORKSTATION TAKES ON A~~

~~WHOLE NEW DEFINITION IN THE OFFICE ENVIRONMENT. THE DEFINITION~~

~~I WOULD LIKE TO USE IS:~~

~~ENGINEERING/SCIENTIFIC WORKSTATION~~

~~DEFINITION:~~

~~• SINGLE-USER~~

~~f SINGLE CENTRAL PROCESSOR~~

~~f SINGLE FUNCTION~~

~~• 1 - ^ MEGABYTE MEMORY., MINIMUM~~

~~0 MID OR HIGH RESOLUTION GRAPHICS~~

~~• 50 MEGABYTE DISK SPACE/USER~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 SECOND^ LET'S DEFINE WHAT WE MEAN BY A NETWORK; AND AGAIN

~~LET'S LIMIT OUR DEFINITION TO A NETWORK USED IN AN ENGINEERING~~

~~AND SCIENTIFIC ENVIRONMENT:~~

~~ENGINEERING/SCIENTIFIC NETWORK~~

~~DEFINITION:~~

~~f A COMBINATION OF VARIOUS COMPUTING^ PERIPHERAL.,~~

~~AND COMMUNICATIONS ELEMENTS.~~

~~• WORKING TOGETHER.~~

~~• To SERVE MULTIPLE APPLICATIONS REQUIREMENTS.~~

~~PLEASE NOTE TWO KEY WORDS IN THE PREVIOUS SLIDE: COMBINATION~~

~~AND MULTIPLE.~~

~~LET ME BRING THESE TO YOUR ATTENTION-AGAIN. A USEFUL~~

~~NETWORK MUST COMBINE WIDE RANGING LEVELS OE COMPUTATION. P0WER>~~

~~AND FUNCTIONALITY WITH PERIPHERALS AND COMMUNICATIONS IN ORDER~~

~~THAT IT FULFILLS THE REQUIREMENTS OF MULTIPLE APPLICATIONS~~

~~ON THE NETWORK.~~

- 23 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 IN ORDER TO ACCOMMODATE MULTIPLE APPLICATIONS^ THE NETWORK

~~SHOULD HAVE AVAILABLE:~~

~~• COMBINATIONS OF ELEMENTS WITH DIFFERENT LEVELS OF~~

~~COMPUTING PERFORMANCE AND FUNCTIONALITY;~~

~~f DIFFERENT OPERATING SYSTEMS TO ACCOMMODATE A WIDE~~

~~VARIETY OF APPLICATIONS SOFTWARE AND TO ALLOW EXISTING~~

~~USERS TO PRESERVE THEIR SOFTWARE INVESTMENT AND STILL~~

~~TAKE ADVANTAGE OF NEW OPERATING ENVIRONMENTS;~~

~~• COMMUNICATIONS BETWEEN DIFFERENT USER ENVIRONMENTS.~~

~~§ STANDARD HARDWARE AND SOFTWARE INTERFACES^ ACCESSIBLE~~

~~BY INDEPENDENT EQUIPMENT VENDORS (E.G.,, ETHERNET^~~

~~TCP/IP);~~

~~• PERIPHERALS FROM A WIDE VARIETY OF VENDORS TO SATISFY~~

~~SPECIFIC USER NEEDS;~~

~~TO SUM IT UP> A SINGLE NETWORK SOLUTION THAT PROVIDES~~

~~FEW ALTERNATIVES TO THE USERS IS GENERALLY A COMPROMISE THAT~~

~~SATISFIES NO ONE OR A SPECIFIC CONFIGURATION THAT WORKS WELL~~

~~FOR A PARTICULAR CLASS OF USERS,, BUT LACKS EITHER THE PERFORMANCE~~

~~OR ECOMONIC JUSTIFICATION FOR A WIDE RANGE OF APPLICATIONS.~~

- 24 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Now THAT WE HAVE DEFINED WORKSTATIONS AND NETWORKS AS

~~THEY PERTAIN TO ENGINEERING AND SCIENTIFIC ENVIRONMENTS., LET'S~~

~~LOOK AT THE ADVANTAGES OF WORKSTATIONS. THEN LET'S LOOK AT~~

~~THE SUPERMINI AND POINT OUT THE DIFFERENCE BETWEEN IT AND~~

~~A WORKSTATION^ AND PERHAPS AGREE ON WHEN A SUPERMINI ON A~~

~~NETWORK MIGHT ACTUALLY BE MORE APPROPRIATE FOR A PARTICULAR~~

~~CLASS OF APPLICATIONS THAN A WORKSTATION. AFTER THATy I WILL~~

~~SUGGEST THAT THE ULTIMATE ENVIRONMENT MAY BE THE COMBINATION~~

~~OF WORKSTATION ELEMENTS AND SUPERMINI ELEMENTS IN ONE INTEGRATED~~

~~NETWORK. A NETWORK OF COMPUTING ELEMENTS MUST SERVE THE APPLICATION~~

~~REQUIREMENTS AND FINANCIAL CONSTRAINTS OF ALL THE USERS.~~

~~AND IT IS VERY PROBABLE THAT THE PROPER CONFIGURATION IS NOT~~

~~WORKSTATIONS ONLYy BUT A COMPATIBLE MIXTURE OF DIFFERENT LEVELS~~

~~OF COMPUTING ELEMENTS.~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 THE INTEGRATED WORKSTATION HAS SOME CLEAR ADVANTAGES^~~

~~AND I BELIEVE THAT IT WILL BECOME THE DOMINANT COMPUTING ELEMENT~~

~~IN THE FUTURE. LET'S EXAMINE THE REASONS:~~

~~1) WORKSTATIONS^ BETTER THAN MINIS., PROVIDE THE USER~~

~~WITH A PREDICTABLE RESPONSE. THE RESOURCES ARE DEDICATED~~

~~AND WITHIN A SPECIFIC CLASS OF APPLICATIONS., THE~~

~~MACHINE ACTS IN A CONSISTENT MANNER.~~

~~2) WORKSTATIONS PROVIDE THE USER WITH A HIGH DEGREE~~

~~OF INTERACTION. THEY ARE GENERALLY EASY AND COMFORTABLE~~

~~TO USE. THEIR BIT-MAPPED DISPLAYS AND,, TYPICALLY,,~~

~~THEIR WINDOW-MANAGED,, MOUSE'ORIENTED INTERFACE,, CREATE~~

~~AN EFFICIENT,, PRODUCTIVE^ AND "FRIENDLY" OR FRIENDLIER~~

~~ENVIRONMENT.~~

~~3) THE WORKSTATION BECOMES THE PRIVATE PROPERTY OF THE~~

~~USER. PSYCHOLOGICALLY-* THE PROGRAMMER^ DEVELOPER^~~

~~OR IMPLEMENTOR OWNES THE MACHINE. FEELING GOOD,,~~

~~IMPORTANT,, AND IN CONTROL ALL MAKE THE WORKSTATION~~

~~POPULAR WITH THE USER.~~

- 26 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 AT TWO OPPOSITE ENDS OF THE SPECTRUM,. HOWEVER,, THE WORKSTATION

~~BECOMES LESS THAN THE IDEAL SOLUTION: THE JOB MAY BE TOO~~

~~BlGj OR THE JOB MAY BE TOO SMALL.~~

~~LARGE COMPUTE-INTENSIVE,, MULTI-TASKING APPLICATIONS VERY~~

~~OFTEN REQUIRE COMPUTING RESOURCES OR SPECIAL PURPOSE ACCELERATORS~~

~~THAT ARE NOT AVAILABLE WITH A TYPICAL WORKSTATION. THE POWER~~

~~OF A SUPERMINI INDEPENDENT OF COSTJ MAY BE REQUIRED JUST~~

~~TO GET THE JOB DONE.~~

~~ALTERNATELY,, SOME APPLICATIONS MAY REQUIRE VERY SMALL~~

~~AMOUNTS OF COMPUTER RESOURCE AND> ALTHOUGH A TYPICAL WORKSTATION~~

~~COULD EASILY HANDLE THE JOB,, THE EXPENSE OF DEDICATING THE~~

~~ENTIRE MACHINE TO A SINGLE'USER IS HIGHER THAN THE APPLICATION~~

~~CAN JUSTIFY OR MORE THAN AN ALTERNATE SOLUTION MAY COST,~~

~~A SUPERMINI SUPPORTING MULTIPLE USERS MAY PROVIDE THE PERFORMANCE~~

~~LEVELS DEFINED FOR SPECIFIC APPLICATIONS WHILE SHARING THE~~

~~COST OF A MACHINE ACROSS A NUMBER OF USERS. THE RESULT: A~~

~~COST/USER LOWER THAN THE COST OF A DEDICATED MACHINE WHILE~~

~~THE USER RECEIVES ADEQUATE OR APPROPRIATE RESPONSE IN A MULTI'USER~~

~~ENVIRONMENT.~~

— 9*7 — Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ADD TO THIS A NETWORKING ELEMENT., PERHAPS SIMILAR TO

~~THAT USED BY WORKSTATIONS AND WE HAVE CREATED WHAT I CALL~~

~~A NETWORK SUPERMINICOMPUTER. LET'S EXAMINE ITS FEATURES.~~

~~FlRST^ IT IS BUILT AROUND A BASIC COMPUTING ELEMENT^~~

~~PERHAPS EVEN THE SAME ONE THAT FORMS THE FOUNDATION OF OUR~~

~~WORKSTATION. NEXT., AND MOST IMPORTANTLY^ IT CAN PROVIDE MULTIPLE~~

~~LEVELS OF PROCESSING POWER USING THE COMPUTING ELEMENT AS~~

~~BASE LEVEL 1 AND COMPATIBLE MACHINES 3-5 TIMES MORE POWERFUL~~

~~THAN THE SINGLE ELEMENT MACHINE.~~

~~ALSO> ANY LEVEL OF COMPUTING PERFORMANCE CAN BE DEDICATED~~

~~TO ONE USER OR SPREAD AMONG MANY USERS^ DEPENDING UPON APPLICATION~~

~~REQUIREMENTS AND USER DEMANDS. GRAPHICS IS CERTAINLY A REQUIRED~~

~~OPTION., BUT ITS FEATURES AND BENEFITS SHOULD BE AVAILABLE~~

~~TO MULTIPLE USERS. ALSO., DEPENDING UPON THE APPLICATION OR~~

~~MIX OF APPLICATIONS^ USERS SHOULD BE ABLE TO SELECT FROM A •~~

~~VARIETY OF GRAPHIC DEVICES COLOR OR MONOCHROME., TIGHTLY COUPLED~~

~~OR TERMINAL DEVICES., 2D AND 3D,, INEXPENSIVE TO FULL FEATURE.~~

~~COMPUTE AND DISPLAY INTENSIVE APPLICATIONS LIKE SOLID MODELING~~

~~ANALYSIS AND RENDERING SHOULD HAVE SPECIAL PURPOSE DEVICES~~

AVAILABLE TO EFFICIENTLY CALCULATE AND DISPLAY DATA. - 28 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ON THE OTHER HAND^ MANY APPLICATIONS CALL FOR A SIMPLE

~~ALPHA-NUMERIC TERMINAL AS THE I/O DEVICE OR PERHAPS A CLUSTER~~

~~OF TERMINALS IS THE APPROPRIATE SOLUTION. No NEED TO INCREASE~~

~~SYSTEM COSTS — LOWCOST TERMINALS MUST BE AVAILABLE.~~

~~MULTI-USER MACHINES OR NETWORKS WITH MANY USERS IMPLY~~

~~DIFFERENT APPLICATIONS. THUS,, A TRUE MULTIUSER ENVIRONMENT~~

~~SHOULD HAVE AVAILABLE A LARGE SET OF TOOLS FOR DEVELOPMENT~~

~~AND EXECUTION. MULTIPLE CONCURRENT OPERATING SYSTEMS^ EASILY~~

~~AVAILABLE TO EACH USER^ WOULD BE IDEAL. DIFFERENT OPERATING~~

~~SYSTEMS SHOULD COMPLEMENT EACH OTHER WITH RESPECT TO THE TYPE~~

~~OF APPLICATIONS THAT THEY ARE MOST IDEALLY SUITED. UNIX>~~

~~FOR INSTANCE^ FOR SOFTWARE DEVELOPMENT^ SMALLTALK FOR GRAPHICS~~

~~AND SIMULATION^ PICK PERHAPS^ FOR DATABASE.~~

~~AND LET'S NOT FORGET COMMUNICATIONS. THE ABILITY TO~~

~~FREELY COMMUNICATE WITH OTHER ELEMENTS ON THE NETWORK^ AS~~

~~WELL AS OTHER NETWORKS OR ENVIRONMENTS., IS IMPORTANT IF WE~~

~~ARE TO RETAIN FLEXIBILITY AND COMPANY-WIDE ACCESS,, SUBJECT~~

~~TO SECURITY CHECKS^ TO ALL INFORMATION.~~

- 29 ~ Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 As WITH MOST THINGS IN THIS WORLD> THERE IS NOT ONE PERFECT

~~SOLUTION. BUT A NETWORK THAT SUPPORTS DIFFERENT ELEMENTS~~

~~THAT MEET THE NEEDS OF A DIVERSE NUMBER OF USERS AND APPLICATIONS~~

~~IS SURELY SUPERIOR TO A NETWORK THAT SUPPORTS ONLY ONE SPECIFIC~~

~~CLASS OF MACHINES.~~

~~FURTHERMORE^ I MAINTAIN THAT A NETWORK THAT NOT ONLY~~

~~SUPPORTS BUT INTEGRATES DIFFERENT CLASSES OF WORKSTATIONS~~

~~AND MULTIPLE-USER SUPERMINIS IS MORE PROBABLE TO MEET THE~~

~~REQUIREMENTS OF THE ENGINEERING AND SCIENTIFIC DEPARTMENTS~~

~~THAT WE SERVE. THE ABILITY TO CONTINUE TO LOWER THE COST~~

~~PER USER^ NOT ONLY BY REDUCING THE PRICE PER ELEMENT^ BUT~~

~~ALSO BY BEING ABLE TO SHARE CERTAIN ELEMENTS^ WHEN APPROPRIATE^~~

~~MAKES THE DAY A LOT CLOSER WHEN THE ENGINEERING AND SCIENTIFIC~~

~~NETWORK IS INTEGRATED^ OR EXPANDED^ WITH THE COMPUTING REQUIREMENTS~~

~~OF OTHER DEPARTMENTS. BUT THAT'S THE TOPIC OF ANOTHER TALK.~~

- 30 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Now,, LET'S GO ONE STEP FURTHER AND SEE WHAT HAPPENS WHEN

~~WE ATTEMPT TO INTEGRATE DIFFERENT COMPUTING ELEMENTS INTO~~

~~ONE NETWORK ENVIRONMENT.~~

~~MoST NETWORKS,, TODAY^ PROVIDE A MORE EFFICIENT,, HIGHER~~

~~SPEED MECHANISM TO TRANSFER INFORMATION FROM ONE NETWORK ELEMENT~~

~~TO ANOTHER.THAN THE TYPICAL SERIAL COMMUNICATIONS AVAILABLE~~

~~WITH MOST COMPUTERS. THIS IS WHAT I CALL A NETWORK OF COMPUTERS.~~

~~THE NEXT STEP IS A COMPUTING NETWORK^ OR THE CREATION~~

~~OF A COMPUTER ACROSS THE NETWORK,, A NETWORK ENVIRONMENT.~~

~~IN THIS CASE,, A USER SEES NOT ONLY THE MACHINE THAT HE HAPPENS~~

~~TO BE SITTING IN FRONT OF,, BUT> RATHER^ ALL OF THE COMPUTING~~

~~RESOURCES ON THE NETWORK APPEAR TO BE AVAILABLE TO ANY ONE~~

~~OF MANY SINGLE USERS.~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 TODAY'S NETWORKS ARE FAST ENOUGH TO ALLOW FOR THE CREATION

~~OF SUCH A COMPUTING ENVIRONMENT. YET VERY FEW COMPANIES HAVE~~

~~DEVELOPED THE SOPHISTICATED SOFTWARE REQUIRED TO CREATE THIS~~

~~ENVIRONMENT. THE DAY IS NOT FAR OFF^ HOWEVER^ WHEN THIS TYPE~~

~~OF SOFTWARE WILL BE READILY AVAILABLE. TRUE NETWORKING SOFTWARE^~~

~~WHEN COMBINED WITH ADVANCED., POWERFUL^ FLEXIBLE COMPUTING~~

~~ELEMENTS^ WILL PROPEL US INTO A NEW ERA OF COMPUTING: THE~~

~~COMPUTING NETWORK.~~

~~AT CURRENT NETWORK SPEEDS., EVEN A COMPUTING NETWORK MAY~~

~~NOT BE FAST ENOUGH TO CLOSELY COUPLE MULTIPLE CPU'S AND PROVIDE~~

~~A LARGE NUMBER OF USERS WITH THE* LEVEL OF PERFORMANCE^ AT~~

~~A NEW LOW LEVEL OF COST,, IN A MULTIUSER ENVIRONMENT^ THAT~~

~~HAS BEEN AVAILABLE FROM LARGE DECy DATA GENERAL^ G0ULD> OR~~

~~OTHER SUCH MACHINES. AND UNTIL 50-100 MBiT PER SECOND NETWORKS~~

~~ARE RELIABLE AND AFFORDABLE,, THE EFFICIENCIES AND PERFORMANCE~~

~~OF A MULTI-PROCESSOR MACHINE MAY BE DIFFICULT TO DUPLICATE~~

~~OVER A NETWORK.~~

- jz - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 THE DAY WILL COME^ HOWEVER., WHEN THE COMBINATION OF SUPER

~~FASJJ EFFICIENT^ SOFTWARE INTENSIVE^ COST EFFECTIVE NETWORKS~~

~~WITH MODERN WORKSTATION AND SUPERMINI COMPUTING ELEMENTS ARE~~

~~JOINED TO FORM A SUPERMINI ACROSS THE NETWORK. A COMPUTING~~

~~RESOURCE OF AWESOME PROPORTIONS WILL BE THE RESULT.~~

~~THUS^ DEPENDING UPON THE APPLICATION^ EVEN A NETWORKED~~

~~SOLUTION WILL REQUIRE AT LEAST TWO LEVELS OF COMPUTER TO MEET~~

~~THE NEEDS OF A DIVERSE NUMBER OF USERS AND APPLICATIONS. A~~

~~SOFTWARE-DEPENDENTy RESOURCE-SHARING NETWORK OF WORKSTATIONS~~

~~AND SUPERMINIS^ SINGLE AND MULTIUSER,, EASILY CONFIGURED TO~~

~~MEET THE REQUIREMENTS OF A PARTICULAR ENVIRONMENT SHOULD BE~~

~~THE ARCHITECTURAL COMMON DENOMINATOR FROM WHICH TO DEVELOP~~

~~THE SPECIFIC SOLUTION FOR AN INDIVIDUAL INSTALLATION.~~

- 33 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 LET'S GO BACK TO BASICS. START WITH A POWERFUL^ FLEXIBLE

~~COMPUTING ELEMENT, USE IT TO CREATE CONFIGURATIONS THAT MEET~~

~~THE REQUIREMENTS OF THE WIDE VARIETY OF APPLICATIONS TO BE~~

~~RUN. CONNECT THE ELEMENTS AND CREATE NOT A NETWORK OF COMPUTERS^~~

~~BUT RATHER A COMPUTING NETWORK.~~

~~BUILD A STRONG FOUNDATION. REMEMBER^ COMPUTERS AND NETWORKS~~

~~SHOULD BE THE SLAVES. THEY ARE THERE TO SERVE THEIR MASTERS^~~

~~THE USERS. START BY DEFINING THE PROBLEM AND THEN PROVIDE~~

~~THE NETWORK ELEMENTS THAT WILL MOST EFFECTIVELY SOLVE PROBLEMS.,~~

~~NOT CREATE NEW ONES.~~

~~IT'S ELEMENTARY!~~

~~THANK YOU FOR YOUR ATTENTION. I WOULD BE HAPPY TO ANSWER~~

~~ANY QUESTIONS.~~

~~- J4. - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dataquest Dataquest~~

~~START-UP: BUILDING THE MANAGEMENT TEAM~~

~~Arthur T. Campbell President Mosaic Technologies, Inc.~~

Mr. Campbell joined Mosaic Technologies, Inc., in late 1983 as President, Chief Executive Officer, and Director. From 1970 until joining Mosaic Technologies, he was employed by Digital Equipment Corporation in various management and marketing and sales positions. Most recently, he was the general manager of Digltuil's Terminals Product Group. Mr. Campbell has been directly involved with the con^uter industry in management, sales eaid marketing, and engineering capacities for more than twenty years. Mr. Campbell received a B.S.E.E. from Tufts University, an M.S.E.E. from Northeastern University, and an M.B.A. from Harvard University.

~~Dataquest Incorporated WORKSTATION FOCUS CONFERENCE August 13 and 14, 1984 Burlington, Massachusetts~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o 0>~~

~~> w~~

~~o BUILDING THE STARTUP TEAM 2 a a o 3 U O I People Make the Difference" a a~~

~~o Mosaic Technologies, Inc~~

~~o 00~~

~~(£> oO o : Jka rm las las lias HQS ims Eins ^QS sae aae QBB • • SQE CO »Bi QOi QSi Qas see < QOB iQQi o. iSQHQOl mm •nm o IJEia lUB o m3 LU~~

~~< - 2 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 tin ilia~~

~~^"iimi ^ms mum i"HI lU i^lOIBBBBB tL II ill ^;^l o Wi '. ,i £^âi ly 'r^n :••.- - i-j> O WJSBEEEB tL •^ fJ3BBEBB 0 S ISâBBBBB •__,;; '--^ 0> ÎSISBBES~~

~~^•1 .i^t E l^'W O giaBBEBi~~

~~;,..••,'• •• ••»• ••:.;••- •^ 5'ISSEEli UJ £ SSllBBEBB •o »^Lir— IL O w,^ BEli (P D•B BD Qsps!^|S^BEii i e € PjfgiBBl•:;.OJ5 S a •S .B S-^^K^BBias~~

~~^ :3 -^ EJataquesu. t Incorporated, A Subsidiary o( A.C. Nielsen Company / 1290 Ridder Park Drive I San Jose. CA 95131 / (408) 971-9000 I Telex 171973~~

~~iiaiBBBB •;:i^~~

•^^^ ••si i^ isaaBBBBi B v-ti o o ?3 o ^ i•' 5 BB EI^IBEBE EBB 11 BB msm 11 IE Sg BB mumsrr,\ BB mmm PI laSBBBB E ^1 o S! ISfiSEEBEi li aEEBBi o Eili (0 iir^ai BBi m^'M S o m IISEB !i CL m WHAT o (f a. > cCO cr

> b D High Performance, Integrated Graphics Technical Workstations o o 3 •o 0> D Covering $25,000 - $60,000 I Price Range 0^ I • New Standards of Price Performance a Parallel Processing Hardware . jB a_ »» _. ^ tL^m^ ^ *^-^ '^''* '*'' **^ ***^' ^''*" ••***' ^*^** *""-' ***" **^ *••** '*''* •MiClbt kBX tsd I«n *M ln>M O> ^Mi^fBtst^ m

~~e ^n o ^ 31 ISEBEB "O iS '^311 Bgg o BB Bl •o (0 o 00-^ li 5CO >l iii O"" B@li x: o 0 CO Bli "c "oL c e 10 o E iii lU o£ BEi (0 C CO>i 0o> <(0 BIS ODD BIO SBiSQ o~~

~~- 7 - Dataquest Incorporated. A Subsidiary of A.C Nielsen Company / 1290 Bidder Park Drive I San Jose. CA 95131 / (408) 971-9000 / Telex 171973 D~~

~~•o o~~

~~> w a. 0) TECHNICAL USER NEEDS > O~~

~~O D Responsive System o 3~~

~~N) g ' 5 00 a •^ I CJ Sophisticated Application -o lU Software Environment o~~

~~-' • v. v-.v^^^^:B D Customlzatloh and O> Growth Capabilities~~

~~v> o o o D so~~

~~> w TECHNICAL USER NEEDS MOSAIC PROVIDES~~

~~> D Integrated Design b D Responsive System D 32*Blt Computer System~~

~~3 o D Bit-Slice Graphics Processor o 3 T3 D Object Oriented Graphics and M O I D Sophisticated Application Data Management Environment VO Software Environment a Software Standards •V 0>~~

~~CO OJ o > •i; _ _ Vski^~~

~~^'U•^-9 ;K8~~

gsa O C0 sig ias a lU rîSSBBBBB E O i^^SlBBBBB o Z- Bîasi o « jîî"' M\ C t^^ÔSLiHlBiCBH .= £ Li^SBBBBB ^ O M^^^BBBBB S- O S ^SSEBEB S H o?iS£EEBB «j T: «S3^1!aBEBSB ^ S »- -cs li^SEBEB f\ O O e '^^SEBEB Z^ '1 i=U^^Siiii§ (P e &I! !.îilBEil O S-ll ÎSIBBBinÊEBSBB CO k. S g ^^SBBSB 7^ Q. W Q :4£|11S1B VJ DO lllBBiS

~~- 10 - Dataques0)t Incorporated , A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose. CA 95131 / (408) 971-9000 / Telex 171973 O o 01~~

~~o o •o MOSAIC UNIQUENESS o~~

~~> en~~

~~a (U~~

~~> b D Integrated Graphics Computer o o System 3 D High Performance 3J a. o. (D O Open Architecture~~

(/) (U D Networiced, Object Oriented 3 o <.%^* '•X^4 i- > Software Architecture^^sssisiai-iii u-^ tat PiA c^t M„>. ^ J L.ifa '^JA M-i-^ b'M WL4 9>^ tîfa idt^ A^ na b^ w(^4 uiM MM mw MM fen mm WmmtmBtmmmtM MU ^•~ :.--.v I..'« i-"fi r-:./a ti.,i enrn fâe n~M u^ Qia CttJl mU OAB BMB183 Bitt CU tliU IwSd ifiM ICdl liUl 013 Bttl Sill lUB UU Udl Ufl fcU fi323q)&Sia9eSSlR3sSa)E«9E23EG^ E3 O EU9 E3 B9 QUS IS 19 B9 ESI BS S9 EO13 Bi Q! Bl nn nn •BHH i 1SSSBH8GBSS .^ miBBBB fS laaBBBB UJ '41iBBBB Qt i Z H oilSEBBB < L. SiaSBBBB ^ ^ 33EBBEB (Q ^ yaiBBBB < liSBBBBB SIIBBBBE IIISlBBBi ^KllBBBi (/) ^ KI3SBBBE — ••• m S1SSBBBE ^ U lU '^lEBBBi U^ 3 Z &]a£iBi° ._ O U SISSBBB r" ^ OC ^^SIBS a.I Z CL îStlBEBi X "^ its yJSEBBB C CC J'JIIBBS ^ I— {!IllBBi O3 Z IJSBBSi - 12 - Dataquest .Incorporated^ , A Subsidiary of A.C, Nielsen Company / 1290 Ridder Park Drivme / Sa n Jose, CAlilSBEE 95131 / (408) 971-9000Q / Tele x 171973 S "" •ilBBBi QMl '-âÊBBBQ i; laiBHB 3BBBBB IBBBBB

.*, OIBIBIB FiiBBBBB âSEBBB PaSBBBB HI EiaSBBBB y^BBBBB «tiaBBBBB o. ^t ly a %E OSBEBBBr^3EBBBB S.«iaBEBBB Q M^ÎSEBESB 1. KîSBSii a ÎSlEBBi «« o IS • ss îgilEBBB UJ E laiBiBi iiilEBBi IL o IJSIBBO yj D D o ;iasgiEB LflgBBi D (-ISBBBfi

~~fli^Bi~~

H - 13 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 m liaiHM ^.laiBBBB 7i ^aBBBBB iSlBBBBB rlSBEElB

~~BSBEEEB qâ MigiiiBEie 'f^lllE^S îllSlBElE ^-^lElEES S^ÊEEEB IB ^SEEBEB O SllBlBI CO S S'ÊEEi yj MS E&^sisii ^ S e 5^ i^SEBSE H o « S o^'SBeii •J ^ ^ ^ T: TMESIS ^ D D D D Qis^ss~~

~~H - 14 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o~~

~~CD THE CULTURE a > en c cr~~

> i:^ I D Common Goal a: IT Tl a Results Oriented a- o D Responsive to Change cr 01 3 o D Realistic > O People OrientedMds^^MiiM«iM2sai£»^jM^ M^^^MJ^WiMM^MMI^WkMMMMfSAMMWMnWiMWkM'UifnW OE3^S9E9ES9^li!^E!9Bll^GQBIEISBSIIE^IfflBli9B9n0iniBHnHnHI i ••••••• aniHniBHHi .1X1

IBBBBfl II SIBBIB CO aSBBBB IBBBEl UJ £ aiBBBi •T A ^ 3BBEBB Q S ^ ^ ^IIBBIB Ss O le 3£BB@§ ^ '^ c S (IlEBBB ^ « I -g '^iiiii SlBSgli X O O SlBBli 111 I- O llBiBi mm X K- - 16 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 (0 UJ o fiC 3 X X o X (L 3

< (0 U X I- - 17 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o o •o o HURDLE #1 Organization and Finance o

~~o o 3~~

~~en 01~~

~~>O a Technical Workstation Focus~~

*»• D Unique Solution Mapped o CD to Focus Need CD oO o PJOtutetandin^^^ t«« Ui SI i»3 iti ¥il ?" UiannmLi.

ID tjlhiJi -si CO •1 -T • - fl Jl ~.^ 0) V3 ix'f4 •r • -m c 3..1 O m m c u O) V2 _cg '5 •••• 0) 13 Q (O •o m «> iii 3:1 E "c 0) lil o c -0c) c I:] "K * O (0 m 3 m VA w a o o £ i..j

~~^^s5^s!:^w ^11~~

~~- 20 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 (0 o 3 •o o (^ (D UJ O <0~~

~~og •o w ^a o JO o '5) ••§• o< 0) CO U. (D ^ O To 0)4= u o'E Q-CO • MB CO ^£ E-D t> D) (D en o c Q UJ C Sl- O (0 H D Q 0) D cc-S (0~~

- 21 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ? (0 a osaic Prttktonl and f CM«f Exccirtiv* orrictr A. CsmptMll o > I 6

~~;, 'ett^MWttlfl o p3 Hanlww* DavvtopiDMit Softwftr* NJ to Dttvolopfiwiit I M«ch«ilcal 5" E Engliwsrtng~~

~~CO (U~~

~~o >~~

-I to o o m '1 .lis laa O c O. K ^^ lU ?]e3 o o> CO "5 roi O c (0 <0 « igj c .2 IgicJ (0 CO (0 a "c E 4) o Wl o ® E 'Si Q.C 2 O ISil 3 4) 3 O iPi O) ' •= > *: 4- O loi So>' 3 > 2& %i(0 (/)liJ a ^ LU o D D mJ CO CC c 3 23 Dataquest Incorporated£S, A Subsidiar y of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose. CA 95131 / (408) 971-9000 / Telex 171973 m:^ • ^ \r:.T 8 £

~~a III I ^1U iS = p III1^ -S^ i.ti_Lj—I S3i ' 1S5 i |fl< iH I l_l~~

~~'(^E^fjtJLdI ^>**P A„^.~~

~~xo So Su J I I~~

~~- 24 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o~~

~~HURDLE #5 Sales~~

~~> ir)~~

~~o o 3 n~~

~~M I (O ^.^ a~~

Tl Bl a Sales Mgmt. Experience - Computer Systems, Workstations o > D Target Account Focus D Outstanding Sales Skills >.:... « ... .4 Lc.« LiJ ::>uJi i..^.'. t.-^ tjiJ LJI Kki Lul a.*^ •.—< luU k^u L^bi kuJ L..^ ti^ li^ E^ iJ^ '*-,,...;: wc. b.^.'. l...U..'.'^ o w oT .• c HURDLE #6 Customer Support

~~> b o o 3 "O B)~~

~~M I~~

~~w DJ D Computer Systems Experience~~

~~o > D Planning and Control Skills~~

~~o•tx CD D Services as Business Leverage~~

<£> O O O D Customer Satisfaction Driven •*... 1 . -.^ •..!.. » .- .a ;^:M t. ..> • • u ti.il 4. A.* r...., ir. ^•4 »'_i i '..:k> Ld.itri k_ j lau 1 I'n >. _i liiid l*a«:t t«t;i.t •.«> t,4jU t...-j K«I t .. •>. - ... J :,. .* fi-.' i*-J i^l L^ .Jj.i It— i^ Cki BLii: 1(1 j LL j \.-i.Ji \K.^ ihirJ WM tU-l ku'> i.,'j £*(«; L^Lt Uu, i*. J k'.U S*. ! i-vJii t i'. -vl

~~CD -J CJ o o o~~

~~> CO HMosaic Pr««ld«nt and c (c^r CN«f Exacutiv* Offle«r Q! OJ A. CamplMll~~

~~> b Ltijiyya.'ftJt'^. ms^j o o 3 T> Haidww y«rk«l KllanirfftctHteo DavalopnMiit D«V*lopHMflt Ouaty 5 to Soltwar* t MsrfcAt ProgrMM AMurane* ci Q. D«ylopwwt I (t t TJ llMhwrieal lU E wi^r^"^~~

~~CO Dl MttlO fMM~'v1 iu.*jiu-.'%..»*^*w.'^ Dtotrtet llMWflM o> hPf S>tot8«yport~~

~~ILMcOiMy Mstrtct 8wTlc« Uanag^r* ll«MlqiMrt«r« Product 8u|ipwl ID -vl EtfucaUon CO E HURDLE # 7 Do it Again~~

~~> b o o 3~~

~~a5 ^ 5 ' -D tu~~

~~V) D Long Term Orientation o D Build a Company, Not > Just a Product D Market Focus <<-^Technology~~

§ Managementj I Krt4 h>. J >...'.A iA»J *.>-'•*! JUii ^tyj bitd CV't. tTi4 !•.*>» k-t.'< ».• ^ lur^Ljt ^^ifiVAAV- -I .-I -J ./•<•, •;>•(•" .f' -( •-| •• •'*' .•ii'"-'.!' • l.^.J K atti l«tl''l "W-l H.r'' »i.>l- • ,' t^s^' i.i—' #»•-;' t..i..J '-vi., * _.. V'JH Jllt'V iit.,ii^f;*4 i-,-l.i..-.rA.Ki' D [u

~~10% Pr»tld«nt and W CM«f Exsctrthr* Offtcar c IT A. CamptMll I~~

SSESSSSaSSTi > _,.4,v-4*'Kiri^B T o ;^f*-^l .•T^JI^gTtffl Vml99 >naHVfn9VL ,^ mwm m^ffj^^iti^^ IfW^^J o HI. .l....t. ..Lt,»»l*.i.^ ilwllj o Hardwar* Markat Manufacturino 3 T3 D«v«lopinwit Davalopmant Oualty Softwnr* t Mariiat ProgrMIM Aaturanca Ottvalopmtnt NO -T»clwle«l Product 10 y«chanleal EEnglnaarlng EEngln*«rlng I*->t*it'r\ni

~~w 01 DIatflct Pra-Salaa Support o t >~~

~~Support KHcQvtif Uttrtct Sanrle* Uanagara Haadquartan Product Support EEducation ^^Tn, s-^^Hii' '~~

~~< O.~~

~~O 'i0m^^ o UJ •y-r H o o i*4iPM~~

~~(n - 30 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 (0 E (0 — CO 55® C Q. coo ©^ 0) o Q. CO O~~

~~- 31 - Dataquest Incorporated. A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Li:.:£x^ I ):;;:a Dataquest Dataquest~~

~~FLEXIBLE SYSTEMS ARCHITECTURE~~

~~Dennis Peck President and CEO Saber Technology Corporation~~

Mr. Peck is responsible for all aspects of Saber's operation including long-range corporate planning and development, and overseeing the company's professional growth. Mr. Peck has an extensive background in sales, marketing, and general management, including almost 20 years experience in the computer systems industry. Prior tio joining Saber in February, 1984, Mr. Peck served as general manager of NCR Corporation's $350 million medium and large-scale computer system business unit in San Diego. He was responsible for NCR's largest development and production business unit. Mr. Peck obtained his M.B.A. and B.A degrees from Bowling Green State University in Ohio.

~~Dataquest Incorporated WORKSTATION FOCUS CONFERENCE August 13 and 14, 1984 Burlington, Massachusetts~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 FLEXIBLE SYSTEM ARCHITECTURE WORKSTATIONS~~

~~DENNIS M. PECK~~

~~SABER TECHNOLOGY CORPORATION SAN JOSE, CALIFORNIA~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 FLEXIBLE SYSTEM ARCHITECTURE~~

~~I. INTRODUCTION~~

~~II. WHAT~~

~~III. WHY~~

~~A. END USER~~

~~B. SYSTEM SUPPLIER~~

~~C. MANUFACTURER~~

~~IV. HOW~~

~~A. APPLICATION PROCESSOR~~

~~B.GRAPHICS SYSTEM~~

~~C. MONITOR~~

~~D. IMAGE PROCESSOR~~

~~E. SOFTWARE~~

~~F. NETWORK~~

~~V. SUMMARY~~

~~- 1~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 I. COMPANY INTRODUCTION~~

~~A. FOUNDED 1982 TO BECOME A SUBSTANTIAL COMPANY BY~~

~~SUCCESSFULLY DEVELOPING AND MARKETING A FAMILY OF~~

~~NEXT GENERATION WORKSTATIONS.~~

~~B. RELEASED INITIAL PRODUCT AT NCGA IN MAY.~~

~~- 2 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o w S i3 C CD 3 O SABER TRIBUS ARCHITECTURE -o^ •D O~~

~~CD a. HARDWARE w c o- en APPLICATION BUS IMAGE BUS d. dj" I 1 ' I 1_ •3 o m > 32032 b GRAPHIC APPLICATION IMAGE ./ •^ PROCESSOR PROCESSOR PROCESSOR~~

~~o 3 •o T. X lu GRAPHICS BUS GRAPHIC >< DISPLAY to (O o LJ 3 5! I 4A- =/KEYBOARD^—--^ a. (D ETHERNET -1 T3 P> (10 MBS) vMB o ra~~

~~CO 5h DT • GRAPHICS PROCESSOR (D Ol • IMAGE PROCESSOR~~

~~.b. O • DISPLAY 00 • BUS STRUCTURE~~

~~(D o n c (D (0 SABER SOFTWARE~~

~~(D a USER~~

~~CO c CT CA QI APPLICATION SOFTWARE 0)' •< o_ > SHELL o UNIX UNIX O o 3 TOOLS KERNEL •o B) &~~

~~(O GKS OR O~~

~~Q. SIGGRAPH a. (D T) CORE (l>~~

~~WINDOW MANAGER CO~~

~~O U VIRTUAL DEVICE INTERFACE O> DEVICE DRIVERS~~

~~HARDWARE oCO o o~~

~~5. CD to CO~~

~~< O~~

~~ZD 2 O O Uj~~

~~< cc LL~~

~~< 2 cc m DQ < CO~~

~~o c w o :; O "_( nO fv%- o ^ o3- r>~~

~~wogu~~

~~• n. •< K~~

~~llj 3 .< a c O ". — I~~

~~•< 3~~

~~- 5~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 II. WHAT IS A FLEXIBLE SYSTEM ARCHITECTURE~~

~~A. A SYSTEM THAT MEETS A WIDE RANGE OF APPLICATION NEEDS~~

~~1. CAD ARCHITECTURAL~~

~~2. CAD MECHCANICAL~~

~~3. CAD ELECTRICAL~~

~~4. CAE A. PCB LAYOUT B. I.e. DESIGN C. LOGIC/CIRCUIT SIMULATION~~

~~B. A SYSTEM THAT CAN GROW IN PLACE~~

~~1. MIGRATION PATH ENGINEERING A. ADD PERIPHERALS, FEATURES, ACCELERATORS AS THE NEED ARISES~~

~~C. A SYSTEM THAT WILL HAVE A LONG USEFUL LIFE~~

~~- 6 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 III. WHY EMPLOY A FLEXIBLE SYSTEM ARCHITECTURE~~

~~A. END USER~~

~~1. SAME SYSTEM INTERFACE ACROSS APPLICATIONS A. LOWER TRAINING COSTS B. ECONOMY OF SCALE IN MAINTENANCE~~

~~2. COMPATIBLE SYSTEM TO MEET FUTURE NEEDS~~

~~3.LEVERAGED PURCHASES~~

~~4. LOWER COST OF OWNERSHIP~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 B. SYSTEM RESELLER~~

~~1. COMPATIBLE HARDWARE /SOFTWARE PLATFORM FOR MULTIPLE APPLICATIONS~~

~~A. ONE SOFTWARE PORT TO A FAMILY OF PRODUCTS~~

~~2. ECONOMIES OF SCALE~~

~~3. LOWER SUPPORT COST~~

~~4. MORE RESPONSIVE SUPPORT ORGANIZATION~~

~~5. LONGER PRODUCT LIFE CYCLE~~

~~6. IMPROVED MARGINS~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 C. SYSTEM MANUFACTURER~~

~~1. MEETS WIDE RANGE OF MARKET REQUIREMENTS~~

~~2. EXPERIENCE CURVE ECONOMIES~~

~~3. IMPROVED OVERHEAD MANAGEMENT~~

~~A. TEST ENGINEERING~~

~~B. MANUFACTURING ENGINEERING~~

~~C. QUALITY CONTROL~~

~~4. IMPROVED SUPPORT COST MANAGEMENT~~

~~5. ENHANCED PRODUCT LIFE CYCLE AND MARGINS~~

~~- 9 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-90001 Telex 171973 HOW~~

~~THE IMPLEMENTATION OF A FLEXIBLE SYSTEM ARCHITECTURE~~

~~- 10 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, GA 95131 / (408) 971-9000 / Telex 171973 LIFECYCLE PLANNING~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 i^ldder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o B>~~

~~1-^ 0) n c.r 0o) o •5 o~~

(D D. > GRAPHICS PROCESSOR Cfl c CT (A a 5" APPLICATION BUS IMAGE OUS •3 o c I 1 > b 32032 APPLICATION (fl (D r-Vt PROCESSOR o 3O •o 01 GRAPHIC DISPLAY rv) CO o (O Q. I Q. (D ug •o ETHERNET (10 MBS)

~~w a> •3~~

~~o • HIGH PERFORMANCE OP CODE/DATA BUS DESIGNED • SCALABLE HARWARE FONT PROCESSING > TO SUPPORT APPLICATION ACCELERATORS: • DIRECT ACCESS TO MASS STORAGE - MATRIX TRANSFORMATION~~

~~o•fc . - ARRAY PROCESSING 00 - HIGH SPEED FILL - TILE GENERATORS CoO o - MESH GENERATORS o~~

~~2. (D~~

~~CO A. EXPANDABLE~~

~~1. MAIN MEMORY 1-16 MB~~

~~2. ADDITIONAL DISC AND TAPE~~

~~3. COMMUNICATION - LANS, COAX, RS 232, SNA~~

~~4. FLEXIBLE I/O - TABLETS, PLOTTER, MICE~~

~~- 13~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 B. LONG LIFE CYCLE~~

~~1; 256 K MEMORY CHIPS~~

~~2. IMPROVED PERFORMANCE~~

~~A. DUAL PROCESSOR - INCREASED PERFORMANCE 1.8 B. CACHE C. UNIX ACCELERATORS~~

~~- 14 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 D. CHIP EVOLUTION~~

~~1,32132-12 (1985) INCREASED PERFORMANCE .2~~

~~2.32232-12 (1985) INCREASED PERFORMANCE .5~~

~~3. 32332 - 15 (1986) INCREASED PERFORMANCE 2.3~~

~~4. 32C432 - 15 (1987) INCREASED PERFORMANCE 4.7~~

~~- 15 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o BJ. 0) .O c (D~~

~~•D O~~

(D a. GRAPHICS PROCESSOR > w c cr w APPLICATION BUS IMAGE BUS a 5' JL T I i •< o_ > b 32032 I GRAPHIC APPLICATION IMAGE (D_ PROCESSOR PROCESSOR cn f^ PROCESSOR (D O O 3 •a L J 01 3 GRAPHICS BUS GRAPHIC DISPLAY I lO DT u o cn ro HIGH PERFORMANCE OP CODE/DATA BUS DESIGNED • SCALABLE HARWARE FONT PROCESSING O > TO SUPPORT APPLICATION ACCELERATORS: • DIRECT ACCESS TO MASS STORAGE CO CJl - MATRIX TRANSFORMATION - ARRAY PROCESSING j^ 3o - HIGH SPEED FILL CO - TILE GENERATORS

~~oCO - MESH GENERATORS o o~~

~~CO ~J A. EXPANDABLE~~

~~1. PLUG IN ACCELERATORS~~

~~2. FIVE SLOT CHASIS~~

~~3. MATRIX TRANSFORMATIONS 120 K MATRIX TRANSFORMATIONS PER SECOND~~

~~4. POLYGON LINE FILL~~

~~- 17 -~~

~~Dataquest incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 D~~

~~IMAGE PROCESSOR~~

~~o o •3 o APPLICATION BUS IMAGE BUS I 1 (D X 1 JCZ 1 , a.~~

~~CO c 32032 CJ GRAPHIC (A APPLICATION IMAGE 5! rAi PROCESSOR PROCESSOR PROCESSOR 0)' —1 •< o_ > X T p GRAPHICS BUS GRAPHIC z DISPLAY (D"~~

~~O o 3 •a p=f KEYBOARD^--^ B> ETHERNET (10 MBS) I~~

T) 0) • 2 MEGAPIXEL IMAGE PU^NES • 180MHz DAC WITH 8BIT RESOLUTION • 2 TO 24 PIXELS • HIGH SPEED VIDEO LOOKUP TABLES FOR CO D) COLOR CONTROL 3 • HIGH SPEED MEMORY ALLOWS MEMORY UPDATE C_ SIMULTANEOUSLY WITH SCREEN UPDATE • FULL LOGICAL OPERATION SUPPORT o (A LAYERED IN BLOCKS DOUBLE BUFFERING

~~.b o 00 o(O o o~~

~~(O •^ CJ B. EXTENDED LIFE CYCLE~~

~~1. ADDITIONAL ACCELERATORS~~

~~A. TILE GENERATORS~~

~~B. MESH GENERATORS~~

~~C. SIMULATION FUNCTIONS~~

~~2. LOWER COST~~

~~A. CUSTOM VLSI~~

~~- 19 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 A. IMAGE PROCESSOR EXPANDABILITY~~

~~1. 2 TO 24 M BYTE IMAGE PLANES~~

~~2. EXPANSION CHASSIS~~

~~3. HIGH SPEED DAC'S~~

~~4. 180 MHZ~~

~~- 20 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 B. EXPANDED LIFE CYCLE~~

~~A. HIGHER DENSITY MEMORY CHIPS~~

~~B. LSI/VLSI IMAGE CONTROL UNIT~~

~~C. IMPROVED BANDWIDTH~~

~~- 21 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o a n c CO o o •o3~~

~~(D DISPLAY OBJECTIVES CL~~

~~CO c CT (A 5: BJ'~~

~~> b~~

~~o o 3 "O 01 -3< CO o NJ Q. Q. CD ETHERNET —1 (10 MBS •D D)~~

~~W D) 3 • FLICKER FREE [NON INTERLACED) o > • ULTRA HIGH RESOLUTION (2M PIXELS) • BRIGHT (FOR OFFICE USE)~~

~~.b. • MINIMAL FOOTPRINT o 3 CO • DIN SPEC DESIGN~~

~~CoO o o~~

~~CO u A. DISPLAY FLEXIBILITY~~

~~1. MONOCHROME~~

~~2. COLOR~~

~~3. BANDWIDTH~~

~~4. RESOLUTION~~

~~5. EROGONOMICS~~

~~- 23 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 B. LIFECYCLE PLANNING~~

~~1. RESOLUTION REVOLUTION~~

~~2. IMAGE DYNAMICS~~

~~3. HIGHER BANDWIDTH~~

~~4. COST REDUCTION~~

~~- 24 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 SOFTWARE~~

~~UNIX (4.2)~~

~~ETHERNET DISTRIBUTED MASS STORAGE ENHANCEMENTS~~

~~SIGGRAPH CORE (2D & 3D) PRIMITIVES~~

~~GKS PRIMITIVES (2D) (3D)~~

~~"C", PASCAL, FORTRAN 77~~

~~PROGRAMMERS WORK BENCH~~

~~WINDOW MANAGER~~

~~- 25 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 A. SOFTWARE EXPANDIBILITY~~

~~1.ADHERANCE TO STANDARDS~~

~~A. 4.2 BSD B. SIGGRAPH CORE C. GKS CORE D. VDI 2. OPTOMIZE FOR PERFORMANCE A. OS AND LANGUAGES~~

~~- 26 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o 0)~~

~~^ o o -1 •o o SABER MAINFRAME COMMUNICATIONS~~

~~(D a. > CO VAX-11/7XX I ° > o (A (D o o 3 •D fi> 3 IDI^ SC-5000 1 CMJtlIfMft CLUSTEft N) MAlNFnAUE § coHTROLLenf Q. a. cs> 3.5 Wb/SEC •o COAX a>~~

~~w B) 3 o > CO~~

~~.b o SS (O to o o o~~

~~2. a>~~

~~(O w-J B. LIFECYCLE PLANNING~~

~~I.UNIX STANDARDS~~

~~2. WINDOWS~~

~~3. FILE~~

~~4. NETWORKING~~

~~- 28 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 A. NETWORKING EXPANDIBILUY~~

~~1.RS 232~~

~~2. LANS~~

~~-3. HIGH SPEED LINKS~~

~~- 29 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Pari< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 B. LIFE CYCLE PLANNING~~

~~INTEGRATED SYSTEMS~~

~~2. GATEWAY~~

~~31 SERVER ORIENTED NETWORK~~

~~- 30 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 SUMMARY~~

~~. FLEXIBLE ARCHITECTURE OFFER ECONOMIES TO:~~

~~- END USER - SYSTEM HOUSE - OEM SUPPLIER~~

~~. LONGER LIFE CYCLE~~

~~- 31 -~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 WHEN EVALUATING WORK STATION FROM A FLEXIBLE SYSTEM ARCHITECTURE

~~- PROCESSOR - GRAPHICS - IMAGE PROCESSOR - MONITOR - BASIC SOFTWARE AND TOOLS - NETWORKING~~

~~- 32 -~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 :>aDataqijest PERFORMANCE ENHANCEMENT VIA APPLICATION-SPECIFIC HARDWARE Arthur J. Collraeyer, WEITEK

Background Advances In Integrated circuit technology, coupled with advances In convnunication/ networking technology have altered significantly the economics of computing. The mIcroprocessor-based workstation has replaced the time-sharing terminal, giving the user the opportunity to interact with a dedicated computing resource, which in turn communicates efficiently with other computers In a networked environment. Apart frtMii Its potential for truly Interactive computing, distributed computing is today an economically viable and often compelling alternative to centralized computing.

To maintain economic viability across a broad range of applications, distributed computing incorporates the concept of servers, which are shared resources, accessible by users via the network. Examples are file servers, peripheral servers (e.g., plotters, printers, etc) and compute servers. Compute servers enable convenient access to more powerful computing facilities, for background processing or computationally-intensive tasks. Problem; Computationally-Intensive Applications The number of applications which stand to benefit from user interaction with a (dedicated) computer is great indeed. Many of these Involve computationally- Intensive tasks, tasks which place extreme demands on the computing resource. Such applications (e.g., CAE, CAO) depend on effective integration of servers into workstation networks. Moreover, In these applications, the capability of the server tends to limit the proliferation of workstations; as servers grow in power, the number of workstations capable of being supported increases. Computationally-intensive tasks can be divided Into two classes: numeric and non-numeric Where arithmetic operations (multiply, divide, square root, etc) predominate or real (floating point) numbers are Involved, the task is described as numeric. Where sorting, searching, decision making (pointer arithmetic, compares, etc) predominate, the task is described as non-numeric Examples of computationally-Intensive tasks - with their descriptors - are:

> Circuit Simulation (numeric) > Logic Simulation (non-numeric) > Placement/Routing (hon-numeric) > Design Rule Checking (non-numeric) > Solid Modeling (numeric) > Structural Analysis (numeric) - 1 - Dataquest incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Solution; Application Specific Hardware To cope with the demands of computationally-intensive tasks, a number of options are available to system designers. The search for efficient (sometimes called "clever") algorithms is the most popular and can yield dramatic results. Novel hardware, optimized for the task, is another popular remedy. The combination of the two can be particularly effective. Indeed, It can be so effective as to enable the construction of extremely compact, practically useful hardware assists. Such compact hardware assists, which derive their power from the workstation and their input/output signals from its backplane, are known as accelerators. While they may or may not be dedicated to their workstation hosts, they rarely serve more than one user at a time.

Where the task, the algorithm, or the technology preclude a compact hardware assist, the economics of scale favor the construction of a high-performance, application-specific system capable of serving multiple users simultaneously. Such systems are referred to as application-specific servers. Cost Effectiveness To measure the cost effectiveness of application-specific hardware, one requires a unit Of throughput. In a particular non-numeric application; namely, logic simulation, this unit Is the event per second. Precise as it sounds, this unit requires considerable qualification, especially if one Intends to make benchmark comparisons. Undaunted, I have attempted to extract from the literature of several vendors, the salient parameters of their logic simulation-specific hardware. Dividing the throughput by the cost of the equipment, servers would appear to be 100 to 1000 times as cost effective as workstations - or, for that matter, as the popular supermini. A recently developed accelerator, burdened with the cost of a workstation host is 100 times as cost effective, and quite adequate to support a single user-designer. Both products have been Implemented without the benefit of custom LSI. Clearly, the potential of application-specific hardware in non-numeric applications is dramatic. In numeric applications, such as circuit simulation, the cost effectiveness can often be approximated by the cost per MFLOP (million floating point operations per second). It is claimed that a popular array processor can execute SPICE five times faster than the popular supermini with floating point accelerator. Inasmuch as these two machines differ in their MFLOP rating by a factor of 40, one must be careful in correlating throughput with MFLOP's. Still, it Is a valid indicator. MFLOP's have long been precious. Until recently, the cheapest MFLOP in the world cost $20,000 ready-to-serve. Today, high-performance VLSI devices are coming into the market with the potential to drastically alter the economics in numeric applications. With these devices, accelerators of 10 MFLOP's and servers of 1000 MFLOP's can be constructed to assist applications such as circuit simulation, at a fraction of the cost of today's computing resources. Arithmetic bottlenecks will continue to be more costly to remove than non-numeric bottlenecks (after all, a MFLOP is the most costly MIP), but complexity of the respective solutions is converging rapidly.

- 2 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Out! ool( In surveying the prospects for application-specific hardware, one must examine the driving forces, technological and commercial. It goes without saying that the increasing availability of highly-integrated, high-performance digital circuitry is a major driving force, with the potential to dominate. Today, however, the prime technological enabler is the application insight developed and honed by the turnkey system cwnpanies. Driven by the immense market opportunity, competition among these companies is fierce. To grow their businesses, these suppliers must (1) justify their systems against the broadest possible ensemble of application- related tasks, and (2) position themselves to capture as much of the customer's budget for computing as possible.

Against this background, the haze begins to melt. Clearly, there are many unresolved computational bottlenecks. With increasing availability of VLSI to resolve these bottlenecks, and the Increasingly stiff penalties assessed against those who fail to respond to the needs of the marketplace, it is likely that application-specific hardware will proliferate. (Who will resist? Certainly not the mainframe makers.) Moreover, VLSI and competition will force marketplace pricing. Finally, the accelerator - favored by existing (workstation) companies - will present formidable competition to the server.

Summary The technological evolution which has made it possible to distribute computing to the point of need will continue to undermine the role of the general-purpose computational server, as it enables powerful 32-bit microprocessors to broaden the application of workstations, and powerful application-specific hardware to augment the capabilities of workstations in the computationally-intensive applications. Application-specific hardware, in particular, cuts two ways. First, by offering substantially greater economy than general-purpose computers, application-specific hardware is bound to erode the mainframe share of the engineering computing budget. Second, by covering effectively for the limitations of workstations in computationally-intensive applications, application-specific hardware will facilitate the growth of distributed computing.

~~- 3 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ao~~

~~EVOLUTION OF DISTRIBUTED COMPUTING~~

~~CO c CT n d S" :SERVER; •3 > b PROCESSOR PROCESSOR ;SERVER; ffi (A (D 3 o o 3 X3 0)~~

8 I a5 ^ £• • -^ T3 B—%) 3C o I • •I Tllllim»aKii ilKii.miiiM»TTiTnnrMWllllimiXXi » oI > QUEUE TERMINCIIEIIIIIALS I WORKSTATIONS o.(^ BATCH DISTRIBUTED 00 PROCESSING TIMESHARING COMPUTING o » B n c a 5»" o o -1 •ao CLASSES OF SERVERS

~~U) c CJ <» a. S •5 o^ FILE SERVERS GATEWAYS > p • FILE • LAN-LAN z 5' SHARING O o •LAN-PBX 3 •o • ARCHIVING D)~~

~~S I 3J Ul CI ( ^ CD X~~

~~TJ A) I/O SERVERS /COMPUTE 3- r^^ o •Rs y »UNIT RECORD /SERVERS CO • GENERAi1/L > g> 3 »HARD COPY /PURPOSE, , o > APPLICATION SPECIFIC/ A y >^ > / O 00~~

~~2. (K~~

~~CO -^l u 3 O o •oo~~

~~ID TWO TYPES OF COMPUTE SERVERS a. > (n 7R a B>' -5 o_ > Y777A o HIGH APPLICATION- / SPECIFIC/ o o 3 /SERVERS^ •o (B TRANSACTIONS g I 3 (Ti $ k I I I Q GENERAL~~

~~U) LOW PURPOSE 0) SERVERS o >~~

~~o 00 LOW HIGH BREADTH OF APPLICATION~~

~~(O o u a J3 c (D a> S o o •3 COMPUTATIONALLY INTENSIVE APPLICATIONS o~~

~~> CO c CT (0 Q! u° -3 o^ > p z 5"~~

~~O NUMERIC NON-NUMERIC o 3 •o 01 •3< CIRCUIT SIMULATION • LOGIC SIMULATION~~

~~KJ S I SOLID MODELING • PLACEMENT/ROUTING Q. S^ > • DESIGN RULE CHECKING TJ STRUCTURAL ANALYSIS lU s-,r O IMAGE PROCESSING... • LAYOUT VERIFICATION...~~

~~^ I o >~~

~~to u-J o~~

~~o o 1 THE SERVER/ACCELERATOR DICHOTOMY~~

~~CO c CT U g.' a •3 > b APPLICATION HIGH SPECIFIC o o SERVERS 3 •o u f INTEGRAL~~

~~M I ADD-IN g ' 3! 00 APPLICATION- \^ ADD-ON a TRANSACTIONS & • ^ SPECIFIC CD~~

~~TJ ACCELERATORS a> jr O~~

~~W Bl C_ O (A (D o LOW WORKSTATION >~~

~~55. a> X~~

~~(O -J CO I3 O •o3 o ECONOMICS OF APPLICATION-SPECIFIC SERVERS > CO (A NON-NUIVIERIC APPLICATION)~~

~~I 10,000 100 /!^^^ a WORKSTATION W/O ACCELERATOR o 3 •o 0) 1.000 —SERVERS 10 •3<~~

~~"J . S ' COST WORKSTATION 3D *0 a. ($000) W/ACCELERATOR EPS' -o B> a- 100 o~~

~~CO 0> 3~~

~~o > 10 ISL .1 * EVENT PER SECOND~~

~~to u-J o s> i c (D SW o o •D O SERVERS VS WORKSTATIONS > O) (A NON-NUMERIC APPLICATION)~~

~~Q. a o^ > p Z~~

~~a O / o 3 y •o 0) FIGURE OF MERIT UNIT 58 ' $ $/EPS ° I-i 5o Q. CL i WORKSTATION 1.0 150 TJ B>~~

~~W/ACCELERATOR 1,5 2 I § o > SERVER($300K) 3.5 1 5/~~

~~(O -.1 tti I o o -^ •o o NON-NUMERIC SERVERS/ACCELERATORS > (EXAMPLE: LOGIC SIMULATION)~~

~~a. MEPS u •< o^ > p 100 z MCREAW $400,000 » fA ID f^'c"5s"V**«^ $2,000,000^ 3 10 o 3 •o SERVERS S) •< I 1 I— J1 MEPS~~

~~X i-- $125,000 I W/ACCELERATOR~~

~~i .01 - WORKSTATIONS a $85,000 I .001 L^ I 5 .001 .01 .1 10 100 • EPS/$ * SERVER COMPUTATIONS DERIVED FROM SERVER PRICE; ACCELERATOR COMPUTATION INCLUDES PRICE OF WORKSTATION~~

~~(D o~~

~~o o -1 •oo COMPUTATIONALLY INTENSIVE APPLICATIONS cU) atj d. o > p z it (A CD o3 o NUMERIC 3 •D 0) • CIRCUIT SIMULATION~~

~~2 to a. • SOLID MODELING Q. I • STRUCTURAL ANALYSIS • IMAGE PROCESSING... 0)~~

~~O (It~~

~~O 00~~

~~it (D~~

~~(O O} o 01~~

~~3 O O •3 o ECONOMICS OF APPLICATION-SPECIFIC SERVERS (D a > (NUMERIC APPLICATIONS) in~~

~~10,000 1,000 I WORKSTATION a SUPER (D COMPUTER SUPERMINI o o 3 •o 1,000 100 B) ARRAY 3 PROCESSOR COST *COST MFLOPS 3 oj ($000) ($000) n. • X 01 jr 100 10 o~~

~~0) 0> 3 o > 10 • 64-BIT OPERATONS~~

~~to -t til o u .Q C (D~~

~~3 O o •3 o SERVERS VS WORKSTATIONS > CO (NUMERIC APPLICATIONS)~~

~~Q. S' o^ > b~~

CD o o 3 •a X FIGURE OF IMERIT y UNIT $ $/MFLOPS i- 2 Ji" a. (B ' WORKSTATION 1 15 "^ T3 ai jr O ACCELERATOR o • (0 m 01 3 o ARRAY > PROCESSOR 10 1 5/ o OD o u » J3 C CO NUMERIC SERVERS/ACCELERATORS c tr CO S. (EXAMPLE: CIRCUIT SIMULATION) D> -3 o^ > b • MFLOPS

o 1,000 SUPERCOMPUTER o 3 $ 1,000,000 •uo 3 100 ARRAY i3 I PROCESSOR/ $59,500 1- 5X VAX W/ FPA °Q-. I 10 <0 ' $350,000 TJ ID COMING 1 — WTL 1064 0) SOON 01 3 C $39,500 o .1 — MATH CHIP o(A > WORKSTATION • .01 J I \ L i o .1 1 10 100 1,000 10.000 00 • FLOPS/$ (£> * 64-BiT OPERATIONS oO o f o o -1 •a o NUMERIC SERVERS/ACCELERATORS (S a w (EXAMPLE: KINEMATICS; SOLID MODELING) c CT a d * MFLOPS S" -3 > h 100 $59,500 o o $72,500 3 ^ 20 MFLOPS •o B> WORKSTATION 3 10 w/ACCELERATOR/ COMING iS I WTL 1032 S M 5 cr> SOON S. a. a I -J TJ a

~~.1 MATH CHIP (ft $39,500 B> 3 WORKSTATION o I i > .01 1 10 100 1,000 10,000 • FLOPS/$ * 32-BIT OPERATIONS~~

~~(D -J U f o o -1 TO o IMPACT OF APPLICATION SPECIFIC HARDWARE~~

> C/) c cr NON-NUMERIC APPLICATIONS (fl a S' -3 o • IMPACT DRAMATIC, EVEN NOW! > o • 100X TO 1000X IMPROVEMENT IN TRANSACTIONS (B o PER DOLLAR. 3 •a lu 3 « SERVERS SUPPORT LARGE USER NETWORKS M I 1- OF LOW COST WORKSTATIONS. d • 10X TO 100X IMPROVEMENT IN TRANSACTIONS PER DOLLAR.

~~cc Ol 0>~~

~~o o -1 •oo ENABLING TECHNOLOGIES~~

~~(/> c CT 01 d S' >< o^ > b PROLIFERATION OF o o APPLICATION 3 •o D) SPECIFIC lO I 1- HARDWARE 2 CXI Q. °-CD I' T) B> HIGH S EFFICIENT i ALGORITHMS PERFORMANCE VLSI~~

~~4>. O OD~~

~~(D o u DT .Q c (V 3 o o •oa OUTLOOK w c cr (A Q! »' -3 o > o~~

~~(D PROLIFERATION OF o o 3 •o APPLICATION B) SPECIFIC to I HARDWARE 13 ^-o a. S^ I TV i 7z^ 1 MANY f BOTTLENECKS CAE/CAD MAINFRAME COMPETITION RESISTANCE 1 UNRESOLVED 5 g~~

~~.tw O 00~~

~~2. (D~~

~~! ^~~

~~o o -1 •oo SUMMARY > (n~~

~~Q. D>' •3 o^ > b~~

~~3 • BY OFFERING SUBSTANTIALLY GREATER ECONOMY o o 3 •o Dl THAN GENERAL PURPOSE COMPUTERS, •3< ° ^o 5 o a APPLICATION SPECIFIC HARDWARE WILL SURELY LIMIT 9^ I T) THE MAINFRAME SHARE~~

~~M 0> C OF THE ENGINEERING COMPUTING BUDOET. o a> a> >O~~

~~(B (P~~

~~(O Ol f I SUMMARY w c cr CO a u' •5 o^ > p z MOREOVER, APPLICATION SPECIFIC HARDWARE « w CD o3 o WILL EFFECTIVELY MASK 3 •D D> •3<~~

~~K3 ' THE LIMITED COMPUTATIONAL ABILITIES O to 3J M a. a I (D "^ OF WORKSTATIONS. TJ 0) HENCE, APPLICATION SPECIFrc HARDWARE w B> 3 WILL FACILITATE GROWTH OF DfSTRmJTED COMPUTING. o >~~

~~•"J (0 (::>aDataquest FOCUS CONFERENCE Electronic Design Automation: The Technology Renaissance~~

~~Sponsored by the CAD/CAM Industry Service~~

~~May 10 & 11, 1984 Sunnyvale Hilton Inn Sunnyvale, California~~

~~^^;ii>aDataquest CONFERENCE PROGRAM~~

~~THURSDAY, MAY 10 FRIDAY, MAY 11~~

7:30 am to 8:45 a.m. 8:15 a.m. Conference Registration Silicon Compilers are a 1984 Technology , 8:45 a.m. Phil Kaufman, President .A Welcome and Silicon Compilers, Inc. Conference Introduction 9:00 a.m. James Newcomb, Director CAD/CAM Industry Service EDA is for PCB CAD Too Dataquest Incorporated Bruce Holland, President ")( Cadnetix, Inc. ' 9:00 a.m. EDA Industry Overview 9:45 a.m. BetIi TucI EDA Design Center Strategies Daisy Systems Corporation ' Del ManIc, Vice PresIdent, IUIarItetIng and Sales^ California Devices, Inc. 12:15 p.m. Conference Wrap-Up 12:15 p.m. to 1:15 p.m. Beth Tucker, Industry Research Manager Lunch CAD/CAM Industry Service Dataquest Incorporated 1:30 p.m. 12:30 p.m. to 1:30 p.m. •Microprocessor Revolution Leads to the ASIC Evolution Buffet Lunch Kenneth IMcKenzIe, AssocIate Director X Semiconductor Industry ServIce Dataquest Incorporated

~~2:15 p.m. Coffee Break~~

~~2:30 p.m. Users' Points of View— Panel Discussion • Expectations • Needs • Futures • Applications~~

~~4:30 p.m. to 6:00 p.m. Cocktails and Hers d'Oeuvres CONFERENCE INFORMATION~~

HOTEL ACCOMMODATIONS Conference attendees are responsible for making their own hotel reservations. For your convenience, the Sunnyvale Hilton Inn is holding a block of rooms for those attending the conference. To make reservations, call the Hilton directly at (408) 738-4888 and mention that you are attending the DATAQUEST Focus Conference.

TRANSPORTATION The Sunnyvale Hilton Inn is located 6 miles north of San Jose Airport and 30 miles south of San Francisco International Airport. A free shuttle service is available from San Jose Airport to the hotel. A direct telephone line to the hotel is available at the airport to arrange for transportation. Taxi and limousine services are also available at the airports. DATAQUEST CALLS IT EDA THE NEED Electronic design automation (EDA) is the CAD/CAM industry's youngest, most explosive, most dynamic segment. EDA products and systems address a broad spectrum of phases within the entire electronic product design cycle. DATAQUEST calls this segment EDA because it encompases more than just computer-aided engineering (CAE), or applications related only to electrical engineers. In addition to logic or circuit design and simulation, EDA may also include: • Full-custom layout • Semicustom layout • Printed circuit board layout • Documentation • Test pattern analysis CAE products also address the needs of mechanical, civil, and architectural engineers. We believe that the EDA segment is significant enough to warrant its own name—ELECTRONIC DESIGN AUTOMATION.

THE FOCUS CONFERENCE The Technology Renaissance Conference is being sponsored by the DATAQUEST CAD/CAM Industry Service in direct response to the growth of a major market segment of the CAD/CAM industry—Electronic Design Automation (EDA). The conference will explore the issues that affect EDA from the vendor's, user's, and investor's points of view. Focus Conference highlights are: • EDA industry dynamics • EDA market size and forecasts • Key vendors' roles • Users' applications and expectations • Hardware and software issues • Product hfe cycles • Design cycles • Futures WHO SHOULD ATTEND Everyone should attend the Technology Renaissance Focus Conference who needs to know and understand the EDA industry and how it is changing the way electronic products are designed. This includes: • Management • Engineers and Product Developers • CAD System Managers • CAD/CAM Marketing Professionals • Workstation Marketing Professionals • Computer Marketing Professionals • Venture Capitalists • Investment Portfolio Managers • Strategic Planners DATAQUEST DATAQUEST provides strategic and tactical information on high-technology industries. DATAQUEST's authoritative timely data and analyses are offered in a four-part format: • Industry-specific data bases • Frequent, concise newsletters on industry developments • Direct access to our research staff of industry experts • The opportunity to meet with industry peers and experts at annual conferences In addition, DATAQUEST offers: • Custom consulting for clients who require speciaUzed information • Financial Services Programs Founded in 1971 with one research group, DATAQUEST has grown to 14 high- technology services and looks forward to continuing expansion. More than 1,500 U.S. and international clients rely on DATAQUEST for critical information on high-technology industries. Major service areas include computers, computer storage, copying and dupUcating, display terminals, electronic printers, graphics terminals, office automation, robotics and manufacturing automation, semi conductors, telecommunications, and CAD/CAM.

CADICAM INDUSTRY SERVICE The CAD/CAM Industry Service, founded in 1981, has a staff of industry pro fessionals devoted to full-time research and analysis of the CAD/CAM industry. The research staff analyzes and reports on all major activities occuring within the CAD/CAM industry, including: • Workstation technology and appUcations • Graphics • Dedicated hardware • Electronic design automation • Integrated circuit CAD/CAM • Printed circuit board CAD/CAM • Mechanical CAD/CAM • Architecture, engineering, and construction CAD • AppUcations software • Personal computer-based CAD/CAM systems CONFERENCE INFORMATION

LOCATION The Sunnyvale Hilton Inn, located in the heart of CaUfornia's Sihcon Valley, is the site of the CAD/CAM Focus Conference, Electronic Product Design: Tlie Tectinology Renaissance. The conference will be held in the Chardonnay Room.

~~Address: Sunnyvale Hilton Inn 1250 Lakeside Drive Sunnyvale, California 94086 Telephone: (408) 738-4888~~

CONFERENCE FEE The fee for the Electronic Product Design Focus Conference is: • Each attendee who is a client of any DATAQUEST industry service $325 • Each attendee who is not a client of any DATAQUEST industry service $395 The conference fee includes a conference binder with copies of the material presented at the conference, and lunches. Conference tees must be paid in advance. Checks, money orders, purchase order numbers, American Express, Mastercard, or Visa will be accepted.

REGISTRATION To register, mail the reservation form on the tear-off card, or telephone Ms. Pam Shook at (408) 971-9000, extension 316, or Telex 171973. Space is limited, so be sure to register early.

CONFERENCE SCHEDULE Conference registration will take place starting at 7:30 a.m. on Thursday, May 10. Thursday's program will begin at 8:45 a.m. and will conclude with a DATAQUEST-hosted cocktail party from 4:30 to 6:00 p.m. On May 11, the program will start at 8:15 a.m. and conclude after a buffet luncheon.

Last Name First . IVI.I. Conference Fee: D $325 DATAQUEST client attendee Title • $395 individual or nonclient company attendee Company. Form of Payment: D Check enclosed D Money Order enclosed Street n Purchase Order*. City . State -Zip. D Am. Express* Exp. date _ n Visa* Exp. date _ Telephone ( ) D Mastercard*. Exp. date_ a Please send me information on DATAQUEST's CAD/CAM Industry Service. Signature NO POSTAGE NECESSARY IF MAILED IN THE UNITED STATES Business Reply Mail FIRST CLASS PERMIT NO. 7279 SAN JOSE. CA

~~POSTAGE WILL BE PAID BY ADDRESSEE: Attention: Pam Shook Dataquest Incorporated 1290 RIdder Park Drive San Jose, CA 95131-9980 FOCUS CONFERENCE~~

~~Electronic Design Automations The Technology Renaissance~~

~~sponsored by the CAD/CAM Industry Service~~

~~May 10 & 11, 1984 Sunnyvale Hilton Inn Sunnyvale, California~~

~~n)::»aE>ataquest CONFERENCE PROGRAM~~

~~THURSDAY, MAY 10 FRIDAY, MAY 11~~

7:30 a.m. to 8:45 a.m. 8:15 a.m. Conference Registration Silicon Compilers are a 1984 Technology 8:45 a.m. PIiiI Kaufman, President Welcome and Silicon Compilers, Inc. Conference Introduction 9:00 a.m. James Newcomb, Director CAD/CAIM Industry Service EDA is for PCB CAD Too Dataquest Incorporated Bruce Holland, President Cadnetix, Inc. 9:00 a.m. EDA Industry Overview 9:45 a.m. BetIi TucIcer, Industry ResearcIi IManager Coffee Break CAO/CAM Industry ServIce Dataquest Incorporated 10:00 a.m. 9:45 a.m. Evaluating Business Plans, Going Public Wliat's All The Excitement About— Grant HeidricIi How Can An Industry Grow So Fast? GeneraI Partner Thomas Bruggere, President MayfIeId Fund Mentor GrapIiics 10:45 a.m. 10:30 a m. EDA: Support is the Big Issue Coffee Break TIiomas BInder, Vice President IMarlteting and Sales 10:45 a.m. Silvar-Lisco Application-Specific Hardware Jared Anderson, President 11:30 a.m. VaIJd Logic Systems, Inc. Coping with a Changing Industry and Changing Technologies 11:30 a.m. Harvey Jones, Senior Vice President EDA Design Center Strategies Daisy Systems Corporation DeI IUanI(, VIce President, IVIarketIng and SaIes CaIIfornIa DevIces, Inc. 12:15 p.m. Conference Wrap-Up 12:15 p.m. to 1:15 p.m. Belli Tucker, Industry ResearcIi Manager Lunch CAD/CAM Industry Service Dataquest Incorporated 1:30 p.m. 12:30 p.m. to 1:30 p.m. Microprocessor Revolution Leads to the ASIC Evolution Buffet Lunch KennetIi IMcKenzIe, AssocIate Director Semiconductor Industry ServIce Dataquest Incorporated

~~2:15 p.m. Coffee Break~~

~~2:30 p.m. Users' Points of View- Panel Discussion * ExpectatIons * Needs * Futures * AppIIcations~~

~~4:30 p.m. to 6:00 p.m. Cocktails and Hers d'Oeuvres CONFERENCE INFORMATION~~

f DATAQUEST CALLS IT EDA THE NEED Electronic design automation (EDA) is the CAD/CAM industry's youngest, most explosive, most dynamic segment. EDA products and systems address a broad spectrum of phases within the entire electronic product design cycle. DATAQUEST calls this segment EDA because it encompases more than just computer-aided engineering (CAE), or applications related only to electrical engineers. In addition to logic or circuit design and simulation, EDA may also include: • Full-custom layout • Semicustom layout • Printed circuit board layout • Documentation • Test pattern analysis CAE products also address the needs of mechanical, civil, and architectural engineers. We believe that the EDA segment is significant enough to warrant its own name—ELECTRONIC DESIGN AUTOMATION.

WHO SHOULD ATTEND Everyone should attend the Technology Renaissance Focus Conference who needs to know and understand the EDA industry and how it is changing the way electronic products are designed. This includes: • Management • Engineers and Product Developers • CAD System Managers • CAD/CAM Marketing Professionals • Workstation Marketing Professionals • Computer Marketing Professionals • Venture Capitalists • Investment Portfolio Managers • Strategic Planners CONFERENCE INFORMATION

LOCATION The Sunnyvale Hilton Inn, located in the heart of California's Silicon Valley, is the site of the CAD/CAM Focus Conference, Electronic Product Design: The Technology Renaissance. The conference will be held in the Chardonnay Room.

~~Address: Sunnyvale Hilton Inn 1250 Lakeside Drive Sunnyvale, California 94086 Telephone: (408) 738-4888~~

CONFERENCE FEE The fee for the Electronic Product Design Focus Conference is: • Each attendee who is a client of any DATAQUEST industry service $325 • Each attendee who is not a client of any DATAQUEST industry service $395 The conference fee includes a conference binder with copies of the material presented at the conference, and lunches. Conference fees must be paid in advance. Checks, money orders, purchase order numbers, American Express, Mastercard, or Visa will be accepted.

CANCELLATION POLICY All cancellations received by DATAQUEST after April 26, 1984, are subject to a $100 service charge unless the registrant sends a replacement. Registrants who do not cancel and who do not attend the conference will automatically be assessed a $100 service charge. Notice of cancellation should be made to Ms. Pam Shook, (408) 971-9000, extension 316, or Telex 171973. DATAQUEST DATAQUEST provides strategic and tactical information on high-technology industries. DATAQUEST's authoritative timely data and analyses are offered in a four-part format: • Industry-specific data bases • Frequent, concise newsletters on industry developments • Direct access to our research staff of industry experts • The opportunity to meet with industry peers and experts at annual conferences In addition, DATAQUEST offers: • Custom consulting for clients who require specialized information • Financial Services Programs Founded in 1971 with one research group, DATAQUEST has grown to 14 high- technology services and looks forward to continuing expansion. More than 1,500 U.S. and international clients rely on DATAQUEST for critical information on high-technology industries. Major service areas include computers, computer storage, copying and duplicating, display terminals, electronic printers, graphics terminals, office automation, robotics and manufacturing automation, semi conductors, telecommunications, and CAD/CAM.

CADICAM INDUSTRY SERVICE The CAD/CAM Industry Service, founded in 1981, has a staff of industry pro fessionals devoted to full-time research and analysis of the CAD/CAM industry. The research staff analyzes and reports on all major activities occuring within the CAD/CAM industry, including: • Workstation technology and appUcations • Graphics • Dedicated hardware • Electronic design automation • Integrated circuit CAD/CAM • Printed circuit board CAD/CAM • Mechanical CAD/CAM • Architecture, engineering, and construction CAD • Applications software • Personal computer-based CAD/CAM systems CX:^^^T)(::)(\ Dataquest 1290 Ridder Park Drive Reply Card San Jose, California 95131 (408) 971-9000 Telex 171973

~~n Please register me for the 1984 Conference. Electronic Design Automation: The Technology Renaissance (Please type, print or attach business card.)~~

Last Name First M.I Conference Fee: • $325 DATAQUEST client attendee Title n $395 individual or nonclient company attendee Company. Form of Payment: Street . D Check enclosed D Money Order enclosed D Purchase Order# City . State .Zip. • Am. Express* Exp. date _ • Visa# . Exp. date_ Telephone ( ) D Mastercard*. Exp. date. D Please send me information on DATAOUEST's CAD/CAM Industry Service. Signature NO POSTAGE NECESSARY IF MAILED IW THE UNITED STATES Business Reply Mail FIRST CLASS PERMIT NO. 7279 SAN JOSE, CA

~~POSTAGE WILL BE PAID BY ADDRESSEE: Attention: Pam Shook Dataquest Incorporated 1290 Rjdder Park Drive San Jose, CA 95131-9980 CANCELLATION/NON-RENEWALICHANGE NOTICE~~

~~Date: Bf .Service Group IVIarketing Manager I O Cancel Invoice Number~~

~~D Non-Renewal (invoice Number last issued~~

~~D Change Invoice Number~~

~~CLIENT NUIVIBER ^~~

~~Name: '* ^ i\'' O V ^~~

~~Company:~~

~~Address:~~

~~Telephone:~~

~~CANCELLATIONSlNON-RENiWALS:~~

~~Amount:~~

~~Term Dates:~~

~~Subscription Type:~~

~~Cancel Date:~~

~~Reason:~~

~~CHANGE FROM: CHANGE TO:~~

~~Price:~~

~~Term Dates:~~

~~Subscription Type:~~

~~Commission:~~

~~Installments:~~

~~Comments: Va^^ "^ost Do. Do not. print invoice~~

~~Originated by:~~

~~Completed by: ;-.i¥5-^.v:.:-i'^-t!'^^>-jS".-,^'-:iy,f~~

KE:^S:^©\Dataquest 1290 mdd«r Park DrIvt Invoice Pag* San JosB, CA 95131 ^ Ottaquwl Incorporatad ©377 ^ (408) 97V90O0 . j;. .ur^ INVOICE NO. A Subsidiary of A.C. Nielsen Company :.;^- ' • ..>-1j

BILL TO: 8HIPTO:(s•Ineas"BIIITo!'unI6S8ottl•^wi8eBt1ow^) v 94O08e^SCTLVDOOO •; J^f^>;>*i!!:.>7;-îiiî^'ii;',fr.i,Ji. . WlttLIAM ESCUDERO '••••'''.' . •'"•-,.'• .i'iW.'-^.Hv '.: ^ >';;:••!••'^;.!''iA''''liiy^'f;'' •;t£LEVIDEO "' • • ••.iS!.-j!S^T;if|t;i;^-i^3r^vit! 'di'-- - •;fIi4''^?5Ji'{j.>: ''.'\i-: n'vtjff 1170 MORSE AV£ ' • ••• r'i»;i'*^ tiiif,itvSj|>o'..:./r 'hr- !j iÂ J-RiJj^rfiVj :»,';jf»^'V^^ SUNNYVALE CA 940B8

~~^.-7^ Vi"^ - •: .•-• pcie ANNUAUt.CONFeRENCr 0S/»3/B4iiH ^f^"^ 383^00 0, 00 323. 00~~

~~i-V'Vi tisBt!'!"^;:^ • •-'••: 1^ tr • K» PLEASE REHiT: TOJ .D6PT #1703 . PO BOX dlOO ;-^'»^i-.v- ••' . •' SAN FRANCISCO CA 94Jdl u.. liS'jJv; \ .V. •• •, ^KH.~~

~~, ;'•' i 1~~

~~•f; "'^f't'i'^^'^'^^'.i'^^'-'-. -f^^ii' f?;toii "••!•• ' t . "\ ' '^••XV~~

~~••; ;•;• • JIB..'!^'' '**' *W tJiri»s?';- .-irg^^/. •'*f>'^ -r'VTi "• •i^ ,• ,3V;.l-i. . ,-.:^iri - '.'j''(i '• PIeaft9M«»"Qf>QCk~~

~~%EASE RETURN THE GREEN COPY WITH YOUR REMITTANCE. THANK YOa Focus Conference Electronic Design Automation: The Technology Renaissance~~

~~Sponsored by the CAD/CAM Industry Service May 10 and 11, 1984 Sunnyvale Hilton Inn Sunnyvale, California~~

~~CESs^gu Dataquest~~

~~1290 Ridder Park Drive San Jose, California 95131 (408) 971-9000 Telex: 171973~~

~~Sales/Service offices:~~

FRANCE JAPAN DATAQUEST SARL DATAQUEST Japan, Ltd. 41, rue Ybry Azabu Heights, Suite 410 92522 Neuilly-sur-Seine Cedex l-5-lO, Roppongi, Minato-ku France Tokyo 106, Japan (01) 758-1240 (03) 582-1441 Telex: 630842 Telex: J32768

The content of this report represents our interpretation and analysis of infor mation generally available to the public or released by responsible individuals in the subject companies, but is not guaranteed as to accuracy or completeness. It does not contain material provided to us in confidence by our clients. This information is not furnished in connection with a sale or offer to sell securities, or in connection with the solicitation of an offer to buy securities. This firm and its parent and/or their officers, stockholders, or members of their families may, from time to time, have a long or short position in the securities mentioned and may sell or buy such securities. Printed in the United States of America. All rights reserved. No part of this publication may be reproduced, stored in retrieval systems, or transmitted, in any form or by any means—mechanical, electronic, photocopying, duplicating, microfilming, videotape, or otherwise—without the prior written permission of the publisher.

~~FOCUS CONFERENCE AGENDA ElBctroirie Dm^n Automation: The Tochnolitgy Ronalssanco Sponsored by the CAD/CAM Industry Service May 10 and 11, 1984 Sunnyvale Hilton Inn, Sunnyvale, California~~

~~Ail activities related to the Focus Conference will tie held in the Chardonnay Room~~

~~THURSDAY, May 10 FRIDAY, May 11~~

7:30 a.m. to 8:15 a.m. Integrated Systems Technology: 8:45 a.m. Conference Registration VLSI Design Automation for Systems Engineers 8:45 a.m. Welcome and Conference introduction Phil Kaufman, President James Newcomb, Director Silicon Compilers, Inc. CAD/CAM Industry Service Dataquest Incorporated 9:00 a.m. EDA Is for PCB CAD Too Buck Feltman, President 9:00 a.m. EDA industry Overview Cadnetix, Inc. Beth Tucker, Industry Research Manager CAD/CAM Industry Service 9:45 a.m. Coffee Break Dataquest Incorporated 10:00 a.m. The World According to Venture Capital 9:45 a.m. What's All the Excitement About— Grant Heidrich, General Partner How Can an Industry Grow So Fast? Mayfield Fund Gerald Langler, Vice President, Marl

~~2:15 p.m. Coffee Break~~

~~2:30 p.m. Users' Points of View- Panel Discussion • Expectations • Needs • Futures • Applications 4:30 p.m. to 6:00 p.m. Cocktails and Hors d'Oeuvres~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dataquest~~

~~CAD/CAM FOCUS CONFERENCE EVALUATION QUESTIONNAIRE Sunnyvale, California — May 10 and 11, 1984~~

~~Thank you tor attending our CAD/CAM Industry Focus Conference. Would you please assist us in planning our next Focus Conference by completing and returning this questionnaire?~~

~~1. Please rate each presentation on a scale of 1 to 10 (where 10 is highest in terms of your approval):~~

~~CONTENT DELIVERY COMH/IENTS (1 to 10) (1 to 10) (Use reverse side if necessary)~~

Hicker, EDA Industry Overview Langler, What's All The Excitement About Anderson, Application-Specific Hardware Jaeger, Design Interface Strategies McKenzie, The Synergy of ASICs and EDA Panel Discussion, Users' Points of View Kaufman, Integrated Systems Technology Feltman, EDA Is for PCB CAD Too Heidrich, The World According to Venture Capital Binder, The EDA Industry Jones, Coping with a Changing Industry

~~2. How would you rate the conference facilities (1 to 10)?~~

~~Location Guest Rooms Meals Meeting Rooms Recreational Facilities .~~

~~3. How would you rate the service quality and staff attitude (1 to 10)?~~

~~The hotel staff DATAQUEST staff~~

~~4. What did you like most about the Focus Conference?~~

~~5. What did you like least abovi the Focus Conference?~~

~~6. Suggestions for improving the Focus Conference:~~

~~(OVBr)~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 7. Topics that would be of interest to you for the next Focus Conference:

8. Do you prefer D longer, D shorter, or D the same length conference? 9. Do you prefer D more, O less, or D the same amount of free time? 10. Your primary interest in the CAD/CAM Industry is as a: D Manufacturer D Service Vendor D User O Financial Analyst D Other

~~Name and Company (optional)~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dataquest~~

~~CAD/CAM Focus Conference~~

~~May 10 and 11, 1984~~

~~Sunnyvale, California~~

~~List of Attendees~~

~~Bob Anderson GenRad, Inc.~~

~~Jared Anderson Valid Logic Systems, Inc.~~

~~Lloyd Anderson Dataproducts~~

~~Phil Arana Calma Company~~

~~Joseph Augustino HHB-Softron~~

~~Raj Badhwar Calma Company~~

~~Alan Bagley Hewlett-Packard~~

~~David L. Bailey Cericor, Inc.~~

~~Hal Barbaur GenRad~~

~~Andy Barnes DATAQUEST Incorporated~~

~~Bill Barret Caddrx~~

~~Tom Binder Silvar-Lisco~~

~~Bill Bowman Hewlett-Packard~~

~~Darko Brodarac Robert Bosch GMBH Reutlingen~~

~~Sharlene Burden DATAQtJEST Incorporated~~

~~Dave Burdick DATAQUEST, Incoporated~~

~~Charles Bures Data General Corp.~~

~~Bob Burkhardt Advance Micro Devices, Inc.~~

~~Gil Burns Teradyne, Inc.~~

~~Draiald Butler General Instrument Corp.~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Nelson Bye Data General Corp.~~

~~Frank Calabretta Holt, Inc.~~

~~Steve Chadima Regis McKenna~~

~~Wade Chanz E.R.S.O.~~

~~Wei-tau Chiang Intel Corp.~~

~~Art Collmeyer Weitek~~

~~Laura Conigliaro Prudential Bache~~

~~Mike Connell LSI Logic Corp.~~

~~Jim Cooper Tektronix~~

~~DOTI Corson Hewlett-Packard~~

~~John Crist Tektronix~~

~~Dave Crockett DATAQtJEST, Incorporated~~

~~P. Csao E.R.S.O.~~

~~Tennyson Csen E.R.S.O.~~

~~Ang Danna Arthur Young & Company~~

~~Sam Daram Calma Company~~

~~Ugo De Riu Olivetti System Development Div.~~

~~John DiGirolamo GenRad, Inc.~~

~~Bob Draeger Tektronix~~

~~Francois Durif GenRad, Inc.~~

~~Bob Ouyn VLSI Techology, Inc.~~

~~Mary Jane ELnore Institutional Investment Partners~~

~~Doug Fairbeirn VLSI Technology, Inc.~~

~~Daniel Fsui Hewlett-Packard~~

~~Buck Feltman Cadnetix, Inc.~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Ralph Pinley DATAQUEST Incorporated~~

~~Suzanne Pinocchio Bank of America~~

~~Kamran Firooz Hewlett-Packard Manufacturing Test Div.~~

~~David Flaningam Test Systems Strategies, Inc.~~

~~Frank Florence DATAQUEST Incorporated~~

~~Rick Fluegel Matrix Partners~~

~~Jim Fogle Teradyne, Inc.~~

~~Curt Francis Advanced Micro Devices, Inc.~~

~~Brian Gardiner Calma Company~~

~~Jim Gay Perq Systems Corp.~~

~~Majid Ghafghaichi Data General Corp.~~

~~John Gilbert Cigna Investment Management~~

~~Tom Gnuse Prime Capital Management~~

~~Chris Goldstein Applicon , •~~

~~Stan Graham Modern Electrosystems, Inc.~~

~~Roberto Grasso Ing. C. Olivetti & C, SPA~~

~~W. RobertGreer Automated Systems Inc.~~

~~Horst Gschwendtner Robert Bosch GMBH~~

~~Dave Gullickson Versatec~~

~~Guy Haas Computervision Corp.~~

~~Kevin Hall Calma Company~~

~~James Harrison Adler & Company~~

~~Gib Battery CAE Systems, Inc.~~

~~Ron Hayman Motorola, Inc.~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Grant Held rich Mayfield Fund~~

~~Henry S. Hofmann Hewlett-Packard Technology Center~~

~~William G. Holt Holt Inc.~~

~~Jeff Hotchkiss Teradyne, Inc.~~

~~Mike W. Hu Televideo~~

~~Paul Huber Megatek Corp.~~

~~Russel Huber Computervision Corp.~~

~~Claudia Huntington Capital Research~~

~~J. E. Iwerson Bell Labs~~

~~John Jackson DATAQOEST, Incoporated~~

~~Debra Jacob DATAQUEST, Incoporated~~

~~Kent Jaeger California Devices, Inc.~~

~~Jan Janse Siemens Research Labs~~

~~Bob J. Jenkins Motorola, Inc.~~

~~Richard Jennings VIA Systems, Inc.~~

~~Alan Jepsen Bank of America~~

~~Harvey Jones Daisy Systems Corp.~~

~~Robert Joseph Philips Test & Measuring Instruments~~

~~Bill Raiser Apollo Computer~~

~~Thanasis Kalekos Versatec~~

~~Ram Kasargod' IDS~~

~~Phillip A. Kaufman Slicon Compilers, Inc.~~

~~Dave Kaverman Hewlett-Packard~~

~~Norman F. Kelly Altos Computer Systems~~

~~Bill Knox CAE Systems Inc.~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Mike Knudsen CalComp~~

~~Michael Kretsch Exlsi~~

~~Rob Kuhling Calma Conpany~~

~~Gerard Langler Mentor Graphics Corp.~~

~~Wendy Ledamun DATAQUEST, Incoporated~~

~~Frank Lee Monolithic Memories~~

~~Kelly Leininger DATAQUEST, Incoporated~~

~~Carol Leralein Teradyne, Inc.~~

~~Robert Leurz Eastman Kodak Company~~

~~Van Lewing Calma Company~~

~~Rhonda Y. Lindsey CADAM Inc.~~

~~Shjjninz Lins B.R.S.O.~~

~~Ed Locke United Technologies MOSTEK~~

~~Jim Long Fujitsu Microelectronics~~

~~Bob Lorentzen Zymos~~

~~Peter Lowell Modern Electrosystems, Inc.~~

~~Francis T. Lynch Valid Logic Systems~~

~~R. Van Maas Philips~~

~~Mark Maltese Versatec~~

~~Bob Martin Tektronix~~

~~Edie Mata Fujitsu Microelectronics~~

~~Roland Mattison VIA Systems, Inc.~~

~~Samuel P. McKay Ventech Partners~~

~~Ken McKenzie DATAQUEST Incorporated~~

~~Daniel C. McNiell Gould, Inc.~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Robert Mclellan Bell Northern Research~~

~~David H. Miller Applicon~~

~~Linda Miller InterWest Partners~~

~~Richard Mirabella Phoenix Data Systems~~

~~Philip Monego Paragon Technology Corp.~~

~~Bob Moore Hewlett-Pac kard~~

~~Eric Mudama GenRad, Inc.~~

~~Carolyn Mullery Calma Company~~

~~Barrie Murray-Upton Racal-Redac Limited~~

~~Karl Nakamura Grid Systems~~

~~John Nam Modern Electrosystems, Inc.~~

~~Richard Nedbal Personal Cad Systems~~

~~Jim Newcomb DATAQUEST, Incorporated~~

~~Dave Niehaus Calma Company~~

~~Jacit Nilsson Hewlett-Packard~~

~~Earl North Circuit Tools, Inc.~~

~~John O'Brien Teradyne, Inc.~~

~~Elizabeth Obershaw Hewlett-Packard~~

~~Dewitt Ong Intel Corp.~~

~~Alan Oppeinheimer Apollo Computer~~

~~Han Park Qume Corp.~~

~~Dennis Patton Ridge Computers~~

~~Nikola Pavicic Modern Electrosystems, Inc.~~

~~Evelyn Peairs Calma Canpany~~

~~Jeff Pepper Perq Systems~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Richard Perry LSI Logic Corp.~~

~~Rich Peters CalCcxnp~~

~~Jewel Peyton DATAQUEST, Incoporated~~

~~Roiald J. Piziali Arthur Young & Company~~

~~Katy Plumb First Interstate Bank of California~~

~~Jay Popper Standard Microsystems Corp.~~

~~Roy Prasad Personal Cad Systems~~

~~Andy Prophet DATAQOEST, Incoporated~~

~~Jamshed Qamar OKI Semiconductors~~

~~F. Ramsay Ferranti Electronics~~

~~Karen Regnante Mosaic Technology~~

~~Douglas Ritchie National Semiconductor~~

~~Carlo Ronea Olivetti System Development Div.~~

~~Frank Sammann DATAQDEST Incorporated~~

~~Mary Ellen Saxby DATAQUEST, Incoporated~~

~~Katherine Schapiro Harris Bretall~~

~~Cesare Segre Ing. C. Olivetti and C, SPA~~

~~C. Seih E.R.S.O.~~

~~John Sheehan Intel Corp.~~

~~Steve Shively International Microcircuits, Inc.~~

~~Pam Shook DATAQOEST, Incoporated~~

~~Chester Silvestri Megatest~~

~~Mike Sisavic Metheus Corp.~~

~~G. Skorup South African Micorelectronics Systems~~

~~Joseph Sliwkowski Telesis Systems Corp.~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 William Smit University of Pretoria~~

~~Jack Spigarelli Cericor, Inc.~~

~~Rich Stamm Data General Corp.~~

~~J. Stevenson Philips~~

~~Gar^ Stump Tektronix~~

~~Leslie Swanson Computervision Corp.~~

~~Tom Teeters Calma Company~~

~~Carol Tegnell Scientific Calculations, Inc.~~

~~Robert S. Tepper A T s T Bell Laboratories~~

~~Ben Thomas Calma Company~~

~~Mel Thanpson DATAQUEST, Incoporated~~

~~Rick Timmons Wyle Labs~~

~~Paul R. Topping McDonnell Dougals Automation (McAuto)~~

~~Beth Tucker DATAQUEST, Incorporated~~

~~Steve Vaughn IBM~~

~~Dave Vogte Teradyne, Inc.~~

~~Kris Vorm Calma Company ^~~

~~Fred Wagner IBM~~

~~Martin Walker Analog Design Technology~~

~~Irving Weiman R&D Funding Corp. Pru-Bache~~

~~Michael Weisberg Prudential Bache~~

~~Patrick Welch Hewlett-Packard~~

~~Jerry Werner VLSI Design Magazine~~

~~Gary Westbrook Calma Cranpany~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 David Wharton GenRad, Inc.~~

~~Alan Winslow Tektronix~~

~~Larry Yamada Calma Company~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 (::)aDataquest Dataquest~~

~~CAD/CMl Industry Conference~~

~~May 10-11, 1984~~

~~Sunnyvalf, California~~

~~Final List of Attendees~~

~~Scott Alan Gould Electronics~~

~~Bob Anderson GenRad, Inc.~~

~~Colleen Anderson Valid Logic Systems, Inc.~~

~~Jared Anderson Valid Logic Systems, Inc.~~

~~Lloyd Anderson Dataproducts~~

~~Phil Arana Calma Canpany~~

~~Joseph Augustino HHB-Softron~~

~~Rajiv Badhwir Calma Conpany~~

~~David L. Bailey Cericor, Inc.~~

~~Hal Barbaur GenRad~~

~~Andy Barnes DATAQUEST Incorporated~~

~~Bill Barrett Caddrx~~

~~Tom Binder Silvar-Lisco~~

~~Bill Bowman Hewlett-Packard~~

~~Darko Brodarac Robert Bosch GMBH Reutlingen~~

~~Sharlene Burden DATAQUEST Incorporated~~

~~Dave Burdick DATAQUEST, Incoporated~~

~~Charles Bures Data General Corp.~~

~~Bob Burkhardt Advance Micro Devices, Inc.~~

~~- 1~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Gil Burns Teradyne, Inc.~~

~~Donald Butler General Instrument Corp.~~

~~Nelson Bye Data General Corp.~~

~~Frank Calabretta Holt, Inc.~~

~~Xaviar Candia Calma Conpany~~

~~Wade Chang E.R.S.O.~~

~~Tennyson Chen E.R.S.O.~~

~~Wei-tau Chiang Intel Corp.~~

~~Art Collmeyer Weitek~~

~~Laura Conigliaro Prudential Bache~~

~~Mike Connell LSI Logic Corp.~~

~~Jim Cooper Tektronix~~

~~Jan A. Cornish Versatec~~

~~Don Corson Hewlett-Packard~~

~~John Crist Tektronix~~

~~Ang Danna Arthur Young & Company~~

~~Sam Daram Calma Company~~

~~Ugo De Riu Olivetti Systems~~

~~Chuck Decker Electro Technology Industries, Inc.~~

~~Bob Draeger Tektronix~~

~~Francois Durif GenRad, Inc.~~

~~Dennis Elkins NCA Corp.~~

~~Mary Jane Elmore Institutional Investment Partners~~

~~William Escudero Televideo~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Daniel Fan Hewlett-Packard~~

~~George Faraco Electronic Trend Publications~~

~~Buck Feltman Cadnetix, Inc.~~

~~Ralph Finley DATAQOEST Incorporated~~

~~Suzanne Finocchio Bank of America~~

~~Kamran Firooz Hewlett-Packard Corp.~~

~~David Flaningam Test Systems Strategies, Inc.~~

~~Frank Florence DATAQtJEST Incorporated~~

~~Rick Flugel Matrix Partners~~

~~Fogle, Jim Teradyne, Inc.~~

~~Curt Francis Advanced Micro Devices, Inc.~~

~~Dan Garms CAD/CAM Specialists, Inc.~~

~~Jim Gay Perq Systems Corp.~~

~~Tom Gnuse Prime Capital Management~~

~~Chris Goldstein Applicon~~

~~Roberto Grasso Ing. C. Olivetti & C, SPA~~

~~Robert W. Greer Automated Systems Inc.~~

~~Horst Gschwendtner Robert Bosch GMBH~~

~~Dave Gullickscai Versatec~~

~~Guy Haas Ccxnputervision Corp.~~

~~Kevin Hall Calma Company~~

~~Bruce Hanson Gould Electronics~~

~~James Harrison Adler & Company~~

~~- 3 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Gib Hattery CAE Systems, Inc.~~

~~Ron Hayman Motorola, Inc.~~

~~Grant Heidrich Mayfield Fund~~

~~Henry S. Hofmann Hewlett-Pac kard~~

~~Sheila A. Holland Hitachi America Limited~~

~~William G. Holt Holt Inc.~~

~~Stefano Horpurgo Olivetti and C; SPA~~

~~Jeff Hotchkiss Teradyne, Inc.~~

~~Mike W. Hu Televideo~~

~~Paul Huber Megatek Corp.~~

~~Claudia Huntington Capital Research~~

~~J. E. Iwerson Bell Labs~~

~~John Jackson DATAQUEST, Incoporated~~

~~Debra Jacob DATAQUEST, Incoporated~~

~~Kent Jaeger California Devices, Inc.~~

~~K. Charles Janac Stanford University~~

~~Jan Janse Philips Research Labs~~

~~Richard Jennings VIA Systems, Inc.~~

~~Alan Jepsen Bank of America~~

~~Harvey Jones Daisy Systems Corp.~~

~~Robert Joseph Philips Test & Measuring Instrument~~

~~Bill Raiser Apollo Computer~~

~~Thanasis Kalekos Versatec~~

~~- 4 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Ram Kasargod IDS~~

~~Phillip A. Kaufman Slicon Conpilers, Inc.~~

~~Dave Kaverman Hewlett-Packard~~

~~Norman F. Kelly Altos Ccnnputer Systems~~

~~Bill Knox CAE Systems Inc.~~

~~Mike Knudsen CalComp~~

~~Michael Kretsch Exlsi~~

~~Gerard Langler Mentor Graphics Corp.~~

~~Lucio Lanza Daisy Systems Corp.~~

~~Wendy Ledamun DATAQUEST, Inooporated~~

~~Greg Ledenbach National Semiconductor~~

~~Frank Lee Monolithic Memories~~

~~Kelly Leininger DATAQUEST, Inooporated~~

~~Carol Lemlein Teradyne, Inc.~~

~~Van Lewing Calma Company~~

~~Shyiuing Lin E.R.S.O.~~

~~Rhonda Y. Lindsey CADAM Inc.~~

~~Robert Lurz Eastman Kodak Company~~

~~Francis T. Lynch Valid Logic Systems~~

~~R. Van Maas Philips~~

~~Leslie Marshall Calma Company~~

~~Bob Martin Tektronix~~

~~Edie Mata Fujitsu Microelectronics~~

~~Roland Mattison VIA Systems, Inc.~~

~~- 5 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Michael W. McDonald CAD/CAM Specialists, Inc.~~

~~Ken Mckenzie DATAQOEST Incorporated~~

~~Daniel C. McNiell Gould, Inc.~~

~~Robert McLellan Bell Northern Research~~

~~Charles Miller Analog Design Technology~~

~~David H. Miller Applicon~~

~~Linda Miller InterWest Partners~~

~~Richard Mirabella Phoenix Data Systems~~

~~Philip Monego Paragon Technology Corp.~~

~~Bob Moore Hewlett-Packard~~

~~Er ic Mudama GenRad, Inc.~~

~~Carolyn Mullery Calma Company~~

~~Barrie Murray-Upton Racal-Redac Limited~~

~~Karl Nakamura Grid Systems~~

~~John Nam Modern Electrosystems, Inc.~~

~~Jim Newcomb DATAQUEST, Incorporated~~

~~Dave Niehaus Calma Company~~

~~Jack Nilsson Hewlett-Packard~~

~~Earl North Circuit Tools, Inc.~~

~~John O'Brien Teradyne, Inc.~~

~~Elizabeth Obershaw Hewlett-Packard~~

~~Dewitt Ong Intel Corp.~~

~~Alan Oppeinheinter Apollo Computer~~

~~Han Park Qume Corp.~~

~~- 6 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Evelyn Peairs Calma Company~~

~~Jeff Pepper Perq Systems~~

~~Michele Perrota Daisy Systems Corp.~~

~~Richard Perry LSI Logic Corp.~~

~~Ross Peters Murray Electronics PLC~~

~~Jewel Peyton DATAQtJEST, Incoporated~~

~~Cathy Pierson Oak Management~~

~~Roland J. Piziali Arthur Young & Company~~

~~Jay Popper Standard Microsystems Corp.~~

~~Roy Prasad Personal Cad Systems~~

~~Jamshed Qamar OKI Semiconductors~~

~~F. Ramsay Ferranti Electronics~~

~~Karen Regnante Mosaic Technology~~

~~Robert Resslhuber CcMnputervision Corp.~~

~~Carlo Ronca Olivetti System Development Div.~~

~~Mary Ellen Saxby DATAQUEST, Incoporated~~

~~Kathrine Schapiro Harris Bretall~~

~~John Seaton Gould Electronics~~

~~Paul P. Seckendorf Silvar-Lisco~~

~~C. Seih E.R.S.O.~~

~~Ariel Sella Softel-Advanced Technology~~

~~John Sheehan Intel Corp.~~

~~Steve Shively International Microcircuits, Inc.~~

~~Pamela Shook DATAQUEST, Incoporated~~

~~- 7 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Scott Shull Intel Corp.~~

~~Chester Slivestri Megatest~~

~~Mike Sisavic Metheus Corp.~~

~~G. Skorup South African Microelectronics Sys,~~

~~Joseph Sliwkowski Telesis Systems Corp.~~

~~William Smit University of Pretoria~~

~~Jack Spigarelli Cericor, Inc.~~

~~Rosann Stach NCA Corp.~~

~~Rich Stamm Data General Corp.~~

~~J. Stevenson Philips~~

~~Mrogan W. Stuart NCA~~

~~Gary Stump Tektronix~~

~~Leslie Swanson Computervision Corp.~~

~~Carol Tegnell Scientific Calculations, Inc.~~

~~Robert S. Tepper AT&T Bell Laboratories~~

~~Ben Thomas Calma Company~~

~~Mel Thompson DATAQUEST, Incoporated~~

~~Rick Timmons V^le Labs~~

~~Paul R. Topping McDonnell Douglas Automation~~

~~Walter Trevaskis AT&T~~

~~Y. C. Tsai E.R.S.O.~~

~~Graham S. Tubbs Intel Corp.~~

~~Beth Tucker DATAQUEST, Incorporated~~

~~- 8 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Ruth Tucker Versatec Bill Unger Mayfield Fund~~

~~Steve Vaughn IBti~~

~~Dave Vogt:e Teradyne, Inc.~~

~~Kris Vorm Calma Ccnnpany~~

~~Fred Wagner IBM~~

~~Irving Weiman R&D Funding Corp. Pru-Bache~~

~~Michael Weisberg Prudential Bache~~

~~Patrick Welch Hewlett-Packard~~

~~Jerry Werner VLSI Design Magazine~~

~~Gary Westbrook Calma Ccxnpany~~

~~Alan Winslow Tektronix~~

~~Larry Yamada Calma Company~~

~~Bryan Yates Calma Company~~

~~- 9 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 (:|)a Dataquest Dataquest~~

~~CAD/CAM Focus Conference~~

~~May 10 and 11, 1984~~

~~Sunnyvaler California~~

~~List of Attendees~~

~~Robert S.Tepper AT&T Bell Laboratories~~

~~James Harrison Adler & Company~~

~~Bob Burkhardt Advance Micro Devices, Inc.~~

~~Curt Francis~~

~~Norman F.Kelly Altos C<»nputer Systems~~

~~Martin Walker Analog Design Technology~~

~~Bill Raiser Apollo Computer~~

~~Alan Oppeinheimer~~

~~Chris Goldstein Applicon~~

~~David H.,Miller~~

~~Ang Danna Arthur Young & Company~~

~~Ronald J. Piziali~~

~~W. Robert Greer Aut(nnated Systems Inc.~~

~~Suzanne Finocchio Bank of America~~

~~Alan Jepsen~~

~~J. E. Iwerson Bell Labs~~

~~Robert Hclellan Bell Northern Research~~

~~Rhonda Y. Lindsey CADAM Inc.~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Bill Knox CAB Systems Inc.~~

~~Gib Battery~~

~~Bill Barret Caddrx~~

~~Buck Feltman Cadnetix, Inc.~~

~~Mike Rnudsen CalCcanp~~

~~Rich Peters~~

~~Kent Jaeger California Devices, Inc.~~

~~Phil Arana Calma Company~~

~~Raj Badhwar~~

~~Sam Oaram~~

~~Brian Gardiner~~

~~Kevin Hall~~

~~Rdb Kuhling~~

~~Van Lewing~~

~~Carolyn Mullery~~

~~Dave Niehaus~~

~~Evelyn Peairs~~

~~Tom Teeters~~

~~Ben Thomas~~

~~Kris Vorm~~

~~Gary Westbrcxsk~~

~~Larry Yamada~~

~~Claudia Huntington Capital Research~~

~~David L. Bailey Cericor, Inc.~~

~~Jacic Spigarelli~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 John Gilbert Cigna Investment Management~~

~~Earl North Circuit Tools, Inc.~~

~~Guy Haas Cc^putervision Corp.~~

~~Russel Huber~~

~~Leslie Swanson~~

~~Andy Barnes DATAQtJEST Incorporated~~

~~Sharlene Burden~~

~~Ralph Finley~~

~~Frank Florence~~

~~Ren McKenzie~~

~~Frank Sammann~~

~~Dave Hurdlek~~

~~John Jackson~~

~~Debta Jacob~~

~~Wendy Ledamun~~

~~Kelly Leininger~~

~~Jewel Peyton~~

~~Andy Prophet~~

~~Mary Ellen Saxby~~

~~Pan Shook~~

~~Mel Thompson~~

~~Dave Crockett~~

~~Jim Newcomb~~

~~Beth Tucker~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Harvey Jones Daisy Systems Corp.~~

~~Charles Bures Data General Corp.~~

~~Nelson Bye~~

~~Majid Ghafghaichi~~

~~Rich St:aniin~~

~~Lloyd Anderson Dataproducts~~

~~Wade Chanz E.R.S.O.~~

~~P. Csao~~

~~Termyson Csen~~

~~Shiminz Lins~~

~~C. Seih~~

~~Revert Leurz Eastman Kodak Company~~

~~Michael Kretsch Exlsi~~

~~F. Ramsay Ferranti Electronics~~

~~Katy Plumb First Interstate Bank of California~~

~~Jim Long Fujitsu Microelectronics Bdie Mata~~

~~Hal Bacbaur GenRad Bob Anderson~~

~~John DiGirolamo~~

~~Francois Our if~~

~~Eric Mudama~~

~~David Wharton~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Donald Butler General Instrument Corp.~~

~~Daniel C. McNiell Gould, Inc.~~

~~Karl Nakamuca GRID Systems~~

~~Joseph Augustine HHB-Softron~~

~~Katherine Schapiro Harris Bretall~~

~~Alan Bagley Hewlett-Packard~~

~~Bill Bowman~~

~~Don Corson~~

~~Daniel Fan~~

~~Dave Kaverman~~

~~Bob Moore~~

~~Jack Nilsson~~

~~Elizabeth Obershaw~~

~~Patrick Welch~~

~~Firooz Kamran~~

~~Henry S. Hoffman~~

~~William G. Holt Holt Inc.~~

~~Frank Calabretta Holt, Inc.~~

~~Steve Vaughn IBM~~

~~Fred Wagner IDS Ram Kasargod Ing. C. Olivetti & C, SPA Roberto Grasso~~

~~Cesare Segre~~

~~Mary Jane Elmore Institutional Investment Partners~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Wei-tau Chiang Intel Corp.~~

~~Dewitt Ong~~

~~John Sheehan~~

~~Linda Miller InterWest Partners~~

~~Steve Shively International Microcircuits, Inc.~~

~~Mike Connell LSI Logic Corp.~~

~~Richard Perry~~

~~Rick Fluegel Matrix Partners~~

~~Grant Heidrich Mayfield Fund~~

~~Paul R. Topping McDonnell Dougals Automation (McAuto)~~

~~Paul Huber Megatek Corp.~~

~~Chester Silvestci Megatest~~

~~Gerard Langlec Mentor Graphics Corp.~~

~~Mike Sisavic Metheus Corp.~~

~~Stan Graham Modern Electrosystems, Inc.~~

~~Peter Lowell~~

~~John Nam~~

~~Nikola Pavicic~~

~~Frsmk Lee MCHiolithic Memories~~

~~Karen Regnante Mosaic Technology~~

~~Ron Hayman Motorola, Inc.~~

~~Bob J. Jenkins~~

~~Douglas Ritchie National Saniconductor~~

~~Jamshed Qamar OKI Semiconductors~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen CoInpany /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Ugo De Riu Olivetti System Development Div.~~

~~Carlo Ronca~~

~~Philip Monego Paragon Technology Corp.~~

~~Jeff Pepper Perq Systems~~

~~Jim Gay~~

~~Richard Nedbal Personal Cad Systems~~

~~Roy Prasad~~

~~R. Van Maas Philips~~

~~J. Stevenson~~

~~Robert Joseph Philips Test & Measuring Instruments~~

~~Richard Mirabella Phoenix Data Systems~~

~~Prime Capital Management Tom Gnuse~~

~~Laura Conigliaro Prudential Bache~~

~~Michael Weisberg~~

~~Han Park Qume Corp.~~

~~Irving Weiman R&D Funding Corp. Pru-Bache~~

~~Barrie Murray-Upton Racal-Redac Limited~~

~~Steve Chadima Regis McKenna~~

~~Dennis Patton Ridge Computers~~

~~Horst Gschwendtner Robert Bosch GMBH~~

~~Oarko Brodarac~~

~~Carol Tegnell Scientific Calculations, Inc.~~

~~Jan Janse Siemens Research Labs~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Tom Binder Silvar-Lisco Phillip A. Kaufman Silicon Compilers, Inc.

~~G. Skorup South African Micorelectronics Systems~~

~~Jay Popper Standard Microsystems Corp.~~

~~Jim Cooper Tektronix~~

~~John Crist~~

~~Bob Draegec~~

~~Bob Martin~~

~~Gary Stump~~

~~Alan Winslow~~

~~Joseph Sliwkowski Telesis Systems Corp.~~

~~Mike W. Hu Televideo~~

~~Gil Burns Teradyne, Inc.~~

~~Jim Pogle~~

~~Jeff Hotchkiss~~

~~Carol Lemlein~~

~~John O'Brien~~

~~Dave Vogte~~

~~David Flaningam Test Systems Strategies, Inc.~~

~~Ed Locke United Technologies MOSTEK~~

~~William Smit University of Pretoria~~

~~Richard Jennings VIA Systems, Inc.~~

~~Roland Mattison~~

~~Jerry Werner VLSI Design Magazine~~

~~Dataquest incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Doug Fairbeirn VLSI Technology/ Inc.~~

~~Bob Duyn~~

~~Francis T. Lynch Valid Logic Systems~~

~~Jared Anderson~~

~~Samuel F. McKay Ventech Partners~~

~~Dave Gullickson Versatec~~

~~Thanasis Kalekos~~

~~Mark Maltese~~

~~Art Collmeyer Weitek~~

~~Rick Tinunons Wyle Labs~~

~~Bob Lorentzen Zymos~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dataquest Dataquest~~

~~WELCOME AND CONFERENCE INTRODUCTON~~

~~James R. Newcomb Director, CAD/CAM Industry Service Dataquest Incorporated~~

Mr. Newcomb is Director of DATAQUEST's CAD/CAM Industry Service. Before joining DATAQUEST, he was with Auto-trol Technology Corporation, where he held various executive positions, including Manager of Marketing. Prior to that, Mr. Newcomb was employed by Xerox Corporation, where he held positions as Manager of Systems Integration, Manager of Advanced Manufacturing Engineeringr and Divisional Program Planner. Earlier, he worked in an engineering capacity with Strippit-Hoidaille. His professional experience has provided him with an in-depth knowledge of CAD, CAM, CAE, robotics technology, artificial intelligence, and manufacturing automation systems. Mr. Newcomb has a B.S.M.E. from Rochester Institute of Technology. He is a member of the American Society of Mechanical Engineering, SME, RI/SME, and CASA/SME.

~~Dataquest Incorporated ELECTRONIC DESIGN AUTOMATION FOCUS CONFERENCE May 10-11, 1984 Sunnyvale, California~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen CoInpany / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 DATAQUEST'S CAD/CAy INDUSTRY SERVICE WELCOMES YOU TO THE 1984 EDA FOCUS CONFERENCE

~~;:-f;:,-:<|A:;,V~~

~~O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 -!• ":• '!'•-' '- ;~~

~~ELECTRONIC DESIGN AUTOMATION "THE TECHNOLOGY RENAISSANCE" WELCOME DATAQUEST CORPORATE OVERVIEW #1 CAD/CAM INDUSTRY SERVICE CAD/CAM MARKET ImI~~

~~^•••tsJiWu^^msm^^^m^m^^m^^^^mMi^C Sj;ai:i^3j!RiSi^î^%;îgî5iSgii^*ji^ iîîî ifejaî^Ji^îiîiiyiiiiiii^siiSiîSiiS^;: j^^ :iiiiliii\iMiii\-4-i^~~

~~-2- C 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 DATAQUEST BUSINESS GROWTH~~

~~1.000-1~~

~~1.700~~

~~1,500~~

~~Ul 2 1,300 /~~

~~flC / lU g 1.100 /~~

~~soo < 700 600~~

~~1979 1980 19*1 1982 1983~~

~~^JMl :X\ Dataquest~~

~~C 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen CompanrV+-290 Ridder Park-Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 DATAQUEST TECHNOLCGY CUENTS (By Type of Business)~~

~~jKlDataQuest~~

~~e 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 /. (408) 971-9000 / Telex 171973 DATAQUEST TECHNOLOGY CUENTS (By Geographic Area)~~

~~BALANCE OF WORLD 2%~~

~~-7- O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 TECHNOLOGY SERVICES SUBSCRIBER GROWTH~~

~~1,900-~~

~~1,700-~~

~~1,500-~~

~~1,300-~~

~~1,100-~~

~~900~~

~~700- 600-~~

~~-T- 79 '80 •81 ^ •83~~

~~ui;'^~~

~~-8- O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park D'rive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD/CAM INDUSTRY SERVICE~~

~~APPLICATION SEGMENTATION~~

~~-11^- e 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD/CAM INDUSTRY SERVICE~~

~~COMPANY/PRODUCT SEGMENTATION~~

~~-12- O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD/CAM TOTAL ANNUAL REVENUE ALL APPLICATIONS/WORLDWIDE Millions of Dollars 15,000 1S83-1988 CAGR 39% 10.000

W^ p-?:E feS:;;;:-; 5,000 E4:ffip>"s*- ^^^^.SA^ ^^^-t^^-s piiii: vc -^ yzzi^i •S^:S3 fe«^-S^jjjjTtTiH f ^•l^i »J>;^^;^;>:J• fcxsita m~:i-' "i*KS^ 'InSffffti. mBu 1983 1984 1985 1986 1987 1988

~~Source: DATAQUEST~~

~~-13- O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD/CAM SYSTEMS SHIPPED ALL APPLICATIONS/WORLDWIDE Units 80,000 1983-1968 CAGR 62% 60.000 m^m s ••;;•• • r:, • 40,000

~~i .• 20.000~~

y^ -r. • •*

~~^'.z-c^yi-yy. i'Z'i 0 y;-:^:::^^::j? ^sas^^^ 1983 1984 1985 1986 1987 1988~~

~~Source: DATAQUEST~~

~~-14- e 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 !saiiSSS3«£S;SS«l»5SSS»»5i;3Ea;iJ;5;* FOCUS ON EDA

~~• INDUSTRY OVERVIEW • EVOLUTION OF AN INDUSTRY • SEMICONDUCTOR REVOLUTION • VENDORS AND USERS • FUTURES. TRENDS BECOME REALITIES~~

~~-15- O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD/CAM REVENUE BY APPLICATION Billions of Dollars~~

~~V? [ZZI 1983 Ea 1388 i~~

~~V?.~~

~~Zl MECH JAE C m.IC PC EDmA MAP CAGR 372 392 312 452 652 342 1983-1988~~

~~Source: DATAQUEST~~

~~-16- 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 -17- © 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited

~~Dataquest Incorporated, A Subsidiary of A.C^ Nielsen Company /1290 Bidder Park Drive / SanJosej CA 95131 / (408) 971-9000 / Telex 171973 L-.£^ I ):::;a Dataquest Dataquest~~

~~EDA IiroUSTRY OVERVIEW~~

~~Beth W. Tucker Industry Research Manager CAD/CAM Industry Service Dataquest Incorporated~~

Ms. Tucker is Industry Research Manager for DATAQUEST's CAD/CAM Industry Service. In this position, she is responsible for analyzing and reporting on the electronics segment of the CAD/CAM industry. Prior to joining DATAQUEST, Ms. Tucker was Market Planning Manager of Calma Company's Microelectronics Division, where she was responsible for identifying market opportunities and developing marketing strategies for the company's electronic CAD product line. She later contributed to the development of marketing information systems at Calma's corporate level. Prior to that, she was a Senior Market Research Analyst with Monroe Systems for Business, where she was responsible for market research and development of small business computer products. Ms. Tucker holds an A.A. degree in Computer Technology from Morris County College in New Jersey and is doing course work in Business Administration at San Jose State University.

~~Dataquest Incorporated ELECTRONIC DESIGN AUTOMATION FOCUS CONFERENCE May 10-11, 1984 Sunnyvale, California~~

~~Dataquest Incorporated, A Subsidiary of A.G. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 INTRODUCTION~~

DATAQUEST is Sponsoring this Focus Conference on Electronic Design Automation (EDA) because it is the youngest, moIst dynamic segment within the CAD/CAM industry today. A five-year compound annual growth rate of 65 percent also makes EDA the fastest-growing CAD application segment. The DATAQtJEST conference will focus on defining the EDA market segment, the companies included in the segment, market size and forecasts, and significant trends.

What is Electronic Design Automation? EDA is an integrated set of design tools for electrical engineers designing electronic circuits and products. The key difference between EDA and other electronic CAD tools is the integration, or linkage, of the different phases within any design cycle, versus a discrete and isolated computer-aided design solution.

The application of EDA systems incorporates more than computer-aided engineering for electrical engineers. EDA has evolved to include not only automation of logic design and analysis, but automation of physical layout, engineering administration, and testing, to some degree. DATAQUEST believes that the accomplishments of EDA vendors, and the product differentiation that they offer are significant enough to warrant distinction from computer-aided engineering applications for mechanical or other engineering disciplines. Thus,

~~DATAQUEST CALLS IT EDA~~

~~THE EDA PERSPECTIVE~~

To put EDA in its proper perspective, we believe that a brief historical overview of DATAQUEST's CAD/CAM applications segments would be valuable. Upon its formation in 1981, the CAD/CAM Industry Service segmented the market into four applications: mechanical; architecture, engineering, and construction; mapping; and electronics. Responding to the market's and our clients' needs, we then refined the definition of the electronics segment to differentiate computer-aided design for integrated circuits and printed circuit boards.

In the publication of our 1983 estimates and forecasts, we further refined the electronics segment to include CAD systems used in electronic design automation. Thus, the CAD/CAM market is segmented based upon the following applications:

~~• Mechanical~~

~~• Architecture, Engineering, and Construction~~

~~• Mapping~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 • Electronic Design Automation~~

~~• Integrated Circuit~~

~~• Printed Circuit Board~~

~~ELECTRONIC CAD DEFINITIONS~~

To avoid confusion, we will define what we mean by IC and PC CAD, and EDA. The frame of reference in which these terms are defined is a simplistic overview of the phases in the electronic product design cycle listed below:

~~• Specification~~

~~• Design and Analysis~~

~~• Physical Layout~~

~~- Full Custom IC~~

~~- Gate Array~~

~~Standard Cell~~

~~•^ Printed Circuit Board~~

~~• Design Verification~~

~~• Analysis~~

~~• Test~~

~~• Manufacturing~~

~~IC CAD/CAM~~

IC CAD/CAM systems are used by layout designers to create the physical geometries of an integrated circuit. The layout designer typically works with hard copy forms of schematics created by an electrical engineer, and graphically translates the schematics into patterns to be recognized by a pattern generation machine for mask or direct-write geometries. As of 1984, most of the IC CAD/CAM systems available offer discrete functionality; i.e., they usually address only drafting and the analysis related to design or electrical rules verification. IC CAD systems typically do not address other phases of the design cycle.

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Companies included in DATAQUEST's IC CAD/CAM data base are:

~~• Applicon~~

~~« Calma Company~~

~~k Computervision~~

~~• Scientific Calculations~~

~~Via Systems~~

With a few exceptions, most IC CAD systems are host-dependent 16-bit computers, running in a centralized environment. Prior to the acceptance of semicustom circuits, IC CAD systems were used to design the geometries of full custom ICs. With the commercialization of semicustom ICs, these CAD systems may also be used to create the geometries of individual standard cells and the manufactured layers of gate arrays.

~~PCB CAD/CAM~~

CAD/CAM systems that are used to design printed circuit boards are going through an evolution, such that DATAQUEST now recognizes two distinct types of systems. Although the boundaries may not always be clearly separate, PCB CAD systems are differentiated on the basis of functionality, end-user, and type of hardware.

The first type is systems used for artwork-only creation, with little to no electrical intelligence associated with the data base. Systems within this category are used by drafting personnel in engineering services departments to create the geometries used for photoplotters, silk screens, numerical controls, drill paths, and documentation. The actual traces may be created interactively (i.e., manually) or automatically, as with packaging and placement functions. Some systems in this category have the ability to create schematics that are used for net list generation as input to an automatic router, back annotation, and documentation. Digitizers are often an integral part of these systems. The environment in which this type of system operates is usually a centralized CAD shop, with host-dependent computers and stations.

The second type of PCB CAD system is more comprehensive in nature. It is typically a 16- or 32-bit standalone system, with communication capabilities for use in a networked environment. In addition to engineering services' drafting personnel, it can also be used by electrical engineers to create PC board logic designs. The data base is often structured so that net lists can be generated as input to other simulation and analysis programs, in addition to placement and routing programs. These PCB vendors may directly sell analysis programs, or they may have joint marketing agreements with software companies.

~~O 1984 Dataquest Incorporated May 9 ed.-Reproduction Prohibited - 3 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Companies in DATAQUEST's PCS CAD/CAM data base include:~~

~~Applicon IBM~~

~~Cadnetix Paragon Technology~~

~~Calay Prime Computer~~

~~Calma Company Racal-Redac~~

~~Computervision Scientific Calculations~~

~~Digital Equipment Telesis Systems~~

~~EAS Vectron Graphics~~

~~Gerber Scientific Versatec~~

The phases in the design cycle that are addressed by PCB CAD/CAM vendors include the physical layout of printed circuit boards, design rule verification, and to some degree, design and analysis, post-layout analysis, and test.

~~Electronic Design Automation~~

EDA systems are used by electrical engineers and engineering staffs to create and design electronic products, including integrated circuits, printed circuit boards, and actual systems. EDA systems emphasize an integrated and comprehensive design tool solution, addressing all phases of the design cycle. The systems themselves are typically standalone, or if not standalone, they are relatively less expensive than traditional host-dependent CAD systems. They typically employ 16/32-bit microprocessor technology, and operate in a decentralized, networked environment.

When DATAQUEST first created a separate applications segment for EDA systems, they were used mostly for logic or circuit design, and for analysis programs such as circuit and logic simulation, timing verification, and fault simulation. Today, however, EDA systems address a broad range of applications, including physical layout of printed circuit boards and full custom ICs, and placement and routing programs for gate arrays and standard cells. And the market continues to expand its scope. For example, an integral part of an EDA system may also include dedicated hardware engines used to speed the process of simulating complex circuitry. Hardware logic analyzers integrated with design data bases are also used and sold by EDA vendors.

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Not only do EDA systems directly address electrical engineers' needs for design tools, but they also include functionality for word processing, technical publications, project management, and electronic mail.

~~Companies in DATAQUEST*s EDA data base currently include the following:~~

~~CAE Systems • Mentor Graphics~~

~~Cadtec • Metheus~~

~~Calma Company • Silvar-Lisco~~

~~Daisy Systems • Valid Logic~~

~~Intergraph~~

DATAQUEST believes that the list of vendors participating in the EDA segment will continue to grow. We believe that other major turnkey companies and computer companies will enter this market during the next several years, and that the market will become somewhat fragmented.

~~EDA MARKET SIZE~~

The electronic design automation segment grew 328 percent in 1983. The worldwide revenues of $90 million were highly concentrated in four of the ten companies currently selling EDA systems. Dataquest estimates that the EDA segment will increase approximately 136 percent during 1984, reaching $211 million in worldwide revenues.

~~Market Forecast~~

DATAQUEST estimates that the EDA market segment will grow at an average compound annual growth rate (CAGR) of 65 percent during the next five years, and that the market will be more than $1 billion in 1988. Workstation shipments are expected to grow by 67 percent per year, increasing from 1,329 units in 1983 to approximately 17,112 units in 1988. The table below represents our yearly estimates:

~~EDA MARKET ESTIMATES AND FORECASTS (Millions of Dollars)~~

~~1983 1984 1985 1986 1987 1988~~

~~Revenue 90 211 357 533 772 1,092 Workstations 1,325 3,223 5,421 8,002 11,774 17,112~~

~~O 1984 Dataquest Incorporated May 9 ed.-Reproduction Prohibited - 5 -~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 The EDA segment has gained a tremendous amount of momentum, considering its relatively early stage of development. We believe that this momentum will carry through 1988 due to:

~~Demand~~

~~Expanded product lines~~

~~Incorporation of new hardware technologies~~

~~New entrants~~

~~Application niches~~

Quarterly revenue of the three market leaders continues to increase at outstanding rates. There has been much publicity about the successful accomplishments of EDA companies, which is in keeping with their general high-awareness marketing strategies. The table below represents the sum of quarterly results of Daisy Systems, Mentor Graphics, and Valid Logic:

~~EDA QUARTERLY SALES (Millions of Dollars)~~

~~1981 1982 1983 1984~~

~~Quarter 1 0 1.17 7.16 39.21 Quarter 2 0 2.08 12.33 N/A Quarter 3 0 3.88 19.24 N/A Quarter 4 0.21 3.81 31.69 N/A~~

~~N/A = Not Available~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Market Share Estimates~~

DATAQtJEST currently follows ten companies in the EDA segment. Four of them comprise 88 percent of the 1983 revenues, while the other six companies comprise the remaining 12 percent. The table below represents our market share estimates:

~~EDA 1983 MARKET SHARE ESTIMATES (Millions of Dollars)~~

~~Revenue Market Share~~

~~Daisy Systems $28.4 31% Mentor Graphics 26.0 29% Silvar-Lisco 9.0 10% Valid Logic 16.1 18% Other* 10.5 12%~~

*Other includes: Cadtec, CAE, Calma, Control Data, Intergraph, and Metheus

~~Electronic CAD/CAM Estimates~~

DATAQireST estimates that the entire electronic CAD/CAM segment is growing at an average CAGR of 47 percent. EDA is growing the fastest, at a 65 percent CAGR, followed by printed circuit CAD/CAM with a 45 percent CAGR and integrated circuit CAD/CAM with a 31 percent CAGR.

DATAQUEST believes that another segmentation of the electronic CAD market is meaningful due to the close relationships and slight overlaps between the IC, PC, and EDA segments. The second market segmentation is based upon the electronic product design cycle discussed earlier in this pape We have estimated the size and growth of the electronic CAD/CAM market-based system usage within the following design phases:

~~Logic design and analysis~~

~~Pull custom IC layout~~

~~Gate array layout~~

~~Standard cell layout~~

~~Printed circuit board layout~~

~~The table below shows our estimates for workstation shipments, either standalone or host-dependent, that will be used in the previously described design cycle.~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ELECTRONIC CAD SYSTEM USAGE (Workstation Units)~~

~~1983 1988 CAGR~~

~~Logic Design and Analysis 1,240 14,253 63% Full Custom 778 3,782 37% Gate Array 80 2,915 105% Standard Cell 54 2,976 123% Printed Circuit 2,072 14,321 47%~~

These numbers represent the sum of total workstations in the IC, PCB, and EDA CAD/CAM segments; however, they were calculated based on design cycle phases, not on the type of company selling the workstation. For example, an EDA system can be used to do logic design, full custom layout, or printed circuit board layout. Conversely, a PC CAD system may be used for logic design and analysis in addition to printed circuit board layout.

We believe that the above design cycle segmentation, in conjunction with our standard applications segmentation, provides an accurate overall picture as to which company is selling a system, and what the system is being used for.

Systems used for printed circuit board layout is the largest segment in 1983 because vendors in the PC CAD/CAM segments have been established longer. We believe that these systems will decrease as a percentage of total system shipments because of increased usage in other design phases.

We expect that systems used for design automation, or logic design and analysis, gate array, and standard cell placement and routing will increase as a percentage of total shipments through 1988. All three have large CAGRs, which is due in part from the small base at which they are starting. Additionally, their growth is spurred by the acceptance of semicustom circuits and design methodologies.

In summary, we believe that the systems usage for companies currently in the IC, PC, and EDA segments will become increasingly difficult, as some start to expand their current product strategies and address design phases outside of their traditional markets.

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 DRIVING FACTORS~~

~~Circuit Complexity~~

Much has been ' said about the increasing complexity of integrated circuits. It is not feasible to create a VLSI circuit today without the aid of some type of computer-assisted design, analysis, or engineering progrem. It is not the intention of this DATAQUEST Focus Conference to discuss the growing complexitiy of circuits, but to draw a parallel to the need for engineering design tools to meet the needs of today's design engineers in working with these circuits.

~~Phrases like "productivity" and "speed to market" are associated with the EDA market. Productivity and speed to market, as it turns out, are EDA products' biggest benefits.~~

The goal is to reduce the time required to design a product, while maintaining or improving circuit functionality and complexity. The end result is a higher return on investment and a faster payback period for the company designing, manufacturing, or marketing a circuit, when compared with the product life cycle. Productivity and speed to market apply not only to semiconductor manufacturers, but to system or computer manufacturers.

~~Semicustom Circuits~~

With the increasing acceptance of gate array ICs and standard cell design methodologies, the time it takes to design and market a product is shortened even more than when compared to using an EDA system for full custom ICs. Not only is the design cycle shortened, but engineers who have not been trained to design specifically on silicon can now create semicustom circuits. The end result is a circuit that is customized to a particular application, has better functionality, and possibly has reduced product cost through the use of fewer printed circuit boards in a given system.

In his speech later today. Ken McKenzie, Associate Director of DATAQUEST's Semiconductor Industry Service, will further explore the relationship and synergy between semicustom chips and electronic design automation systems.

We believe that one of the goals of an EDA vendor with semicustom design capability is to put the task of creating integrated circuits into the hands of engineers designing entire systems, who are accustomed to using TTL parts. The total available market would then increase from approximately 3,000 silicon design engineers to approximately 450,000 total design engineers.

~~O 1984 Dataquest Incorporated May 9 ed.-Reproduction Prohibited - 9 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Hardware Technology~~

To effectively penetrate a market of approximately 450,000 worldwide electrical engineers, two actions must be taken simultaneously: decrease system costs and maintain an equal or better level of system functionality.

~~The typical system hardware of an EDA vendor today consists of the following:~~

~~16/32-bit microprocessor~~

~~1- to 4-MB memory~~

~~60- to 16G-MB Winchester-type disk drive~~

~~1-MB flexible disk drive~~

~~High-performance graphics terminal, preferably color~~

~~Applications software~~

Peripherals such as plotters, printers, communications, file servers, or tape drives, are available in addition to the above, depending on user requirements. The average selling price of the system hardware described above is approximately $60,000. The cost of the hardware continues to decrease over time as performance continues to increase.

DATAQUEST believes that to penetrate the market to the point where the majority of engineers have some type of interactive design tool on their desks, hardware prices must decrease more steeply than the predicted 15 percent per year for OEM cost, and 7.5 percent per year for end-user price. This is discussed further under the Trends section.

~~TRENDS~~

Due to the embryonic state of the EDA segment, practically everything that occurs within it is a significant trend. Major events such as product announcements, mergers and acquisitions, bankruptcies, third-party software agreements, public offerings, and new entrants, occur on a regular and frequent basis. The following trends are those that DATAQUEST believes have the most impact on the present EDA market.

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Marketing/Sales~~

To say that the EDA segment is highly competitive is a gross understatement, but it is true. DATAQUEST expects the EDA segment to maintain this high level of competition during the next several years, for the following reasons:

~~'Buy market share~~

~~New entrants~~

~~Many vendors (in relation to market life cycle)~~

~~Lean and meem~~

~~Functionality race~~

~~Market educational requirements~~

~~Market needs and demands~~

The market leaders have set the industry pace. New entrants must be willing and able to meet this pace. Maintaining intense sales and marketing positions, however, can be a financial drain on a company, or even a barrier to entry.

The following figures represent the sums of Daisy Systems', Mentor Graphics', Silvar-Lisco's, and Valid Logic's fiscal expenses as a percentage of revenue for each of their respective fiscal years.

~~EXPENSES AS A PERCENTAGE OF REVENUE (Millions of Dollars)~~

~~1982 1983~~

~~Revenue 10.6 63.7 Marketing, General and Administrative 66% 40% Research and Development 43% 16%~~

Expenses are expected to be high during the first several years of operation (notice 1982 was not profitable). However, DATAQUEST believes that the EDA segment will continue to reflect higher than average marketing and sales expenses as a percentage of revenue. We do not believe that competition will become less intense than it is today. The

~~O 1984 Dataquest Incorporated May 9 ed.-Reproduction Prohibited - 11 -~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 companies analyzed in the above figure, besides being the only public, nonsubsidiary EDA companies, are also the 1983 market leaders. We believe that they will continue to fight for their present positions as new companies enter the market.

~~Functionality~~

~~Closely related to the market competitiveness is product functionality, as evidenced by the major product announcements of almost all EDA vendors during the past seven months.~~

Product introductions are not haphazard, although some might be reactionary. Perhaps it is pure coincidence that three companies introduced a hardware simulation engine within five months of one another. Perhaps it was because they were all at relatively similar stages of organizational development and maturity. Nonetheless, it seems safe to conclude that what one company starts, others will follow.

~~In addition to the applications offered on current EDA systems, DATAQUEST believes that the following system functions will be added to EDA vendors' product portfolios:~~

~~Enhanced behavorial-level simulators~~

~~Enhanced modeling tools~~

~~Emphasis on structured logic design~~

~~Operating system software development tools~~

~~Application system software development tools~~

~~Automatic test equipment interfaces~~

~~Automated testing software~~

~~Integration of analog circuit design~~

~~Integration of PCB routers and system functionality~~

~~Silicon compilers~~

~~Integration with mechanical CAD data bases~~

~~Interfaces to CAM/CIM, if not the actual software products~~

~~Standard data base interfaces~~

~~Artificial intelligence~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 EDA vendors continue to strive toward a "complete" product. With 1983 R&D expenses approximately 16 percent of revenue, not only are vendors making enhancements (and bug fixes) to existing products, but the leaders continue to reach for state-of-the-art. Going back to the system benefits of increased productivity and speed to market, DATAQUEST believes that development efforts will continue to focus on automating the entire electronic product design cycle.

~~INTERFACES~~

In order to accomplish complete design automation, EDA vendors must contend with two forces. First, they must work with their target market's existing design tools. Although low-cost, interactive graphics systems have not been used very long for design automation, users do have a considerable investment in either in-house developed tools or tools purchased from software companies. To ease the acceptance of EDA systems, vendors must provide capabilities that interface with their users' existing tools.

The second force, and one that works against EDA vendors, is time. Given EDA's relatively early stage of development, it is virtually impossible to develop all software at once. Hence, there is additional need to provide interfaces to existing design aids and to other CAD/CAM systems. DATAQUEST does not believe that this is a negative factor, since it is ultimately most beneficial to the user and to the vendor that can support the widest capabilities.

Finally, interfaces in terms of libraries and data base structure are critical to the widespread acceptance of semicustom circuits. It is in the best interest of both EDA vendors and silicon foundries or semiconductor manufacturers to determine and agree upon standard interfaces. This way, EDA vendors would not lose sales because of the lack of a particular foundry's library. The foundry would benefit from increased distribution channels through each of the EDA vendors' sales forces and installed bases.

~~Application-Specific Hardware~~

The underlying goal to increase productivity affects many aspects of EDA. Interactive software tools are available on low-cost workstations to aid the design engineer. Interfaces to batch-oriented software analysis programs bridge the gap between the standalone user and the host-dependent user. Finally, hardware dedicated to and designed for specific computer-bound tasks is available to work with EDA design data bases.

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Simulation engines, for example, can complete a simulation pass in seconds, where it might have taken software methods hours to complete. Or. Jared Anderson, President of Valid Logic Systems, Inc., will address the use of application-specific hardware in the EDA market segment.

The EDA segment is not alone in its use of application-specific hardware. The CAD/CAM, computer, and graphics industries are applying more and more hardware dedicated to specific tasks. Examples of such hardware may include: pipeline processors; geometry engines; tiling engines; array processors; and co-processors, in addition to simulation engines.

~~DATAQUEST believes that this is a significant trend occurring within the entire CAD/CAM industry, and that it will be especially evident within the EDA segment.~~

~~PERSONAL ENGINEERING WORKSTATIONS~~

There is no question as to acceptance of electronic design automation systems as the solution to engineering design tasks. However, vendors and users agree that the average selling price of workstations must decrease in order to penetrate the electrical engineering community.

In addition to traditional host-dependent systems and standalone workstations, personal computers (PCs) are emerging as an important type of hardware in the entire CAD/CAM marketplace. DATAQUEST believes that PCs will play an increasing role in electronic CAD applications.

Based on current performance levels, PCs are best suited for applications that are less compute-intense, such as schematics drafting and physical layout. There are, however, several vendors that presently sell PC-based CAD systems that include forms of analysis and simulation. DATAQUEST believes that the very low-end market will be segmented based on performance and end use. We believe that low-end PCs will be used mainly for drafting purposes in large companies that use networking capabilities to communicate with the design automation process, or by small companies with less need for higher performance, higher cost, comprehensive systems.

We also believe that the low-end PC market will serve engineers and their requirements for analysis and simulation, as well as for schematic creation. To do this, however, will require different hardware technology than what is currently available from standalone workstation and personal computer manufacturers. The price of a standalone workstation must decrease to the $10,000 to $20,000 range, and the performance of the personal computer must increase to be competitive with that of a standalone workstation.

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 DATAQUEST recognizes this new hardware segment as personal engineering workstations (PEWS). We believe that PEWS will include medium- to high-performance graphics capabilities, local area networking capabilities, 32- or 16/32-bit microprocessors, application-specific integrated circuits (ASICs), and low to medium storage capabilities. PEWS will be configured to attain high levels of productivity, with the inherent ability to share resources. DATAQUEST believes that volume shipments of PEWS will begin in late 1984 or early 1985.

~~It is through PEWS, and their networking with personal computers, standalone workstations, and host-independent computers, that the mass of the engineering market will be penetrated.~~

~~NICHE APPLICATIONS~~

The EDA segment is growing at a phenomenal rate. Product introductions occur regularly and frequently. Different companies are entering the market. Product and company differentiation is becoming increasingly more difficult to recognize. DATAQUEST believes that the EDA segment is large enough to support many vendors, as the other CAD/CAM application segments currently do. However, we believe that product differentiation will begin to occur, as new entrants address niche applications.

The current EDA vendors and the companies expected to enter the market in the near future have similar product portfolios. Though they differ in corporate and marketing strategies, and actual product functionality, DATAQUEST believes that there is presently little true product differentiation between the major vendors.

~~We believe that the market will evolve to include companies addressing specific application or target market niches, that may include the following:~~

• Printed circuit board design automation: We believe that companies selling high performance standalone systems with schematics entry capability and an intelligent data base will expand their product offerings to include more design automation software.

.• Analog circuit design: Although analog circuit design is not the critical bottleneck in a design cycle, we believe that companies will begin to offer standalone and/or integrated systems with capability to incorporate analog and digital circuit design.

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Silicon compilers: We believe that silicon compilers are the next generation of design automation, and that they will become commercially used and accepted. Vendors may incorporate compilers into their current product lines, or new companies will emerge with compilers as their main products.

~~Semiconductor distributors: With the availability of standalone workstations and semicustom circuits, distributors will emerge as an important distribution channel for both types of products.~~

Software-only companies: The ability to accomplish a design task through application software is the true added value to any CAD/CAM system. We believe that more companies will emerge with software-only products as design tasks become increasingly complex. We also believe that more and more third-party software activity, mergers, and acquisitions will occur.

~~CONCLUSIONS~~

The electronic design automation segment of the CAD/CAM industry is complex, and undergoing constant change. DATAQUEST believes that it will play an ever-increasing role in the electronic CAD/CAM application segments and in the design of electronic products. We believe that successful vendors recognize the importance of truly integrating all phases of the design cycle with common data bases, interfaces, and systems. Engineers creating products in 1988 will enjoy the conveniences and productivity increases that EDA systems will provide them. It is these engineers that will make the EDA segment climb to more than $1 billion in 1988. It is these engineers that will create state-of-the-art electronic circuits, systems, and products through the use of electronic design automation tools.

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 17 -

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 18 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 EDA INDUSTRY OVERVIEW

~~• DEFINITIONS • MARKET SIZE • DRIVING FACTORS • TRENDS~~

~~O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 19 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD/CAM APPLICATION SEGMENTS~~

~~MECHANICAL AEC MAPPING ELECTRONICS~~

O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 20 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ELECTRONIC CAD/CAM SEGMENTS

~~EDA~~

~~10 PC~~

© 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 21 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 SPEC DESIGN & ANALYSIS FC GA SC PC VERIFICATION ANALYSIS TEST I MFG

O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 22 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 !l«a8!»SSSS!«Sl««S«Sl^^ IC CAD/CAM

~~• LAYOUT DESIGNER • PHYSICAL GEOMETRIES • CENTRALIZED CAD SHOPS • RJLL CUSTOM CIRCUT • TYPICALLY HOST-DEPENDENT~~

C 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 23 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 SPEC DESIGN & ANALYSIS iSpSiia l^ramm ^S PC .•^•,A;^."A''VXV,-.J K-iUtttvMi'-dM^ iSwMWai-tfci

~~IBfgp^iJPfpy^~~

~~ANALYSIS TEST i MFG~~

~~O 1984 Dataquest Incorporated Hay 8 ed.-Repcoduction Prohibited - 24 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 PCB CAD/CAM~~

~~ARTWORK ONLY • DRAFTING PERSONNEL • ARTWORK CREATION • CENTRALIZED CAD SHOP • STANDALONE OR HOST-DEPENDENT~~

O 1984 Dataquest Incorporated Hay 8 ed.-Reproduction Prohibited - 25 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ^isitzsssfiseiSttsiissiSsstiPSifss^^ PCB CAD/CAM

~~COMPREHENSIVE~~

~~• ELECTRICAL ENGINEER~~

~~• SCHEMATIC CAPTURE • DECEISTTRALIZED~~

~~• TYPICALLY STANDALONE~~

O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 26 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 SPEC

O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 27 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 EDA ii'sssssissssssssiiiifsisssiisiiss^^

~~• ELECTPICAL ENGINEER • LOGIC AND CIRCUIT DESIGN AND ANALYSIS • PHYSICAL LAYOUT • DECENTRALIZED • STANDALONE~~

O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 28 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 29 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 30 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 EDA MARKET ESTIMATES (Millions of Dollars)

~~1,500~~

~~1,000 ^~~

~~500~~

~~1983 1984 1985 1986 1987 1988~~

~~Source: DATAQUEST~~

O 1984 Dataquest Incorporated May 8 ed.-Repreduction Prohibited - 31 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 EDA QUARTERLY SALES DAISY, MENTOR, VALID (Millions of Dollars}

~~92% COMPOUND GROWTH 40 mmmmmm^^m^H^^^^^^^^^^^^m^^^^^i^^maam mmmmmttm^m^m^mm^m^^^m^^m^mm^^i^^m~~

~~30 — I " ' h ,"_• :^^^^^^^^^^^^^^^^^ ^^^^^^^^^ ^^^^^^^^^ ^^^^_^_L^ ' L- r .. ^M~~

~~• • -.U^——J.it<<»^ J- •••••••••••••••~~

~~Q4 Q1 Q2 Q3 Q4 Q1 Q2 03 Q4 01 1961 1982 1983 1984~~

~~Source: DATAQUEST~~

~~O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 32 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 1983 MARKET SHARE ESTIMATES~~

~~OTHERS DAISY SILVAR-LISCO~~

~~MENTOR VALID ^~~

~~Source: DATAQUEST~~

~~O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 33 -~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ELECTRONIC CAD/CAM MARKET ESTIMATES (Millions of Dollars)

~~1963 1964 1965 1986 1987 1988 • EDA D IC Q PC ii TOTAL~~

~~Source: DATAQUEST~~

~~O 1984 Dataquest Incorporated May 8 ed.-Repreduction Prohibited - 34 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 1983 ELECTRONIC CAD SYSTEM USAGE (Workstation Units)~~

~~FULL CUSTOM LOGIC DESIGN AND ANALYSIS GATE ARRAY STANDARD CELL~~

~~PRINTED CIRCUIT BOARD~~

~~Source: DATAQUEST~~

~~O 1984 Dataquest Incorporated May 8 ed.-Repreduction Prohibited - 35 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 1988 ELECTRONIC CAD SYSTEM USAGE (Workstation Units)~~

~~FULL CUSTOM/ \ X ^°A^"O ANALYSIS~~

~~GATE ARRAY~~

~~STANDARD CELL~~

~~PRINTED CIRCUIT BOARD~~

~~Source: DATAQUEST~~

e 1984 Oataquest Incorporated May 8 ed.-Reproduction Prohibited - 36 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 37 -

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 •^s»a^«sguisasis««aafeKa)B»aKtafti^^ DRIVING FACTORS

~~• CRCUT COMPLEXITY~~

~~O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited -38-~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 S9IwRiSjI^ DRIVING FACTORS '~~

~~• CIRCUIT COMPLEXrrY • SEMICUSTOM ORCUTTS~~

O 1984 Dataquest Incorporated May 8 ed.-Repreduction Prohibited -39- Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 DRIVING FACTORS •iXSISvSXilBiliSliiSiXuSufSSwSlSXa^^

~~• CRCUIT COMPLEXnrY • SEMICUSTOM CRCUTS • HARDWARE TECHNOLOGY~~

~~O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 40 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 TRENDS SSSSSSIeiSIS^SSSSSSSSSSSSSSSS^~~

~~• MARKETING/SALES~~

O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 41 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973

~~• MARKETING/SALES • FUNCTIONALITY~~

~~O 1984 Dataquest Incorporated Hay 8 ed.-Reproduction Prohibited - 42 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 QvSdVp^ TRENDS~~

~~• MARKETING/SALES • FUNCTIONALITY • INTEFFACES~~

~~O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 43 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ijjSJjiiWjjbbwaiiSBifeii^igiiaiSiSiiiwS TRENDS~~

~~• MARKETING/SALES • FUNCTIONALITY • INTERFACES • APPUCATION-SPECFIC HARDWARE~~

~~O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 44 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 TRENDS~~

~~• MARKETING/SALES • FUNCTIONALITY • INTERFACES • APPUCAHON-SPECFIC HARDWARE • PERSONAL ENGNEERNG WORKSTATIONS~~

~~O 1984 Dataquest Incorporated May 8 ed.-.Reproduction Prohibited - 45 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Conn pany /_1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 /Telex 171973 . PERSONAL ENGINEERING WORKSTATIONS~~

~~TIME~~

~~PRICE~~

~~PERFORMANCE~~

O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 46 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 »S!«$iimSSv&lSSSS&SiS!fS$S9» TRENDS tSSSSSiSf^SS^

~~• MARKETING/SALES • FUNCnONAUTY • INTERFACES • APPUCAHON-SPECIRC HARDWARE • PERSONAL ENGINEERNG WORKSTATIONS • NICHE APPUCATIONS~~

O 1984 Dataquest Incorporated May 8 ed.-Reproduction Prohibited - 47 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 SUMMARY ^PSS^SS^^

~~• TRBSDS • DRIVING FACTORS • MARKET SIZE • DERNTTIONS~~

~~O 1984 Dataquest Incorporated Hay 8 ed.-Reproduction Prohibited - 48 -~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, GA 95131 / (408) 971-9000 / Telex 171973 i O 1984 Dataquest Incorporated May 8 ed.-Repredaction Prohibited - 49 -

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973~~

~~Dataquest~~

~~WHAT'S ALL THE EXCITEMENT ABOUT OR HOW CAN AN INDUSTRY GROW SO FAST~~

~~Gerard H. Langler Vice President, Marketing Mentor Graphics Corporation~~

~~THIS BIO SHEET WAS NOT MADE AVAILABLE AT PUBLICATION TIME.~~

~~IF A COPY IS MADE AVAILABLE TO DATAQUEST, WE WILL MAIL IT DIRECTLY TO YOU FOLLOWING THE CONFERENCE.~~

~~Dataquest Incorporated ELECTRONIC DESIGN AUTOMATION FOCUS CONFERENCE May 10-11, 1984 Sunnyvale, California~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 What's All The Excitement About or How Can An Industry Grow So Fast

~~Gerard H. Langeler MENTOR GRAPHICS CORPORATION 8500 S.W. Creekside Place Beaverton, Oregon 97005-7191 (503) 626-7000~~

The market for Computer Aided Engineering products didn't exist in 1981, yet many industry observers predict a market size of over $1 billion by 1987! How can an industry grow SO fast? To answer this question we need to look at what is fueling the growth, and just as importantly the limits to growth. By focusing our attention on the limits to growth, we can determine whether all this euphoria is realistic or whether this market is just a high technology flash in the pan. The limits to growth for any industry or company are a direct function of both internal and external elements. Internal limits center on such things as cash position, manufacturing capacity, market coverage, management ability and something the Boston Consulting Group calls the "Maximum Sustainable Growth Rate" which is approximated by return on net assets minus the interest rate on debt times the debt to equity ratio. External limits are focused on such issues as degree of pent-up demand, economic cycles, capital avail ability and most importantly the return on net assets a customer can expect from newly acquired CAE systems. It turns out that in the short term (e.g., next one to two years) the limits to growth for the CAE industry are largely internal, not external! While the significant CAE players are public companies with excellent cash positions, and all have such large capitalizations that return on assets is not yet a meaningful number, the other internal limits are very much in play.

~~- 1 -~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dramatically lengthening semiconductor lead times put an artificial constraint on manufacturing capacity in the short term. The leading CAE companies have relatively small sales and support organizations (though growing fast) so market coverage limits industry growth at present. Finally, some CAE companies have the management experience and depth to handle over $100 million in sales, growing at over 100% per year. Some will struggle under such pressure. In contrast, there are few external limits to growth in the short term. There is a large amount of pent up demand, stemming from years of "shoemaker's children" phenomena among electronic engineers. We are in the middle of a strong upturn in the business cycle, led by capital spend ing.

Over the long terra, the more enduring factors of return on investment come into play. Given the very high software content to the vlaue added portion of the business, internal return on net assets is likely to be quite high. Coupled with the traditional aversion to debt by successful high technology firms, the maximum sustaninable growth rate for leading CAE firms is likely to remain quite high. The real limit to long term growth in this industry will be found in the return on investment customers ex perience when they purchase a CAE system. Most customers focus on measuring productivity gains to determine the value of their investment. While productivity gains are very important in terms of a shorter time to market serving to lower costs and improve market sha_re, there is a larger mostly unrecognized factor - price flexibility. If you give me a choice of getting to market three months early, or getting there on time with product in novation able to support a 5% or 10% higher price, I will normally choose the second option. CAE systems which not only promote productivity but also provide tools for in novation allow for such gains. This kind of thinking can lead to sophisticated an alysis which can prove beyond a shadow of a doubt that CAE is a good investment, often yielding 50% to 100% annual return on investment. However, even this may pale in com parison to the real return on CAE-survival. It is the

~~- 2 -~~

Dataquest incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 impression of many people (not just vendors) that failure to adopt CAE tools will spell economical doom for any firm competing in the fast moving electronics industry. With that attitude the growth potential for leading CAE firms is indeed spectacular. So, how fast can the industry really grow? Our best estimates are 200%-300% in 1984. 40%-70% per year in the 1985-1990 time frame and at least 25% to 35% per year during the 1990's. Along the way, you are going to see some interesting changes. First of all, during 1984 and 1985 some of the walking wounded will stop walking. Some will just go away, Others will merge or get acquired. For the long term leaders, look to Mentor Graphics, Daisy, one of the in strument companies (probably Hewlett-Packard) and one of the CAD/CAM companies (probably Computervision). Rapid growth and rapid change, it makes for a most exciting business.

~~- 3~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 a 01~~

~~•o o~~

~~Q. aj' •< o^ > p 2~~

~~O o 3 x> 01 WHAPS ALLTHl EXCITEMENT ABOUT? to 4 lit to ' I 7 or ? HOW CAN AN INDUSTRY GROW so FAiif?~~

~~CO Iu 3 Gerard H. Langeler Mentor Graphics Corp. o >~~

~~.u o 00~~

~~(O ~j CO o 0)~~

~~> w c cr en n.~~

~~> b (D W . § * O o LIMITS TO INDUSTRY GROWTH 3 •a o> 3~~

~~N3 to t A. Internal 3O) 1 Q." to Q. O -t "D (U B. External~~

~~01 3~~

~~o >~~

~~{>. o a> o(O o o~~

~~(D •vl CO o w ST n c (D~~

~~O O •5 O~~

~~> w c CT CO a £" > b INTERNAL (COMPANY) LIMITS TO GROWTH O o 3 •o D> 1. Short term rv) (D o u a a. Cash a CD -0 p> b. Manufacturing capacity c. Market coverage (» 01~~

~~C- o 0) 2. Long term a> O > to a. Maximum sustainable growth rate o (RONA- INTEREST ON DEBT) 00 b. Management ability o(O o o~~

~~(O -J CO o su~~

~~•o o~~

~~> V) c cr CO Q.' to'~~

~~> b~~

~~O EXTERNAL (MARKET) LIMITS TO GROWTH o 3 X3 a> 3 lO U3 O 1 1. Short term a! il^ a.~~

~~CO Bl 3 2. Long term o > a. RONA (eg. productivity gains & innovation gains) o b.Capital availability 00 oto o o~~

~~5. CD~~

~~CO --J CO o~~

~~c (D~~

~~O O~~

~~W THE MENTOR GRAPHICS GOAL c cr u Q.' 0)" "Enable its customers to compete better by improving > productivity and innovation in electronics design." b~~

o o 3 •o IVIanu- m 3 facturing tvj Test —*- (O Test o A ! ! Q. UI • I I a 1 • (S 1 TJ 0) Product Product Phtysical Manu Analysis — r Concept ~ Design " Layout facturing 1 1 1 w Ql 1 1 1

~~O (0 DRC/ _ ERC~~

~~(D O O O~~

~~(O w~.i O) 0)~~

CO +- 1^ (0 |i sg" CO 4.* ^ CO (/i «*- lU O) cc c 'w Q 0)0) «Q Q ^rf ss < O 0) (0 QE 0) 1 >« o l^ 1 O 21S 1 1 <<^ 1 • ojm w1 *co coo e •i^^M 75 cc ^ OCL r c CO < 1 1 o 1 • c 1 *^ X 3 O) 1 CO c Q. D) L—. m^m^ S a> p 0) CD E < c'5) Q 0) Ui CC 0)0) CO 1 CO c «Q 1 c o o CO o 0) = •- ?i •^B O 0) a o CO S "3 •c oc •o ^ 4^ QO o S oc o 0) ""S jfe- o a> 4ii* LU E3 CO Q. o o c k. OT3 CL o e CO 3 II * oco

~~- 6 -~~

~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 D Si c CD~~

~~3 O o -X •oo a DATABASE MANAGEMENT IS AS IMPORTANT CO c CT AS THE DESIGN PROCESS ITSELF a(A S'~~

~~> b o o 3 x> Q) 3 Manu~~

~~K3 CO I facturing O -J Test a 1 a 1 (D t) • Manu w facturing~~

~~O U> (D >O~~

~~•li. O 00~~

~~CO O O O~~

~~t-»o< u o (u~~

~~o o "O o~~

~~(D Q.~~

~~cn c cr CONCEPTUAL DESIGN 01 Q! Bl'~~

~~> The difference between engineering b and schematic drawing. (D 3 o o 3 •D a~~

(O o Manu a facturing a. 00 Test T) Test 0) 1 1 1 1 1 1 1 1 1 • Product Physical Manu w Analysis D> Design — Layout facturing O W (0 1 1 1 O 1 1 > ^0; • i DRC/ 1 1 ERG J j 1 1 .CK o i CO Relational DBMS i 1 Project Management olO o o User Interface S5. (D Ui 0) ' C C3 .±• ^-«t 3-Z m 3 0> c 5 (0 4i.« « (0 •*~ «»»

o - I- *-• 8 3 (0 tf) z . «•• r—'"^ (0 tc OE o c \L < ' 1 Li. 1 1 1 'sf^jîiîW • UJ y- 1 uj o L... li^^lw^ 1 âi^s 'jn c l- Q 0) DQ (/) O i Q E QC 1 O0}) -J CL C9 CO o U u. c c(0 0) LU 3 Q) «o o T3 O (0 S u CD I- 2 A- O ts 0) 0. o 1 0) •i* Z H • •MK c CL ^ LU P 0) C/> LL (0 LU LL 3 O LU

~~'• 5 Dataquest Incorporated, A Subsidiary of A.C. Ni, eisen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Te\e^ 171973 o w XJ c~~

~~o —. 2J. a a. > w Q. ANALYSIS TOOLS: WHAT ENGINEERS REALLY CARE ABOUT > b~~

~~3 o o 3 •a 0) •<~~

to O 1 Manu 1 S (S Test facturing tu —I 1 1 Test X- O • 1 1 • 1 1 • 1 (n 33SfiSs8E«I(E$»e3S»Kacot^'•.^ (U Product Product Physical Manu 3 C_ — O Concept Design ^^^^^^^^ Layout facturing CO tl 1 >O 1 (O 1 1 1 Ol 1 1 u DRC/ ERC o CO Relational DBMS 1

~~(O u•-J o SI~~

~~3 O o •3 o a PHYSICAL LAYOUT TOOLS CcO cr w Q! 0)' •< o Taking theory to reality. > b~~

o o 3 •o 0> 3 Manu lo CO o Test facturing Q2. M Q. Test (D 1 1 I -o 1 1 1 ' 1 1 I Product Produc1 t Manu Analysis —. Concept Design - — facturing m 3 1 1 1 1 1 • - o 1 1 > DRC/ ERG Relational DBMS g Project Management User Interface oCO o

~~CO u-J o a~~

~~"O o (D Q. > w~~

~~Q. TEST AND MEASUREMENT TOOLS: 0)' VERIFYING THE END PRODUCT > O~~

~~CO * :3 O o 3 Manu~~

~~M facturing (O "• 'TdiiK; ^ O 1 '. :;(({'•;••'••: M 1 1 Test Q. to 1 1 a 1 1 CD 1 1 1 "• 01 Product Product Physical Manu Analysis Concept Design Layout facturing 1~~

~~CO 1 1 01 1 1 1 3 • 1 DRC/ 1 o ERG > Relational DBMS~~

~~4S. Project Management O 00 User Interface~~

~~(O O O o o S. .tou c o o •5 o PROJECT MANAGEMENT TOOLS w c (T (O Q.' Ql' •-<^ Covering the 60% of the day O > P engineers spend outside "design engineering" 2 ro'~~

O o 3 •a D> Manu lo CO facturing •T Test ..»..-.«. o r- LJ I Test a I CL I I I o I _L • T] Product Product Physical 0) _i_ " Analysis -i- Manu Concept Design Layout CO facturing a> T c_ O I r » I o I > L-.. I ...J

~~Ji. O 00 I Project Management~~

~~£5.~~

~~USER INTERFACE TOOLS > w Accessing all tools through a unified portal~~

~~> b~~

~~o o 3 •o Dl Manu~~

~~M facturing CO O 3) Test ai. a. il^~~

~~O >~~

~~CD O O O~~

~~2. (B o w w .o c: (D~~

~~•a o~~

~~(D a~~

~~w c cr~~

~~0)' > b CUSTOMER RETURN ON NET ASSETS (D o O3 (Payback On CAE Systems) •o s> 3~~

~~to oto 1. Cost savings are nice~~

~~i 2. Time to market reductions are nicer~~

~~w ai 3. Price flexibility is the largest o > potential return (O~~

~~§ o o -i CD CO~~

~~(O ->l CJ o 5) n c (S~~

~~•o o~~

<» >a w I a > o WHAT TO WATCH FOR o o 3 •o o> 3 1. Some dead bodies

~~M to 2. Some acquisitions/consolidations O a 9> i? ,1 3. Continued exciting growtli •V Dl~~

~~CO lu o >~~

~~o 00 oto o o~~

~~CO --J CJ fi] S3 C w cCT (A CL '• Bj' •<~~

~~> P z 2! 0) WHO WILL LEAD CAE? (D 3 0 0 3 •D0> •3<~~

~~•TI ^ S) —t ?r D Mentor Graphics Hewlett-Packard Computervision <' (0 w ai Daisy Systems Tektronix Calma (/0> a> Racal-Redac 0 GenRad~~

~~CO 01 u—A Intergraph 3: 30~~

~~1 CO 0 0 0 H (D (D X~~

~~-NI~~

~~CO t. D)~~

~~O O •3 o~~

~~(D a > 03~~

~~> P Z eo' HOW FAST CAN CAE GROW?~~

~~O o 3 •o Dl~~

~~Ivj ° * 1984 ^ 200%-300% Q. 00 (D 1985-1990 ^ 40%- 70% 0) 1990*8 *- 25%- 35% w (U~~

~~.b. O 00~~

~~to u-J Dataquest Dataquest~~

~~APPLICATION-SPECIFIC HARDWARE~~

~~Dr. Jared A. Anderson President and Chief Executive Officer Valid Logic Systems, Incorporated~~

Dr. Anderson is a founder of Valid Logic Systems, Inc. Prior to founding Valid, Dr. Anderson was President, Director, and Chief Executive Officer of Two Pi Corporation. He was also the inventor of Two Pi's 32-bit IBM-compatible minicomputer. Dr. Anderson personally negotiated the sale of Two Pi to N.V. Philips in 1976. Previously, Dr. Anderson was Vice President of Research and Development at Computer Machinery Corporation, a manufacturer of shared-processor data entry systems. Dr. Anderson was also President and co-founder of Decision Inc., a supplier of disk and tape controllers for Data General minicomputers. Before forming Decision Inc., Dr. Anderson headed major research projects involving multidisciplinary teams of scientists, engineers, and contract personnel. Dr. Anderson holds a Ph.D degree in Physics from the University of California at Berkeley, where he participated in a research project that culminated in the awarding of the 1968 Nobel Prize in Physics to Professor Luis W. Alvarez.

Dataquest Incorporated ELECTRONIC DESIGN AUTOMATION FOCUS CONFERENCE May 10-11, 1984 Sunnyvale, California Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 lU

~~o DESIGN AUTOMATION EVOLUTION~~

~~CO c CT 80's~~

~~> b~~

~~in o B o 7G's Design Workstation •o 0) 60*8 CAD Layout M D) Minicomputer Manufacturing MIHMIMMnBHIBHMMBt ' ^Mainframe ©atch~~

to u •ataqueajaflTwe.t ' Tyi 1^ nhiStif Ham urmoi ban JOSS, UAyyi3l 1(408) 971-UOOOi Telex 171973 (D LO ID Dataquest inoorpofated, A SubaidIary ot A.C. Nielsen Company /1290 l=liddef Park Drive t San Jose, CA 95131 / (408) 971-9000 / Telex 171973 1 c a f CAE CAPABILITES i Cl I o >l C Workstation

~~Qraphics~~

~~Analysis~~

(O •^4 lO Dataquest Incofporated, A tiubsidlafy ot A.b. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 . o u w .a c o iro •o• t u c a APPLICATION SPECIFIC WORKSTATIONS ti A o > f

~~c o 3 '^T^MMR i|m,^tni0lufail€tc. ••~~

~~^^.f^f^*^. to 'S5!;'-'*!;#^^t«5^,W'^?r^^Pf,M^^§>^»*! UD fM-^Si~~

Q ^iBTflflBBmiSpBpiillrnprDvaTnents Qalned By O. (S T) the JMdition i^|yp3riioation-S>peciflc Hardware SI •-,•'.*•;• • .'T-'^-'?-; '-•'T V-.'-^'^^^y^ j^.>™.:^^^i-i^,:^'^T>" S^ Ss.^'; < Q lU

~~O An Appllcatlon-Speclfic Workstation Can Execute the PCf> o > Target Program Faster than a Large Super Computer~~

~~.c^ o CO o(O S/\LID o £-10292 o~~

~~5. (D X~~

~~-vl (O ~J CJ . o 111 § c y^~~

~~tj c c EXAMPLES: ASW"s s~~

~~c )~~

~~Speed Improvement of 600 x in Simulation Speed f c 5 •c Valld's Realfast"'~~

~~JH 50 • (° I T Dl Functional Enhancement JUlowIng Hardware Modeling~~

~~o u (P VallilliiRealchii a > g~~

~~r,.;G2!)'l S/\LID ^ sV~~

~~CO -vt o> Dataquest I o iro •5 o u p > a SIMULATION PROBLEMS~~

9 Bl -5 wr-^w^.i-K o r iloddsare c " ••* r*'* ' ^- 1 ^ • • fti,TB^ •-•-•• - •^.•^.f • »•-- o 3 •o IMar^^' -5.:-v/ TooOoniplex V .^l+"H:/!f!;^^^!î^•.-;i.•îL•:i-'li*-rc^^^^^>^.•'«v 3 .^^^.•i:.^.^;?-/^^.^- •..• •. V r-*'??^' •*'>-^ • •'--•^•i -. '4b^MMiWMyiHBO S I 31- Q M \t^r ^ M^ ^^^r^-^^^^sm^^^ îrrn^r^^f^'^^^^^^Pf^^^Kî^îf^^ T Bl

~~oJ^^rl^^ nr-^rr^^^^g^^^s-^^ ;^:^*-^r ^^S^f^^^^^^ Simulation is U ^uidpowj lloolisiow o > JUiipmft Design j (O Ol ;S»£^(^j££^:U£~^ *^,':it^.iF^'A~~

~~S S/ILID o(O o o~~

~~-si (O •-J CJ 1^ - Dataquest Incorporated. A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 a 01~~

~~<" J-~~

~~o o •oo~~

~~> SIMULATION PROBLEMS w a u' o > Poardljevel O~~

~~wjp>T-fJ..-.^>^*y~~

~~3 o o 3 ^^ '^Ma^lP iHiMHlP; ^#J#%l5|Tirj(i^||# ^^ #'i^Hundred ''~~

~~CO ?^" •'."• "J'i/'.':':^"^''• •^^•^^'J • -^^': ^. •"••••• • o ^ '; -• • .-i;- •--"'•-.^-^ i'-:: •.;.• • r, ' • .' '^ -. ! a" CD •O~~

~~V^\LID~~

~~CD O O O~~

CD -^J CJ Dataquest incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o V w 1 REALFAST™ > w SIMULATJON ACCELERATOR Q. 5j' -3 o^ > o Purpose: Speed o o 3 "O 01 D Allowing Laiii&r, f

~~(O O d. (Ti More Detailed and Q. ID imulations w u o >~~

~~B40234 V\UD oCO o o V o~~

(0 J— ^ E REALFAS"r" > w HOW FAST IS FAST? c cr (0 Q: 0)' o_ > o Executes 500,000 Circuit Evaluations Per Second • Up to 1,000,000 Circuit Elements (Primitives) o o 3 •o 0> 3 Otiier Accelerators ro - Comparable Speeds - Five Times the Cost

~~B40235 S/^LID~~

~~CD Xfc_ I •zc REALFAST™: WHY IS THE COST u t a PERFORMANCE SO MUCH BETTER? c B~~

~~Simple Architecture Allows for c c T. ta High-Speed Execution 1 - RISC (Reduced Instruction •a Set Computer) u Wide Memories Allowing High o Information Bandwidth >~~

~~t- o CB 840236 \^U\> I~~

~~im -J o w .Q C ir i\ REALFAST™ in c CT HOW DOES IT WORK? (A a! 0)'~~

~~> Schedule p Z. n Advance Time Dl~~

~~8 go BttlMUellemoiy ci kO Q. (D I T3 lu Evaluates Primitives Keeps Tracic of Time~~

~~en (U 3 » >• a o >~~

~~840236 S^LID § CO u•-J CD - 22 -~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 - 24 - Dataquest IncorporateIJ, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 s V ir "^

~~o -1~~

~~> w c tr REALFAST™ VREALCHIP o 2 ni' n o 3 •o Dl ^ 500,000 Events/Second Makes Sequential Fault Simulation Practical REALCHIP™ CO 111 o Provides Simulation Models of~~

~~>O Complex VLSI Parts Ol CJ o 00 oCD V\L! o o ^V t c c TYPICAL TARGET APPLICATIONS c • Logic Simulation Electrical • Circuit Simulation c c •Routing~~

~~•3-0 Modeling Mechanical •Simulation I: •Thermal Analysis c~~

~~1 0 I ^Finite Element Analysis (NASTRAN) Structural •Loading Calculations d r Hydrodynamic ^ Navier^tc^es Equation~~

~~a (i: V\LID iV - 28 - J Dataquest Incorporated, A suosIdia^V "1' !^-^. NIelsen Uompany I 1290 Rldder Harh Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 .KEfe^OK:' Dataquest~~

~~DESIGN INTERFACE STRATEGIES~~

~~Kent Jaeger Manager of Design Interface California Devices, Inc.~~

Mc. Jaeger joined California Devices, Inc., in July 1983 as Area Marketing and Major Programs Manager. His present responsibilities as Manager of Design Interface include management of all product and technical marketing areas. Prior to joining California Devices, Inc., Mr. Jaeger worked at Texas Instruments where he was responsible for major programs in both the Bipolar Gate Array and Standard Cell product groups. Later he had management responsibility for CMOS Gate Array progreuns. Mr. Jaeger received a B.S. degree in Industrial Distribution from Clarkson College of Technology in Potsdam, New York.

~~Dataquest Incorporated ELECTRONIC DESIGN AUTOMATION FOCUS CONFERENCE May 10-11, 1984 Sunnyvale, California~~

Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 D Si. 5) .o c (D 3 n o ••o CyVLIFORNIA DEVICES INO. w CDI ac

~~> b o o 3 •D • INTRODUCTION TO CDI (O o 3) a • DESIGN AUTOMATION CAPABILITIES • INTERFACE ALTERNATIVES _(D >O~~

~~o 00~~

~~•vl CO I~~

~~4 COf OyV.I-IFORNI.A. DEVICIBS INC. > w CDI EVOLUTION (140) DESIGN I •* a AUTOMATION~~

~~s ::• MANAGEMENT o o 3I •D 0) SERVICE (47) 8 SERVICE DLM SERIES I 5- HCS SERIES HCS SERIES (8) (12)~~

~~w HC SERIES HC SERIES HC SERIES HC SERIES (u {4) O (0 _(D O> CDI SERIES CDI SERIES CDI SERIES CDI SERIES CDI SERIES # EMPLOYrEE S (1)~~

~~DESIGN DESIGN DESIGN DESIGN DESIGH DESIGN~~

~~1978 1979 1980 1981 1982 1983 ID -e- o 2. (D~~

~~-J CO I oi o —1 •• o— . OJ > CAL.IRORNI/V DEVICES INO. CD CDI c: cr 9. ai~~

~~> MANAGEMENT TEAM o~~

~~01 z> O G. GIBSON o 3 •o PRESIDENT 0) :j~~

~~10 ' TI A.LEUPP A. CAPPUCCiO B. CAPARELLI B> VICE PRESIDENT VICE PRESIDENT VICE PRESIDENT I OPERATIONS FINANCE MARKETING/SALES~~

~~O I 1 > M. GULLETT J. BUNIK L. ROFFELSEN R&D STRATEGIC ENGINEERING SYSTEMS o 00 t (S~~

~~CD o~~

~~o o •3 o~~

CD CALIRORrsllA DEVICES liSJC_ a (/> cor c CDI PRODUCTS CT a> a. 0)' PRODUCT FEATURES TECHNOLOGY > p CDi HIGH VOLTAGE 6,3/^ METAL GATE CMOS 2 RADIATION TOLERANCE SINGLE LEVEL METAL O lOOrSOO GATES o 3 •o 22-72 l/O'S 3 MHz CLOCK N3 lO I O iC^ LSTTL COMPATIBLE I HCS 4/1 SILICON GATE CMOS I 300-1780 GATES SINGLE LEVEL METAL % 40-92 l/O'S o 20 MHz CLOCK w a> =1 DLM 6 TTL LOAD DRIVE 3/JC SILICON GATE CMOS SMALL SILICON DIE DUAL LEVEL METAL O > SIZE 225-10, 260 GATES o 00 30-156 l/O'S 52 MHz CLOCK §

~~01 •K o S. n c CD (A O O ? QQI CALIFORNIA PEVIOES INC~~

~~« c IT (0 Q.' 5' DESIGN AUTOMATION -< o > b • HISTORICAL PERSPECTIVE (D o O3 •o - MAJORITY OF CAPABILITIES DEVELOPED FOR lu INTERNAL USE (IBM, BELL LABS, Tl)~~

~~HAVE SPAWNED MORE EFFICIENT COMMERCIAL CAPABILITIES w o to CD >O WORKSTATIONS~~

~~o 00 TIME lO & o <2. (D~~

~~-vl u o 0)~~

~~•o o~~

~~> CAL-IRORNIA DEVICES INC. CO c CDl cr~~

~~> O DESIGN AUTOMATION OBJECTIVES~~

(D o o I • PROVIDE TOOL ACCESS AND CAPABILITY TO ENSURE LOGIC FUNCTIONALITY, ELECTRICAL M § ' INTEGRITY, AND TESTABILITY 3) a\ d. S- ' i(D • STRUCTURE DESIGN SYSTEM TO ALLOW WIDE VARIETY OF DESIGN INPUTS AT VARIOUS POINTS IN THE DESIGN PHASE

~~o VALIDATE AND VERIFY EACH ELEMENT > OF DATA REGARDLESS OF ENTRY FORMAT OR POINT~~

~~I i~~

~~ID o~~

~~O O -, x> o -1 .flPJc/VLIFOFtNIA DEVIOES IN Ifa a. W c tj u> a u'~~

~~> p 2. ADVANTAGES OF COMMERCIAL (A (D o SOFTWARE TOOLS o 3 •o D) 3 CO I • RAPID ESTABLISHMENT OF FULL CAPABILITIES 2 -J a °- I ' • INCREASED SUPPORT AND CONTINUED DEVELOPMENT~~

~~w • MLRGE USER BASE FOR RAPID DEBUG a>~~

~~O 9 TECHNQLOGYTNDEPENDENT > • PORTAelUTY FOR IN-HOUSE CUSTOMER USE o.t^ 00~~

~~to 8- o 2. fo~~

~~CO CJ D~~

~~n c a>~~

~~3 O o "-•1 o (D Q.~~

~~C C>eVL-ll=ORNIA. DEVIOES INC. o- u> CDI a. B' -1 •< o > O DESIGN AUTOMATION CAPABILITIES~~

o DESIGN ANALYSIS AND DEVELOPMENT: o 3 ! - DAISY LOGICIAN WORKSTATIONS 8"^ I - GENRAD'S HILO SIMULATION SYSTEMS 33 00 S. S- ' (PRIME 750 VAX 11/780) CD T) 0) -1 AUTOMATED PLACE AND ROUTE: O - DAISY GATEMASTER WORKSTATIONS CO DJ - SCIENTIFIC CALCULATIONS SCIMEDS o SYSTEM (VAX 11/780) > AUTOMATED VERIFICATION: - PHOENIX DATA SYSTEMS' MASKAP PROGRAM - APPLICON DRC PROGRAM oCO o o

~~•^ 0)~~

~~CO CJ a w~~

~~a. C/VL-IRORNIA DEVICES IN > CDl w~~

~~a 0)'~~

~~> O SEMICUSTOM DESIGN CENTERS O Io / HISTORICAL PERSPECTIVE N3 TREND TO OPERATE AS PROFIT AND LOSS CENTERS~~

~~o.b . CO~~

~~•-( D Si nw c 0)~~

~~Ia. ^^1^ C/VLIFOF^IMiA. DEVICES ifsiO,~~

~~W a cr CO Q: 0)' >-<1 o > b~~

~~(D o o 3 •o 01 SEMICUSTOM DESIGN CENTERS 3~~

~~IV) (O o • WITH THE EVOLUTION OF D.A. CAPABILITIES (HARDWARE a! CL n> & SOFTWARE), ESTABLISHMENT OF LARGE OF REGIONAL 1) 0) DESIGN CENTERS NOT ESSENTIAL FOR SUCCESSFUL~~

~~CO COMPLETION OF SEMICUSTOM DESIGN 0) o u> _(D o >~~

~~u> CJ o 01~~

~~o o •o3 w-t ^^^^^^^^ > CAl-IFCDRNIA DEVICES I O) c~~

p z 5" - CAPACITY TO TRAIN 48-60 PEOPLE PER MONTH TJ DJ - APPLICATION AND DESIGN ASSISTANCE AVAILABLE FUTURE PLANS W a> 3 - EXTENDED REGIONAL SUPPORT THROUGH NATIONWIDE o DISTRIBUTION NETWORK > - SUPPORT INCLUDES APPLICATION ASSISTANCE, PRODUCT SPECIALIST, TRAINING, DOCUMENTATION BANK, DESIGN TOOLS AND ASSISTANCE
(D O O
-A to o oT -CcI a
o o o 5 CALIFORNIA DEVIOES INC. CO _a CDI > CUSTOMER w INTERFACE METHODOLOGIES c cr LOOK: DESION DOCUMENTATION (» LDGK SCHEMATIC GEN. MACRO UBRART a fu' TEST PATTERN GEM. TRAINING DEVICE SPECIFICATION LEVEL I ACCESS TO TOOLS
> LOGIC SCHEMATIC. FUNCTIONAL TEST PATTERNS. DEVICE SPECIFICATIONS LOGIC CONVERSION T MINIMIZATION O CUSTOMER APPROVAL CONVERTED SCHEMATIC LOGIC VERIFICATION TIMING ANALYSIS TESTABILITY ANALYSIS o MACnO LIBRiftRY FAULT GRADING o 3 •o LOGIC CONVERSION REVERIFY LOOK B> VERIFIED NET LIST. TEST PATTERNS. DEVICE SPECIFICrtTioNS MINIMIZATION TIMING ANALYSIS TESTABILITY ANALYSIS ro LOGIC VERirlCATION FAULT GRADING (O TIMING ANALYSIS - o RHin. Hl^ln« 3 LEVEL II d TESTABILtTY ANALYSIS Q. FAULT QRAOINQ n> T) D)
OT D) 13 CFNTFR LINF RATA BASF AUTOMATIC PLACE S ROUTE C- AUTOMATIC EXTRACTION o PARTITIONINQ i" OF ROUTING STATISTICS O PLACEMENT LtiOOT DATA BASE. NET LIST TEST PATTERNS HEVICE SPECIFICATIONS PERFORMANCE VERIFICATION > ROUTING LEVEL III LAYOUT VERIFICATION PROTOTYPE SHIPMENT DESIGN RULE CHECK PROTOTYPE APPROVAL TEST PROGRAM GENERATION PC TAPE GENERATION .tv PRODUCTION PURCHASE ORDER O PRODUCTION RELEASE MANUFACTURE S TEST a to •i^ I o(O o o
(D
(O ~J U D w m J3
o o •—o1 o a a. > w L-IRORNIVX DEVICES INC. 01 —1 >< C0( o_ > b INTERFACE OPTIONS
a> :> LEASE WORKSTATION CDI MAIN FRAME O o 3 •o IB to too
Q. a CD -^ TJ 0)
5" 3 o >
O S CO -vl
o
CO a o
•o o CDi C/VLIFORIMIA DEVICES INO, Q. 0)' •2 o_ > DLM MACRO LIBRARIES b
(D 3 MACRO LIBRARIES ARE AVAILABLE FOR THE o FOLLOWING DESIGN SYSTEMS:
N> I - DAISY LOGICIAN AND GATEMASTER WORKSTATION . d. - MENTOR WORKSTATION 5" -1 - VALID LOGIC WORKSTATION X- O - GENRAD'S HILO SIMULATION SYSTEMS W lu
C_ O u - TEGAS SIMULATION SYSTEMS n> O> - FUTURENET SCHEMATIC DESIGN SYSTEM - ADDITIONAL COMMERCIAL AND SPECIFIC CORPORATE o.b. 00 DESIGN LIBRARIES UNDER DEVELOPMENT oCO o a «u a c (0 u> o o g
> OAHFORtsll/V DEVIOES INC. w COI a. LEVEL II SI' -3 g_ > DAISY b UENTOR
o o 3 •o VALID u FUTURE NET R E lO F CO FLOPPY DISK V R o A XI Ul E L E B 5: R A ^ V ^ a I (D I Y E —t l"^ î^ •D F o R Y U 1 'A D T F T w D> A Y E 3 C- T O u CD A O > •
MAGNETIC TAPE
9-
to ~* a o w ST c (D CO
O O •o O
CD a. w c LEVEL III IT CO 5! B)' o^ DAISY > QATEMASTER b R E CO ^P V o E 3O •o MENTOR CAD! R a> W I F FLOPPY DISK F A IVJ CO Y o en Q! D "I a I CD A -^ '»s •D 01 T -1 5r A D ST T E W (PHYSICAL) o> 3
o >
4!>. O MAGNETIC TAPE 3 U3 •vl oCO o o
CD CD
-J CO -si Dataquest Dataquest
THE SYNERGY OF ASICs AND EDA
Kenneth V. McKenzie Associate Directorr Semiconductor Industry Service Dataquest Incorporated
Mr. McKenzie is Associate Director of DATAQUEST's Semiconductor Industry Service. He is responsible for all research activities on semiconductors and related products. His other duties include internal data processing coordination for the Semiconductor Industry Service and research into specific end-user markets. During Mr. McKenzie's 14 years in the electronics industryr he has held management positions in both design engineering and marketing. His most recent position was as Marketing Manager at Zilogr Inc. Prior to thatr Mr. McKenzie was Marketing Manager for 8-bit microprocessors at Intel Corporation.
Dataquest Incorporated ELECTRONIC DESIGN AUTOMATION FOCUS CONFERENCE May 10-11, 1984 Sunnyvale, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 |4^4$I^Sf%§rJt THE SYNERGY OF APPLICATION-SPECIFIC INTEGRATED CIRCUITS AND ELECTRONIC DESIGN AUTOMATION
KEN MCKENZIE ASSOCIATE DIRECTOR SEMICONDUCTOR INDUSTRY SERVICE DATAQUEST INCORPORATED i pS^3IJîpKW^m!!!J!!;!3S!pp ij;:/ iiJKl.OT:
1 - 1984 Dataquest incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 >S!S*T*R!S!^!SHl<^<^:iS!
• CUSTOM • GATE ARRAY • STANDARD CELL • PROGRAMMABLE ARRAY LOGIC (PAL®)
*: , - 2 - ® 1984 Dataquest Incocporated-Reproduction Prohibited
•i
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 WHY PREVIOUS ASIC GENERATIONS FAILED
ISa<«S
• LACK OF ECONOMICAL USER DESIGN TOOLS • USER/VENDOR LOGISTICS • LIMITED SECOND-SOURCING • PRIMITIVE HIGH-PIN-COUNT PACKAGING
- 3 - © 1984 Dataguest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ASIC TECHNOLOGY--TODAY
• EVOLUTION TO STRUCTURED ELEMENTS • CMOS FOR LOW POWER AND SPEED • CHIP CARRIER AND PIN-GRID PACKAGING • TREND TOWARD INDUSTRY STANDARDS
- 4 - © 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MARKET OVERVIEW WORLDWIDE INTEGRATED CIRCUITS (BJllions of Dollars) 60 50 40 30 20 10
J—1 79 '80 "81 '82 '83 '84 '85 '86 87 '88 '89 Year
Source: DATAQUEST
- 5 © 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MARKET OVERVIEW WORLDWIDE IC AND MEMORY (Billions of Dollars)
'79 '80 '81 '82 '83 '84 '85 '86 '87 '88 '89 Year . IC MARKET .. MEMORY MARKET
Source: DATAQUEST
6 - © 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen CoITipany /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MARKET OVERVIEW WORLDWIDE IC AMD MICRO (Billions of Dollars)
'79 '80 '81 '82 '83 '84 '85 86 '87 '88 '89 Year . IC MARKET .. MICRO MARKET
Source: DATAQUEST
-7 - © 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MARKET OVERVIEW WORLDWIDE IC AND ASIC (Billions of Dollars)
'79 '80 '81 '82 '83 '84 '85 '86 '87 '88 '89 Year . IC MARKET •• ASIC MARKET
Source: DATAQUEST
- 8 - (g) 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MARKET OVERVIEW WORLDWIDE MICRO AND ASIC (Billions of Dollars) 15 p 10
0 '79 '80 '81 '82 '83 '84 '85 '86 '87 '88 '89 Year . ASIC MARKET .. MICRO MARKET
Source: DATAQUEST
- 9 -
1984 Dataquest incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MARKET OVERVIEW WORLDWIDE ASIC MARKET 1983 SHIPMENTS ($)
GATE ARRAY FULL CUSTOM 21.3% 75.5% STANDARD CELL 3.2%
Source: DATAQUEST
- 10 -
@ 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MARKET OVERVIEW WORLDWIDE ASIC MARKET 1989 SHIPMENTS ($)
GATE ARRAY FULL CUSTOM 25.2% 57.2%
STANDARD CELL 17.6%
Source: DATAQUEST
- 11 -
© 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 WHY ASICS ARE WINNING- -TODAY
• POWERFUL INEXPENSIVE MICROPROCESSORS • SOPHISTICATION OF USERS • ASIC-FOCUSED START-UP COMPANIES • ACCEPTANCE OF CMOS VLSI • SHORTAGE OF TTL "GLUE"
- 12 - © 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 THE IMPACT OF CMOS VLSI
HIGH-SPEED VLSI MUST BE IN CMOS • PACKAGING TECHNOLOGY • POWER DISSIPATION • RELIABILITY
- 13 - © 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 i««SSK«SKHKS»!KffiHHSS{«n«^^ THE IMPACT OF CMOS VLSI
^'Sp3Isi£^k?-%^HistaK^
LOW POWER LOWERS TOTAL SYSTEM COST • SWITCHING POWER SUPPLY $ • COOLING FAN $ • FEATURE CONTENT VS. SYSTEM FOOTPRINT • NOISE IMMUNITY AND SERVICE
- 14 - © 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 *S»!HS8SniHHSH»?»SKSH!^^ EDA DRIVEN BY VLSI ADVANCEMENTS Tsmsisssi^
• MICROPROCESSORS • MICROCONTROLLERS • MICROPERIPHERALS • DYNAMIC RAMS • ASICS
- 15 -
© 1984 Dataquest Incorporated-Reproductlon Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MICROPROCESSORS - - TODAY
8-BIT • MATURE (1972) • GIVING WAY TO 8/16-BIT • CMOS FAMILIES DELAYED • FUNCTIONAL INTEGRATION
- 16 -
© 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen CoInpany /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 «««««<«« MICROPROCESSORS- -TODAY
16-BIT • 4 PLAYERS--2 CLEAR WINNERS • GOING 16/32-BIT • CMOS--THE HOT TOPIC • FUNCTIONAL INTEGRATION
- 17 -
© 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MICROPROCESSORS- -TODAY
»SHSiS!SSSiS»«««SS!SS!Si»SiK-«^^
32-BIT • LITTLE REAL ACTION • SUPERSETS OF EXISTING 16-BIT MICROS • VIRTUAL MEMORY SUPPORT • CHIP CARRIER AND PIN-GRID PACKAGING
- 18 - © 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MICROPROCESSORS--THE FUTURE
• HIGH-SPEED (20 MHz) • TOTAL CMOS FAMILIES • TRUE 32-BIT MPU • ON-CHIP MEMORY MANAGEMENT • PARALLEL PIPELINED ARCHITECTURES • APPLICATION-SPECIFIC MRUS
- 19 - © 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 DYNAMIC RAMS--TODAY
64K BIT • MANY SUPPLIERS • U.S. GAINING MARKET SHARE • PRICING FLAT TO UP • SUPPLY/DEMAND BALANCE--YEAR END
256K BIT • LITTLE NON-JAPANESE ACTION • HIGH PRICES ($30 ASP-1984) • VOLUME IN 1985-1986
- 20 -
© •• 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 DYNAMIC RAMS--THE FUTURE ^î«iSiixyîiSisiS6ixixiz<^<}!iSx:f>!six^ss!Xi^
• SHORTER PRODUCT LIFE CYCLES • CMOS • X4 AND X8 ORGANIZATIONS • MINIMAL AREA PACKAGING (MAP) • HIGH SPEEDS (50-100 NS)
- 21 - © 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 •«!KSKKSHKHffiK»:SS«SiSK^^ MICROCONTROLLERS--THE FUTURE
• CMOS FAMILIES • HIGH-PIN-COUNT PACKAGING • 16/32 BIT ARCHITECTURES • ANALOG I/O • ON-CHIP LAN • APPLICATION-SPECIFIC MCUS
- 22 - © 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 MICROPERIPHERALS--THE FUTURE x^saSfisssxsiiiSf
• LONGER PRODUCT LIFE • MPU INDEPENDENCE • CONFIGURATION LANGUAGE • RELATED FUNCTION INTEGRATION • ON-CHIP FIFO AND PROCESSING
- 23 -
(g) 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 APPLICATION-SPECIFIC ICS--THE FUTURE xasm
• COMBINED GATE ARRAY AND STANDARD CELL • MRU. MPR, MCU AS STANDARD CELLS • SOFTWARE DATA SHEETS • INDUSTRY STANDARDS • COMBINED DIGITAL AND ANALOG FUNCTIONS
- 24 -
© 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD TOOLS "TODAY
• MULTIUSER • SHARED RESOURCE • STAR OR CLUSTERED COMMUNICATIONS • TASKS VS. USERS--PERFORMANCE LIMITS • COMPLEX MULTIUSER SOFTWARE
- 25 - © 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 •SHKSKSHSnSSKHHSSaKSISRH^^ CAD TOOLS - - THE FUTURE vtiiOijfij^vilvivivititvOv^vduovîiSiiSiiSi^^
• SINGLE DEDICATED USER • LOOSE-COUPLED 16-BIT AND 32-BIT MRUS • HIGH-SPEED LAN COMMUNICATIONS • STANDALONE/MULTITASKING • OPERATOR-SENSITIVE SOFTWARE (NLI. AD
- 26 - © 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 S!?$i?S!S««^ EDA WORKSTATIONS--THE FUTURE
• WORKSTATIONS THE SIZE OF TODAY'S PCS • INTERNAL 1MB TO 4MB RAM • TRANSPARENT VIRTUAL MEMORY • VOICE I/O • LOCAL OPTICAL DISK
- 27 -
© 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAD TOOLS- -THE FUTURE
• CAD TOOLS SMALLER THAN TODAY'S PCS • INTERNAL 1MB TO 4MB RAM • TRANSPARENT VIRTUAL MEMORY • VOICE I/O • LOCAL OPTICAL DISK
- 28 - © 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 !SSSKSSI«SHSH«Sn«fl»«!iK!^^ EDA AND ASICS lv*5&SK&-J!SH2-a»t^«S««»!?-
• ASIC GROWTH DRIVES EDA FEATURES • EDA "EASE-OF-USE" OPENS ASIC MARKETS
THANK YOU--VLSI
- 29 - @ 1984 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Dataquest
USER PANEL DISCUSSION
Bob Burkhardt Manager/ Bipolar Design Automation Advanced Micro Devices Incorporated
Mr. Burkhardt has been Manager of Bipolar Design Automation at Advanced Micro Devices, Inc., for the past 2-1/2 years. He has been employed at AMD for more than 10 years. His previous positions at AMD include Design Engineer, Test Engineer, Project Engineer, Project Engineering Supervisor, General Foreman (Test Area), and Manager of Test Engineering. Mr. Burkhardt received a B.S.E.E. degree from California State Polytechnic University. .
Dataquest Incorporated ELECTRONIC DESIGN AUTOMATION FOCUS CONFERENCE May 10-11, 1984 Sunnyvale, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ELECTRONIC DESIGN AUTOMATION
I. APPLICATIONS
II. WHY AUTOMATE
III. EXPECTATIONS
IV. FUTURES
- 1 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 I. APPLICATIONS
- CUSTOM INTEGRATED CIRCUITS
- LSI/VLSI COMPLEXITY
0 SYSTEM DESIGN
0 LOGIC DESIGN
0 MASK LAYOUT
0 VERIFICATION
0 PROTOTYPE DEVELOPMENT
- 2 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 II. WHY AUTOMATE
- COMPLEXITY
- Too CUMBERSOME
- MANY BENEFITS
- 3 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 I '^':^?^^^3:^
"•*!,..i--,- •*"•• • HW^^^ -^ J^ T^«-'*cTt.,.i •jn» ^ •rafT .—"Thv SC3ra
—•"-- * . L-*-" .'".:"•' '''^-^ "•' - •-• -% —-'-..'^^ -5 "^ *^ ^ .-». - * L' •*-•—•• -v • -^ti- *;—r-^ -.V- * ^ - ^
1-: .'•
•AW^^^'Î' •'. ^*^- i, ttm *mf^'^. ... -.-^ Vt' fte AmZ8163/AmZai 67, which measures 15 x 19 feet in layout, will only measure 155 x TS5 mils on the finished chip.
- 4 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 - WE DETERMINED OUR OLD APPROACH WAS TOO CUMBERSOME.
0 WE WOULD PLOT THE ENTIRE CIRCUIT AT 1015X.
0 WE CHECKED THE CIRCUIT BY HAND ON THE FLOOR.
0 WE TOOK DAYS TO DO DRC CHECKS ON OUR SMALL COMPUTER.
CHIP SIZE (MIL) 280 X 280 280 X 280 SCALE 1015X 254X RESOLUTION (UM^ 1 4 DRAWING SIZE (FT.) 2A X 24 5 X 5 # MYLARS 49 4 PLOTS PER LAYERS 15 1
- 5
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 DESIGN VS. TECHNOLOGY DEVELOPMENT CYCLE TIMES 8 r
6 -
4 - YEARS
2 -
1970 1975 1980 1985 1990
A. LEAD TIME FROM RELEASE OF DESIGN RULES TO PRODUCTION B. DESIGN TIME FOR A LOGIC PRODUCT. C. DESIGN TIME. FOR A MEMORY PRODUCT. As one can see in figure 1, the time to develop a new technology is in the 1-2 year time frame. However, the time to design a new logic circuit is in the 4 year time frame.
- 6 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 A
--a
--P
n C A ?s s - a |si s •I p- o cn u n 5 So 3a <
- 7 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o a J2 c (D CO A CHIP Cofplaxitu ? in thouoanda \ of campananta \ clag scalB> > m c CT CO Q! B)' •5 > o
o o 3 •c fi> PRODUCT t^ftftQSXiJCTlOH 1 l\5 00 CO O a. OESIGN STARTS a. CD IBB T3 0)
CO u 7a c_ o CO CD O > CO CJ1
o 00
> 80 i iB7a 75 ea es CD^ CD
CO -.J CJ BENEFITS OF DESIGN AUTOMATION
0 SOONER TO MARKET
0 LOWER DESIGN COST
0 FEWER ERRORS
0 PRODUCTS BUILT ON CURRENT TECHNOLOGY
0 HIGHER ENGINEERING LEVERAGE
- HIGHER RETURN ON ENG'G INVESTMENT - MORE PRODUCTS OUT OF EACH LOGIC/CKT ENGINEER - FASTER CORPORATE GROWTH
9 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 III. EXPECTATIONS
- ESSENTIALS
- HIGHLY DESIREABLE
- OTHER FEATURES
- 10 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ENGINEERING WORKSTATION HISTORY
5/82 ESTABLISHED SELECTION COMMITTEE
8/82 DEMONSTRATIONS
9/82 FIRST CUT SPECIFICATION
10/82 - 2/83 TRIAL EVALUATIONS
5/83 SECOND DETAILED SPECIFICATION BASED ON "CURRENT" CAPABILITIES
2/83 DECISION MATRIX FAILED TO QUALIFY ANY ONE
8/83 REDUCED SPECIFICATION
9/83 DECISION MATRIX
10/83 INITIAL SYSTEM
- 11
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ENGINEERING WORK STATION SELECTION CRITERIA I. E??FNTiAL FEATURES 1. POWERFUL, FAST GRAPHICS EDITOR 2. PLOT CAPABILITY 3. INTERFACE TO NETLIST DRIVEN SOFTWARE A. MASK LAYOUT CAPABILITY 5. DATA BASE CAPABLE OF STORING ALL INFORMATION KNOWN ABOUT A DESIGN EXCEPT FOR LAYOUT DETAILS 5. CAPACITY SUFFICIENT TO HANDLE 100 D-SIZE DRAWINGS 7. HIGH SPEED IBM INTERFACE 8. ABILITY TO SUPPORT A HIERARCHICAL DESIGN 9. 9-TRACK TAPE AVAILABILITY 10. LOCAL APPLICATIONS AND MAINTENANCE SUPPORT
II. HIGHLY DESIREABLE FEATURES 1. PORTABLE SOFTWARE 2. RELEASE CONTROL AND SECURITY SYSTEM 3. ABILITY TO SUPPLY MANY SYSTEMS PER YEAR 4. LOCAL GENERAL PURPOSE PROCESSING 5. INTERACTIVE LOGIC SIMULATION WITH BEHAVIORAL MODELS 5. ABILITY TO RUN UNIVERSITY SOFTWARE III. OTHER DESIREABLE FEATURES 1. LOCAL NETWORKING 2. WORD PROCESSING 3. ELECTRONIC MAIL 4. TIMING VERIFICATION
- 12 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ENGINEERING WORKSTATION HISTORY
5/82 ESTABLISHED SELECTION COMMITTEE
8/82 DEMONSTRATIONS
9/82 FIRST CUT SPECIFICATION
10/82 - 2/83 TRIAL EVALUATIONS
5/83 SECOND DETAILED SPECIFICATION BASED ON "CURRENT" CAPABILITIES
2/83 DECISION MATRIX FAILED TO QUALIFY ANY ONE
8/83 REDUCED SPECIFICATION
9/83 DECISION MATRIX
10/83 INITIAL SYSTEM
- 13 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 DECISION MATRIX
SCHEMATIC EDITOR
NETLIST GENERATOR
LOGIC SIMULATOR
TIMING VERIFIER
PLA OPTIMIZER
LAYOUT EDITOR
AUTO ROUTER
NETWORK EXTRACTOR
DESIGN RULES CHECKER
CONTINUITY CHECKER
GENERAL PURPOSE COMPUTING POWER
- 14
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ENGINEERING WORKSTATION HISTORY
5/82 ESTABLISHED SELECTION COMMITTEE
8/82 DEMONSTRATIONS
9/82 FIRST CUT SPECIFICATION
10/82 - 2/83 TRIAL EVALUATIONS
5/83 SECOND DETAILED SPECIFICATION BASED ON "CURRENT" CAPABILITIES
2/83 DECISION MATRIX FAILED TO QUALIFY ANY ONE
8/83 REDUCED SPECIFICATION
9/83 DECISION MATRIX
10/83 INITIAL SYSTEM
- 15 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ENGINEERING WORK STATION SELECTION CRITERIA I. ESSENTIAL FEATURES 1. POWERFUL, FAST GRAPHICS EDITOR 2. PLOT CAPABILITY 3. INTERFACE TO NETLIST DRIVEN SOFTWARE 4. MASK LAYOUT CAPABILITY
- 16 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ENGINEERING WORK STATION SELECTION CRITERIA I. ESSENTIAL FEATURES 1. POWERFUL, FAST GRAPHICS EDITOR 2. PLOT CAPABILITY 3. INTERFACE TO NETLIST DRIVEN SOFTWARE 4. MASK LAYOUT CAPABILITY 5. DATA BASE CAPABLE OF STORING ALL INFORMATION KNOWN ABOUT A DESIGN EXCEPT FOR LAYOUT DETAILS 5. CAPACITY SUFFICIENT TO HANDLE 100 D-SIZE DRAWINGS 7. HIGH SPEED IBM INTERFACE 8. ABILITY TO SUPPORT A HIERARCHICAL DESIGN 9. 9-TRACK TAPE AVAILABILITY 10. LOCAL APPLICATIONS AND MAINTENANCE SUPPORT
II. HIGHLY PES IREABLE FEATURES 1. PORTABLE SOFTWARE 2. RELEASE CONTROL AND SECURITY SYSTEM 3. ABILITY TO SUPPLY MANY SYSTEMS PER YEAR 4. LOCAL GENERAL PURPOSE PROCESSING 5. INTERACTIVE LOGIC SIMULATION WITH BEHAVIORAL MODELS 5. ABILITY TO RUN UNIVERSITY SOFTWARE III. OTHER DESIREARLE FEATURES 1. LOCAL NETWORKING 2. WORD PROCESSING 3. ELECTRONIC MAIL 4. TIMING VERIFICATION
- 18 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ENGINEERING WORK STATION SELECTION CRITERIA I. ESSFNTIAL FEATURES v/ 1, POWERFUL, FAST GRAPHICS EDITOR v/ 2. PLOT CAPABILITY ?• 3. INTERFACE TO NETLIST DRIVEN SOFTWARE - ^. MASK LAYOUT CAPABILITY ~ 5. DATA BASE CAPABLE OF STORING ALL INFORMATION KNOWN ABOUT A DESIGN EXCEPT FOR LAYOUT DETAILS \/5. CAPACITY SUFFICIENT TO HANDLE 100 D-SIZE DRAWINGS ? 7. HIGH SPEED IBM INTERFACE \/8. ABILITY TO SUPPORT A HIERARCHICAL DESIGN v/9. 9-TRACK TAPE AVAILABILITY v/10. LOCAL APPLICATIONS AND MAINTENANCE SUPPORT
II. HIGHLY PES IREABLE FEATURES SOME 1. PORTABLE SOFTWARE SOME 2. RELEASE CONTROL AND SECURITY SYSTEM v/ 3. ABILITY TO SUPPLY MANY SYSTEMS PER YEAR \/ A. LOCAL GENERAL PURPOSE PROCESSING - 5. INTERACTIVE LOGIC SIMULATION WITH BEHAVIORAL MODELS - 6. ABILITY TO RUN UNIVERSITY SOFTWARE III. OTHER DESIREABLE FEATURES v/ 1. LOCAL NETWORKING - 2. WORD PROCESSING \/ 3. ELECTRONIC MAIL SOME 'I. TIMING VERIFICATION
- 19 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ENGINEERING WORKSTATTON FUTURES
0 HIGH SPEED COMMUNICATIONS 0 HARDWARE SIMULATION 0 CIRCUIT EMULATION 0 PLACEMENT/ROUTING 0 DESIGN DATA BASE MANAGEMENT
- SECURITY - PROJECT MANAGEMENT - RELEASE CONTROL
- 20 -
Dataquest Incorporated, A Subsidiary of A.C. Nielse'n Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ENGINEERING WORKSTATION SUCCESS
- HEAVY VENDOR/CUSTOMER INTERFACE
- THIS IS AN ITERATIVE PROCESS
- IT IS VERY DIFFICULT
0 TO CLEARLY SPECIFY NEEDS
0 TO BUILD WHAT YOU WANT VERSUS WHAT YOU SPECIFY
- RESPONSIVENESS
- 21 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Dataquest
CAE FOR SEHICUSTOH INTEGRATED CIRCUITS
Steve Shively Computer-Aided Engineering Manager International Microcircuits Incorporated
Mr. Shively is Manager of Computer-Aided Engineering at International Microcircuits Inc., and is currently chairman of the Mentor User's Group. Mr. Shively has worked at IMI for 15 months and has integrated the Mentor system into the IC design process at IMI to include schematic capture/ simulation, test vector generation, and schematic to layout verification. Mr. Shively was previously with Intel Corporation, where he designed bipolar IC processes and was later responsible for developing the graphics methodology for using Ultratech IX reticle stepper masking technology.
Dataquest Incorporated ELECTRONIC DESIGN AUTOMATION FOCUS CONFERENCE May 10-11, 1984 Sunnyvale, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAE for SEMICUSTOM INTEGRATED CIRCUITS
NTERNATIQNAL MICRQCIRCUITS INCORPORATED
International Microcircuits incorporated (I Ml) has been manufacturing semi-custom integrated circuits ( gate arrays) since 1974. Since then I Ml has grown to include our own facilities for manufacturing of CMOS integrated circuits. I MI provides extensive design engineering support to customers who are unfamiliar with IC's, logic verification of customer's designs, and prototype parts for customer evaluation before production begins. IMI has complete in-house design automation tools, production wafer fabrication capabilities, assembly facilities, and a production test floor. IMI did around S12 million in sales in 1983, and growth for 1984 looks very promising. IMI maintains over 100 employees, in excess of 25.000 square feet housing all phases of manufacturing. IMI has had over 300 customers to date comprising well over 750 unique, customer defined designs. Nearly all of IMI's current business stems from gate array products; however, standard cell designs will soon be available, and several standard parts have been designed C such as multipliers. 290rs. etc ). Late in 1982. IMI decided that CAE workstations had the potential to be a major part of the logic designer's future; furthermore, that IMI should be prepared to interface cleanly with a large segment of the logic design community.
- 1 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAE for SEMICUSTOM INTEGRATED CIRCUITS
SFMI CUSTOM IC TRENDS
Growth of semi-custom IC markets is projected to continue through 1991 at an annual rate of 32/1 (according to SI A & EDN). The size and number of semi-custom IC vendors has grown tremendously in recent years. Manufacturers will need to continually improve the speed and effiency of their design process to provide a relatively quick turn-around time to remain competitive. Increasing design complexity will make the job of reducing design times more and more difficult. The number of gates per IC has been increasing consistently; this trend will continue. Larger designs require more time for nearly every step in the design process; logic specification, simulation, fault simulation, layout, and test generation are all included. We are constantly looking for a faster, more efficient way to perform each of these tasks. The increasing performance of CMOS semi-custom IC's has caused several additional burdens on computer aided design: delays introduced by metal lines must be carefuliy monitored to insure correct circuit performance, transistor delays must be carefully characterized and modelled, and fabrication techniques must evolve to handle the shrinking geometries. These trends clearly present the need to reduce the length and number of design cycle iterations. Most logic designs today are changed many times in their design lifetime; but the catch is that changes appear throughout the design cycle which will waste resources. More powerful design analysis tools must be used to prevent erroneous designs iterations.
- 2 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAE for SEMICUSTOM INTEGRATED CIRCUITS
my USE CAE WORKSTATIONS?
Why are semi custom IC manufacturers so eager to use CAE workstations ? There are dozens of resons why they're important to us, but generally the reasons fall into one of three categories: 1) Design Aids for the Customer. 2) Enhancement of the Customer interface, and 3) Enhancement to internal design methodologies. As a design aid to the customer. CAE workstations are at home in their intended environment. These tools have evovled to replace breadboards for design and paper for documentation. These systems are relatively low cost compared to the alternatives of purchasing batch-oriented computer equipment to perform simulations and other verification needs. Use of CAE systems will enhance a vendor's interface with customers. If both vendor and customer design, verify . and store their design data electronically, then the vendor can Insure the customer that the manufactured part will meet the specifications of the customer - even on large, complex designs. Also, the customer can have easy access to large Macro Cell libraries containing primitive and high order logic functions to use in his design. Information can be exchanged through a common media. Semi custom vendors may enhance their design flows by using CAE systems to drive automatic placement and routing programs. A key aspect of IC layout is relating layout geometries to the logic schematic. CAE workstations also provide a ability to electronically capture and document their designs; the large number of designs projected for the next 5 years will demand some effective documentation techniques. But overall, use of CAE systems by semicuston IC makers is the ability to provide relatively simple access to IC silicon technology to its customers.
- 3 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAE for SEMICUSTOM INTEGRATED CIRCUITS
CAE TOOL NEEDS
In a semi custom environment, throughput is an extremely important issue; therefore, the single most important need from CAE systems is more power at a lower cost. Although the CAE system manufacturers currently boast awesome power at your fingertips, the fact is that design verification needs will continue to be greater than the CPU cycles available to the designer in his own real-time environment. I Ml experience has shown that simulation alone will take more than the power available in any workstation; and let's not forget the many other functions such as timing verification, fault simulation, test vector generation, etc that must also be performed. Faster machines will help reduce the time to complete a typical design; and/or allow much larger designs to be performed. But most critically, more powerful systems will allow the user to perform more verfication steps which is key to making a design work correctly the first time in silicon. Another area of need is in phsyical layout tools for gate arrays. Efficient phsyical layout automation is a prime concern for gate array manufacturers. Although some people are impressed by today's offerings in this area, I Ml has found severe drawbacks to current auto-routing techniques. Silicon costs money; it is not the largest cost of all IC's. but it is always significant. Furthermore, remember that semi custom IC designs are going to continue to grow in size. Current auto-routing uses much more si I icon than hand-mapped gate arrays. As noted earlier, changes to a design happen frequently. When this occurs, it would be optimum to make the change in one place and have it propagate throughout the entire design cycle in a timely fashion. Reality dictates that one change the necessary steps and reduce the amount of the process it takes to get parts; in other words, the designer will want to make the change on a layout. Auto-routing results in a jumble of wires which defeats fast fixes to most circuits. A small but rapidly growing need is a graphical interface between CAE systems. Perhaps this should be accomplished by establishing a standard intermediate format containing graphical information within a net list; CAE vendors would then have to provide a means to input this information to their
- 4 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAE for SEMICUSTOM INTEGRATED CIRCUITS CAE TOOL NEEDS
respect I vedatabases. An option to the standards is for a set of database trnaslation programs to allow conversion of. for example, a Daisy schematic onto a Valid workstation. This, however, seems to be the least likely since each CAE vendor would naturally like to protect the proprietary nature of the database. Database translation is needed for easier interface to a growing set of customers using a variety of CAE systems. The advantage of translating schematic databases to other systems is clear for semi custom IC vendors: it would allow the vendor to customize his CAE interfaces around one system, instead of many types of systems. A more global need is for a set of integrated conceptual design tools. This design environment would aid the designer in defining his logic problem, allowing the problem statement to be some kind of algorithmic description. Furthermore, this initial description should contain any specifics which become design implementation constraints. But most importantly, the initial specification would be used to verify the design as it progresses. Th i s type of too! wou Id he I p prov i de assurance that the design is error free from the initial logic specification to test vector generation and prototype testing of parts by testing the final product based on the initial problem specification. This would, of course, demand that testabiIity of logic integrated circuits be considered as part of the original design phase.
- 5 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAE for SEMICUSTOM INTEGRATED CIRCUITS
EXPECTATIONS OF CAE MARKET
The greatest expectation I have about the CAE industry is that more economical mrarkstations will become available. The use of IC technology in workstations guarantees continued price reduction, and demand from the market will continue to drive down prices as welI. We In the electronics industry have seen major improvements in the performance of computer systems since the introduction of semiconductor transistors. It is clear that this trend will continue for many years to come. My expectation of CAE vendors is to continue to provide more powerful hardware aa they themselves grow. Finally. I expect CAE vendors to provide more verification too I s and some soph i st i cated des i gn a i ds. The mar ket will demand these as the use of CAE workstations spread and users become familiar with what these systems can and can not do for them. As a result of these expectations, most logic designers will perform their complete design function on the CAE workstation. Currently, only a part of the design work can be performed on a workstation; while conceptual design, performance specification, and test generation are missing, and other features (such as auto-routing, prototype testing, etc) are very weak. These expectations are necessary to fuTfill the needs of logic design on a CAE system.
- 6 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CAE for SEMICUSTOM INTEGRATED CIRCUITS
FUTURE TRENDS
There are two future trends In the CAE industry which are evident at this tiine: 1) the move to specialized hardware for typically CPU intensive tasks, and 2) more powerful & less costly systems. Specialized hardware for logic simulation is already here, evidenced by products like the Megalogician & Zycad Logic Evaluator. Prototype test equipment has been recently introduced, also. We will probably see fault simulation, analog simulation, and test vector generation in hardware in the near future. The advantages for this approach is that hardware is much faster than software and can be integrated into a workstation environment. The trend towards more powerful and less costly systems should continue. In fact, an acceleration of this trend is expected because of the effects of technology feeding technology. That is, most workstation hardware is now being designed on workstations, thereby increasing the effiency of the design of the next generation products. In my knowledge, this is the case for Daisy. Valid, and Mentor-Apollo.
- 7 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 knl P»i IHI
-J W Q < I- lU 2 5 •- O g < z ^ =^ i ^ £ fi P; O I- 2 o Z S
- 1 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 a w a c (D CO O O -1 •D O
> w a. INTERNATIONAL MICROCIRCUITS ni'
> INCORPORATED b o o • INCORPORATED 1972 3 • MANUFACTURING GATE ARRAYS -^ I to § 1^ 5 I SINCE 1974 d a (D •, $12M SALES 1983 • 100 PEOPLE w [U 3 • 25,OOO SQUARE FEET o > • 750+ INTEGRATIONS TO DATE • 3OO+ CUSTOMERS oto o o I.T
CD
CO -^1 w BHI s? iMig (0 oi o5" U.3Q mOUi
111 ^^ W<3 H
- 3 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 D 01
(t> a w c IT CO a.' 0)'
> b 2. (fl (D o o 3 SEMICUSTOM 10 TRENDS •o g>
lo ' (O 0 ji. • INCREASING DEMAND FOR SEMICUSTOM CIRCUITS 3] 01 ' a CD •D CU • INCREASIGING DESIGN COMPLEXITY I
CO • INCREASING PERFORMANCE OF IC'S 0) o u> (D o > CO
j:. o oa oCD o o IMI 5E. (D X
-NI CO •>J OJ Iniill ss ?>Q u o o 0.1 w
oc GO <^ CO ^ (3 - O I- OC UJ o 00 w oc OC (0 < OO m <
m o OO
OO
- 5
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o aS. n c
o o •3 o o. INTERNATIONAL MICROCIRCUITS w c cr CO 9! ARRAY DENSITY TREND tu'
> GATE CELLS PER CHIP o
o o 3 •o (U 10,000 3
to to o 0> a. a. 1,000 (D -, TJ 0) 100
w Q)
O (n _(D 10 >O
1 1 .c^ o CD *74 '76 "78 '80 •82 •84
CO YEAR O O IMI O
<5. CD to ->l CO fflnl -• S SIS' UuMi
M CO
o CO (0 LU >- Q < QC CC < CC LU CC CO > < m CO I- o QC LU s o (0
h-
o o (0 c
7 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 /i408) 971-9000 / Telex 171973
•tV- 1-. D 0)
E
> W c cr — a. 0)' -1 •< o^ > b ro (0 0) o o 3 SEMICUSTOM 10 TRENDS •o 01 /' lo ' (O • TRENDS IMPLY NEED TO REDUCE THE o 00 3) a. I LENGTH AND NUMBER OF DESIGN CYCLE Q. (D TJ OJ ITERATIONS IN ORDER TO MEET
CO GROWING DEMAND
O > (O Ul o•t. 00 oCO o I J. o
(O -«l u o 0)
o3 o -1 •oo
CD a > CO
a.
> b
o o WHY USE CAE WORKSTATIONS? 3 •a 01
M I CO • DESIGN AIDS FOR CUSTOMERS O lO X ci. i a n> - LOGIC SPECIFICATION & DOCUMENTATION -, TJ a> - INTEGRATED LOGIC VERIFICATION TOOLS
W (u - RELATIVE LOW COST VS ALTERNATIVES O u> (S o >
.c o OS
CO u->l o u
(D Q.
W C cr 9: D)' •< > o
tD^ (D o o WHY USE CAE WORKSTATIONS? 3 •uo ENHANCEMENT OF CUSTOMER INTERFACE
a Q. (D —. - CORRECT TRANSFER OF LARGE DESIGNS T) B> - ACCESS TO LARGE MACRO CELL LIBRARIES
O) 01 - FEEDBACK INTO CUSTOMER'S DESIGN FLOW 3
o >
o 00 oto -T o
(D^ CD X
-4 CO ->I G} o 01
•a o
> w c a w a a> •< o > b a
.»> o CO
"li^? (D o O O m (D
CO u o 0>
E
> w c cr (0 a! 0)' -t •< o > o o CAE TOOL NEEDS 3
i>j i_,
TJ a> - LARGER DESIGN CAPACITY
W - MORE DESIGN VERIFICATION o >
3o (O •vl o(O o o iE. (D
(O D D>
O o •o5
(D Q. > w
g. BJ' •2 o > b
o 3 CAE TOOL NEEDS •o a> 3
l\> ,_,
o CO o(O o fMf o ^^Ti.
(O ~J CO o 0)
o o •oo
Q.
w c cr (A Q.' 01 •< o > b CAE TOOL NEEDS o o 3 •a D> • GRAPHICAL INTERFACE BETWEEN CAE SYSTEMS i\) CO o 3) d - CAE INDUSTRY STANDARDS OR Q. CD 01 - CAE SYSTEM DATABASE TRANSLATIONS - SEMICUSTOM VENDOR INTERFACE SUPPORT W 0) REQUIREMENTS
O >
.0. o CD s I !E. CD X
--J
CO -NI u D
BT •Oc (0 (/> o o •o o o _a > « c cr en a oi'
> b CAE TOOL NEEDS (S o o 3 •o DJ 3 • CONCEPTUAL DESIGN TOOLS
l>0 § 33 - LOGIC SPECIFICATION a Q. n> - TEST VECTOR GENERATION 5" • TESTABILITY IS A DESIGN ISSUE CO 0> - PROTOTYPE TESTING o > (0 Ol
.». o 00
o(O o o LMl
2. (D
u D 0)
"O o
CO c CT en Q! o>' •< > b
CD CD EXPECTATIONS OF CAE MARKET (D o o 3 •o 0) 3 • MORE ECONOMICAL WORKSTATIONS ^ I oto etn; • MORE POWERFUL WORKSTATIONS Q. Q. (D T3 (U • MORE DESIGN AIDS & VERIFICATION TOOLS
W • RESULT: MOST LOGIC DESIGNS B> 3 PERFORMED ON WORKSTATIONS o >
to o o o IMI
?
n> Q.
W c cr
> b FUTURE TRENDS
I • SPECIALIZED HARDWARE 3 ^ I JS M - LOGIC SIMULATION IS HERE O .^J
CL I •. - PROTOTYPE TESTING n •D RECENTLY INTRODUCED OJ - FAULT SIMULATION W 0> 3 - ANALOG SIMULATION o > - TEST VECTOR GENERATION
•fc. o 00
o(O o o I
(O u o lu
^
> 1/3
a 0)'
> b FUTURE TRENDS
o o 3 MORE POWERFUL & LESS COSTLY SYSTEMS •o D>
N) I-" - IC PRICE-PERFORMANCE TRENDS 8 00 5 I Q. WILL CONTINUE Q. (0 "O lu - TECHNOLOGY FEEDS TECHNOLOGY - MENTOR - APOLLO 0> - DAISY
O> - VALID
o*. 00
CO o o o £J»£
USER PANEL DISCUSSION
Karl Nakamura Senior Member of the Technical Staff GRiD Systems
Mr. Nakamura is a Senior Member of the Technical Staff at GRiD Systems. He has been with GRiD for more than three years, and has participated on the GRiD Compass design team. Previously, Mr. Nakamura was with Hewlett-Packard as a Research and Development Engineer in the Scientific Instruments Division. Mr. Nakamura received an M.S.E.E. degree from Stanford University.
Dataquest Incorporated ELECTRONIC DESIGN AUTOMATION FOCUS CONFERENCE May 10-11, 1984 Sunnyvale, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 THIS PRESENTATION WAS NOT AVAILABEL AT PUBLICATION TIME.
IF A COPY IS MADE AVAILABLE TO DATAQUEST, WE WILL MAIL IT DIRECTLY OT YOU FOLLOWING THE CONFERENCE.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Dataquest
INTEGRATED SYSTEMS TECHNOLOGY: VLSI DESIGN AUTOMATION FOR SYSTE34S ENGINEERS
Phillip A. Kaufman President Silicon Compilers Incorporated
Mr. Kaufman is President of Silicon Compilers Incorporated and has 18 years of experience developing and managing software, systems, and integrated circuit programs and organizations, and strategic and product planning. He was previously with Intel Corporation, where he served as Director of Engineering, in Corporate Strategic Planning, and as General Manager of the Microprocessor Operation. Mr. Kaufman was the driving force behind both the development and acceptance of the IEEE Floating Point Standard and the organization of the Xerox/Digital Equipment/Intel joint program that resulted in the Ethernet standard. Prior to joining Intel, Mr. Kaufman was Director of Engineering and Director of Planning at Computer Automation Incorporated. Mr. Kaufman received B.S.E.E. and M.S.E.E. degrees from the University of Michigan.
Dataquest Incorporated ELECTRONIC DESIGN AUTOMATION FOCUS CONFERENCE May 10-11, 1984 Sunnyvale, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 INTEGRATED SYSTEMS TECHNOLOGY:
VLSI DESIGN AUTOMATION FOR SYSTEMS ENGINEERS
Phillip A. Kaufman Silicon Compilers Incorporated
Dataquest Incorporated, A Subsidiary of A.C. Nielsen CoInpany /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 INTEGRATED SYSTEMS TECHNOLOGY: VLSI DESIGN AUTOMATION FOR SYSTEMS ENGINEERS
INTRODDCTION Since the advent of the transistor, semiconductor technology has had a profound effect upon the design and construction of all types Of systems. The impact of semiconductor technology has grown rapidly as the technology evolved from transistors to integrated circuits and as the available integrated circuit density has increased. With the evolution from SSI to MSI to LSI to VLSI, the complexity of all types of systems has increased dramatically while costs have declined and reliability has improved. Today, it is a well accepted fact that competitive systems products must make maximum use of Very Large Scale Integration.
Despite the revolution in available silicon density, there has been surprisingly little change in the past twenty years in the way in which systems and integrated circuits are designed. Systems designers still rely primarily upon standard products developed by integrated circuit suppliers. The integrated circuits themselves are developed by a relatively small number of IC design craftsmen. Computer Aided Design (CAD) and Computer Aided Engineering (CAE) have helped to address the problems of designing ever more complex components. However, CAD and CAE have focused primarily upon individually computerizing many of the classical steps required to implement VLSI components. As a result, despite all of the efforts • that have gone into CAD development for IC design, the best that has been accomplished has been to keep the already ve-ry long development times from growing totally out of sight as complexity has increased.
The advent of the microprocessor caused a dramatic revolution in the way in which systems were designed. No longer was it necessary for the system implementor to design the entire product from functionally small individual components. Now the system implementor could design much of the functionality through software and use powerful low cost microprocessors to implement the heart of the system. The first users of microprocessors found that they had a substantial advantage over their competitors. But the microprocessor revolution is over. With microprocessors appearing everywhere in virtually every system, their use has become evolutionary rather than revolutionary. Microprocessors also achieved one goal that is often overlooked. They became the way that . integrated circuit suppliers could define very complex components which could exploit the available silicon density and still address a very broad market. This characteristic is extremely important since, other than memory devices, microprocessors are in general the only very complex Integrated Circuits with a large enough market to justify the enormous development costs. Standard product manufacturers find - 1 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 it very difficult to define a large number of products that have sufficient functionality - to fully exploit the available silicon density while still having a broad enough market to justify the investment. This difficulty arises from the fact that today's VLSI devices are not so much individual components as they are complete systems and subsystems. Unless many, many systems designers build virtually the same system, it is very difficult to sell large volumes of a- complex integrated circuit. With VLSI components becoming major parts of end systems, it is no longer possible for the integrated circuit design engineer to also have sufficient applications knowledge to define the components needed. Thus, we have come to somewhat of an impasse in the industry. The systems designers who have the applications knowledge and have the need to integrate their systems into one or more VLSI circuits do not have the IC design knowledge or the tools to execute the design. At the same time the IC designers do not have the systems applications knowledge or the volume markets necessary to develop components that are actually major parts of complex systems.
THE CUSTOM REVOLUTION Custom and semi-custom technologies have lately received a great deal Of attention. The attention, from the systems developer, arises from a recognition of the many benefits of using high density silicon in systems products. The attention, from the integrated circuit suppliers, comes from a recognition of the difficulty of defining many high density standard products and the large potential market represented by custom and semi-custom designs. Dataquest has estimated that by 1990 nearly half of all integrated circuits shipped will be some form of Application Specific Integrated Circuit (ASIC) .(Figure 1) . The ASIC market segment includes all forms of custom, semi-custom, and mask programmable devices. The custom and semi-custom segments alone will account for in excess of $10 Billion in 1990 shipments. These forecasts, however, beg the question of just who will design all of these custom and semi-custom ICs and what type of design tools the designers will use. The forecasts of very rapid growth in the ASIC market are certainly justified in terms of the number of systems built per year and the potential that exists for penetrating this market. The forecasts can not be met, however, unless systems engineers are able to do the designs and unless tools exist to execute the designs rapidly and cost effectively.
Existing approaches to user designed VLSI have focused upon ways to eliminate the need to design polygons and transistors by using predesigned cells such as are used in Gate Array and Standard Cell systems. These approaches make significant sacrifices in silicon density in order to use predesigned (and often preplaced) cells. Little has been done to make the task of designing complex systems more efficient.
- 2 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 What has actually happened with the use of Gate Arrays and Standard Cell systems is to set the systems design engineer's design methodology back ten years. Over the last ten years, the systems engineer has been given ever more complex building blocks from which he could construct very complex systems. He could use Medium Scale Integrated circuits (MSI), Large Scale Integrated circuits (LSI), and Very Large Scale Integrated circuits (VLSI) and not have to decompose his design into literally hundreds of thousands of gates or transistors. The effect of most Gate Array and Standard Cell systems is to take away from the engineer his MSI, LSI, and VLSI catalogs and give him back a catalog of Small Scale Integrated circuits (SSI) while asking him to design systems substantially more complex than those designed ten years ago.
For the custom revolution to be realized, a new design technology is required. A technology that allows systems engineers to rapidly and efficiently design integrated systems. One that addresses the entire-design engineering process focusing upon design methodology and a complete integrated set of tools that not only implfements efficient silicon realizations of the design but that substantially shortens the design time. In today's rapidly evolving markets, time to market is often the single most critical edge that a manufacturer can have. With the ever increasing complexity of VLSI, time to market becomes a design- time issue far more than a manufacturing-turn-around time issue. The optimum solution is a systems-integration technology that combines the shortest possible design time with the most cost- effective silicon implementation.
INTEGRATED SYSTEMS TECHNOLOGY Integrated Systems Technology encompasses the products, services, and expertise that allows the systems engineer to take advantage of the benefits of VLSI and rapidly and efficiently design end systems products. With Integrated Systems Technology, the systems engineer can fully exploit the reduced size, increased speed, lower cost, lower power, and product uniqueness potential of VLSI. No longer is it necessary to have a large team of IC design craftsmen in order to implement efficient integrated circuits. The systems engineer can develop systems on silicon without the need to be retrained as an integrated circuit designer and without paying the cost and chip-size penalties associated with compromise technologies. There are two key aspects of any design technology: design efficiency and implementation efficiency. Design efficiency refers to the time and manpower required to perform the design. Implementation efficiency refers to the cost effectiveness of the resulting design. Unless both types of efficiency are available, the applicability of a design technology is severely limited. The basis for a real Integrated Systems Technology is silicon compilation. Silicon compilation provides both an extremely
- 3 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 productive design methodology and integrated circuit implementations that are essentially as dense as hand-crafted designs. Design efficiency is achieved by addressing the entire design cycle/ from concept to tape-out, and providing a complete set of tools that maximizes the design engineer's leverage. Silicon implementation efficiency is achieved by actually constructing the exact required functionality using a built-in set of knowledge based rules.
SILICON COMPILATION Silicon compilation is far more than Computer-Aided-Design (CAD) or Computer-Aided-Engineering (CAE). It is true design automation. This distinction is both qualitative and quantitative and results in a very different focus in both developing and using the silicon compilation system. Classical CAD and CAE approach the automation of design by looking at the way in which design and implementation have been done in the past and looking for opportunities to shorten each step. Silicon compilation, on the Other hand, looks to the design cycle and asks the question "Where can the engineer's knowledge and expertise be best exploited and where can the capabilities of the computer not only ease the task but even eliminate major parts of the design effort?" The difference between CAD/CAE and design automation is illustrated in Figure 2. The top line in the Figure portrays the classical IC design cycle from definition to tape-out. Each step of the process involves both design and verification or simulation of the design work. In addition, of course, each step must be verified to be sure that the work done matches the work done in the prior step. CAD and CAE address primarily the implementation steps after the higher level design work has. already been completed. We even speak of "design entry" meaning capturing of a design that is essentially accomplished without the assistance of the system. The only real assistance that the CAD/CAE system provides is to replace paper with a CRT screen, maintain a database of the design, and perform basic error checking. Thus, the definition phase is unaffected while each of the subsequent steps is compressed slightly. In sharp contrast to this approach, the design automation of silicon compilation not only shortens the definition and architectural phase, it improves the quality of the work and totally eliminates the later stages. The result is complex designs executed in a few months by one or two engineers, compared to similar designs done by tens of engineers over a multi-year project.
Silicon compilation has many features and attributes of interest to the systems design engineer. These are summarized in Figure 3. Silicon compilation is a complete design automation capability. It covers the entire design cycle and eliminates the need for the design engineer to perform the routine and labor-intensive implementation tasks. The design begins at the functional or micro-architectural level where the engineer can think in terms
- 4 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 of major functional blocks and the way in which they are connected. From the block level design the compiler synthesizes three different views of the design: layout geometry, functional simulation modelSf and timing models. Thus, the design is immediately ready to simulate and verify. In addition, the design is foundry-independent. The silicon compiler yields an optimum combination of both design efficiency and silicon efficiency.
DESIGN EFFICIENCY Design efficiency in the silicon compilation methodology is achieved by matching the tools to the way in which an engineer really works. Engineering is a process of balancing often conflicting design goals and successive refinement of a design. An initial specification for a design is never absolutely complete and seldom adequately reflects all of the metrics to be used in making design tradeoffs. Real designs are accomplished by selecting an implementation approach, partially implementing the approach in order to evaluate the results, and then altering the approach as needed to best match the project's goals and constraints. Unfortunately, until now, the task of doing a partial implementation in order to get real results to evaluate was SO expensive that it has been impossible to evaluate several alternative approaches and select the optimum one. Because of the ability of a silicon compiler to take an initial design specification and immediately turn it into an implementation, it is now possible to explore multiple implementation alternatives. This exploratory-design phase yields many benefits. A compiler designed to accept minimal input information and interactively return design parameters such as speed, power, and die size allows the design engineer to use his creativity to explore many alternatives in ways that were previously impossible. The result is often a far superior design to any that could be achieved before.
An additional benefit of the concept of exploratory design is that it leads directly to a design methodology well suited to iterative design. Real engineering is iterative. It is necessary to incrementally refine definition and implementation while simultaneously evaluating and verifying the implementation. A silicon compiler developed with the philosophy of being the design engineer's partner accepts a little input and provides maximum feedback on the results of that input. It also treats the layout geometry, functional simulation, and timing analysis of the design in progress as three views of the same design. This approach enables and encourages exploratory design and iterative design resulting in rapid convergence on an optimum implementation. HIERARCHICAL DESIGN Hierarchical design, or structured design, is a design approach developed to make it possible to manage very complex design tasks, be they integrated circuits, software, or any other type
- 5 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 of design. The key attribute of hierarchical design is the breaking down 'of a design into a small number of relatively large functions and then recursively breaking down each of these functions to lowerest level until the design is completed. Figure 4 illustrates a pyramid of design levels, beginning at the top with large VLSI-level functions and descending through LSI, MSI, all the way to polygons. While CAD/CAE systems stress the need to perform hierarchical design in order to manage the immense complexity of VLSI design and provide tools to assist the designer in this process, the pyramid must still be traversed from top to bottom. This process of successive decomposition to the levels of gates, transistors, or polygons, is very time consuming and error prone. A silicon compiler, on the other hand, provides the designer with maximum leverage by allowing the design to focus at the highest possible levels and letting the design-automation system perform the necessary decomposition. This leverage is illustrated in Figure 5. By allowing the designer to completely execute a design at the level of large modules and blocks, without descending the hierarchy into a sea of details, the silicon compiler provides dramatic leverage for the designer. Without this leverage, no design system is adequate to the enormously complex task of designing a complex VLSI system.
IMPLEMENTATION EFFICIENCY Implementation, or silicon, efficiency is determined primarily by the quality of circuit and interconnect wiring execution. A silicon compiler has a substantial advantage in silicon efficiency over other approaches to custom or semi-custom implementation. In either gate array or standard cell design systems, small predesigned cells are interconnected to meet the needs of the design. In these approaches, cells are quite small and are designed without regard for how the cells will be combined to construct larger functions. The result is that a very high percentage of the chip area ends up being used for interconnect wiring. Wiring, however, is the most expensive chip resource. The heavy use of inter-cell wiring in gate array and standard cell systems results in very inefficient implementations. Another factor affecting silicon efficiency is the appropriateness of the available cells for the particular design being executed. A library-based system will always be limited by the available library. Often a 'next larger' cell is used because exactly the function needed is not available. A silicon compiler, however, is not based upon a library of objects (cells). Rather, it is based upon a knowledge base of rules for the construction of large families of functions. Thus, a compiler constructs exactly the function required. The compiler composes large functions that are exactly the functions required for the design. Thus, the amount of interconnect wiring is minimized and the compiler is able to
- 6 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 achieve essentially the same densities as a hand-crafted design. Actual density comparisons of production designs developed with a silicon compiler show .that the total chip area per transistor and the percentage of chip area devoted to interconnect are comparable to the best hand-crafted designs. With the additional capability of doing exploratory design with the compiler, typical designs have the opportunity to be even denser than hand-crafted designs.
In the past, custom or semicustom designs could be justified by the" size and cost savings achieved in the overall system. These designs did not themselves have to compete solely as stand-alone circuits. They could not compete because of the density penalty involved in the design technique. Seldom, if ever, did one see merchant-market standard parts designed with the same design tools used for custom or semi-custom design. The picture changes dramatically with a silicon compiler. The compiler implements integrated circuits of "merchant-market quality". The result is that custom parts need no longer be much more expensive than equivalent functionality standard parts. In addition, standard parts can now be implemented with the time to market advantage of the most effective custom design tools.
APPLICATIONS AREAS A silicon compiler has a very broad area of application. The unique combination of design efficiency and silicon efficiency makes a silicon compiler appropriate across a wide range of complexities and volume. Figure 7 illustrates application areas as a function of both volume and complexity. At low volumes and relatively low complexities gate arrays are an appropriate implementation technique. They are, however, quite limited in density and complexity and are not as cost effective as compiled or hand-crafted implementations. At slightly higher complexities standard cell systems become more desirable than gate arrays. The somewhat larger cells of a standard cell system and the greater density compared to gate arrays makes them more attractive. At the extremes of ultra high volumes and complexities hand-crafted designs will always prevail. Across the broadest range of volumes and complexities, however, a silicon compiler is optimum. Its high densities make it preferable to gate arrays or standard cells at moderate to high volumes. Its rapid design capabilities, including exploratory design, make the silicon compiler the only feasible tool for high complexity designs. While gate arrays will persist as a viable technology for low complexity, low volume, quick turnaround time, the growth of gate arrays into higher and higher densities will be limited by the much more attractive silicon compilation technology. Standard cell systems are apt to be replaced entirely by the much more capable silicon compilation technology.
REAL DESIGNS
- 7 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Parl< Drive / San Jose,' CA 95131 / (408) 971-9000 / Telex 171973 No design technology can be proven effective without actually using it. Prototypes of Silicon Compilers Incorporated's compiler have actually been used to develop several major VLSI components. These components illustrate the power of Integrated Systems Technology: dramatic reductions in development time and very high density implementations. Figure 6 shows the MicroVAX 1 (tm) component developed by Silicon Compilers Incorporated for Digital Equipment Corporation. This part brings the VAX (tm) computer out of the computer room and onto a desk top. The entire VAX processor, including virtual memory management, is contained on just two Quad boards. Development took just seven months from contract inception to working silicon. Transistor density is comparable to that of the best commercial microprocessors on the market. By simultaneously developing both the MicroVAX system and the custom part. Digital Equipment Corporation was able to bring the final product to market in approximately one year versus the multi-year typical development cycles associated with other custom or semi-custom approaches.
Other components developed by Silicon Compilers Incorporated include the world's first commercial Ethernet (tm) controller and a RasterOp graphics chip. The Ethernet controller is sold under license from Silicon Compilers by SEEQ Technologies Inc. The RasterOp chip allowed Sun Microsystems Inc. to reduce the size of their advanced workstation product and bring the product to market just ten months from the beginning of its definition.
SUMMARY Integrated Systems Technology, based -upon silicon compilation, allows the systems engineer to exploit the capabilities of Very Large Scale Integrated circuits. Without retraining to become an IC designer, the systems engineer can produce components with densities that arefully competitive with commercial hand-crafted designs. By addressing the entire design cycle and, in particular, focusing upon the definition and architecture phases, the total time from concept to market is dramatically reduced. Silicon compilation brings a totally new level of design capability to systems developers - true Design Automation. The impact will be a dramatic increase in the number of custom designs and significant reductions in time to market. Standard part suppliers will also benefit from the use of silicon compilers. For the first time they will have available design tools capable of rapidly designing merchant-market-quality parts.
VAX and MicroVAX trademark Digital Equipment Corporation Ethernet trademark Xerox Corporation
- 8 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o oT c (D (0
o O -1 •o o
> OT Custom IC Design Market
Q. • »• —1 •< 1980 1985 1990 > p ($8B) ($17 B) ($43 B) z ro' STANDARD PARTS STANDARD PARTS
O o 3 •D 0) 3
lo I CO o VO 11 Q! Q. o —1 T3 0) MOS CUSTOM OT GATE ARRAYS MOS CUSTOM ' MOS CUSTOM GATE ARRAYS O (A GATE ARRAYS mci^m^ ASIC = 38% ASIC = 48%
ASIC = Application Specific Integrated Circuits I Source: Dataquest, incorporated 1982
<£> O 01B O O Silicon Compilers Inc.
1
> W c Design Leverage Connparison !T (A Q.' B)"
> DEFINITION DECOMPOSITION VERIFICA- b AND AND CIRCUIT FLOOR TlON RE- TAPE ARCHITECTURE LOGIC DESIGN DESIGN PLAN LAYOUT (ERC-DRC) WORK OUT CLASSICAL o 3 DESIGN •o a> •<
IS3 CO O 3) Q: Q. CD
1) 0) CAD/CAE
w 0) o > CO SILICON Ol COMPILATION o•t^ 00
(D
CO CO Figure 2 o 01
•a o
(D Q.
03 c Silicon Compilation-Overview CT
Q.' to'
> b • Complete Design Automation
CD 3 o o • High Level Micro Architecture Input 3 •o D) rj • Synthesis of All Required Views i;S ' (O O h- - Layout Geometry I • - Functional Simulation Models I - Timing Analysis Models D Technology Independence at 3 • Optimum Combination of o > - Design Efficiency to - Silicon Efficiency o•c- 00
S 068 o o Silicon Compilers Inc.
2. (D X
--I _L CO u•^ Figure 3 o £. nDT c (D 0>
"• O
(D a. > 03 Design Efficiency: Implementation Medium 9. 0)'
> b o o 3 •o 01 3 ro (O o Q! D. CD TJ 0)
O)
O (A a> >O CO 01
I COMPLEXITY
<£> O O O 025 Silicon Connpilers Inc.
CO ~g u Figure 4 o a
3 O o •3 o Q. Design Efficiency: Complexity Management > w c CT cn a 0)"
> O INTEGRATED SYSTEMS CD TECHNOLOGY o o 3 •o fi>
g M
Q! Q. I (D -t TI 0) -, a CAE/CAD w fi] D o >
o (O 024 Silicon Compilers Inc.
£5.
-NI —i. CO
Figure 5 OJ i. 3 O)
- 14 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o » n5) c (D
1
> Applications Areas w c cr <2 5! 0)'
> ECONOMIC LIMIT b
O o 3 •D 0)
(O O t-" UJ 5 "^ Q^ I Q. ' 3 (D •D fi) O > m [U
HAND o > CRAFTED
-COMPLEXITY I
55. m X ~J
CO ! 0) Q_ O U o OuO
- 1 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o a n c (B
O o
> w Integrated Systems Technology:
a. Dj'
> O
o o 3 VLSI System Technology •D 01 M • (O for Systems Engineers O l^J a. Q. CD T3 0)
w 0>
o >
o.tw 00 oCO o o °^° Silicon Compilers Inc. !5. (D
(O a (U ST .o c O) 5(A" o -t •o o
a.
u> c a- Today, Competitive Systems Require... (0 5! u' -I o > o 2. (A CD o O 3 XJ a •< Maximum Use of Very CQ O U) 3J a • Large Scale Integration a •o Q)
w tu o (0 CD o >
o o o 060 Silicon Compilers Inc. 2. (D o 2- a c (D
?
(D a. > W cr (fl VLSI Benefits
> 6
0) I • Smaller Size 0) 3
N3 O<0 .it.. i • Lower Cost I • Higher Performance 01 Higher Functionality I • Higher Reliability (A to >O
o CO
(O o 8 061 Silicon Compilers Inc.
(O ~4 CO D
O O •o O
> CO Standard Products are Limited c cr (fl S. 0)'
> P !5. (A (D Force Fit for Applications o o 3 •D 0> Do Not Offer Product Uniqueness
§ J, • Difficult for Vendors to Define Q. I CL Sufficient Functionality and Generality a> TJ (B o Best for Memories and Microprocessors
05 m Insufficient for Optimum System Integration c_ o CO _(t) >O The Microprocessor Revolution is Over
o 00
CO
CD ' 8 062 o Silicon Compilers Inc.
CD X
-~l _L CO ~J CO D lu
o o •3 o
> Custom...The Next Systems Revolution en c CT w 9! 0)'
> b o • Required for a Competitive Edge o 3 T3 0) • Silicon Density Allows i5 ' O Jl Truly Integrated Systems
Q. a a> • All Of the Benefits of Very TJ 0) Large Scale Integration
w • Foundries Make Production Available 3 o >
I
§ 063 o o Silicon Compilers Inc.
2. ID X
->l CO o 0)
3 O o -\ •o o CD Q. > Custom IC Design Market' CO
I 1980 1985 1990 a ($8B) ($17 B) ($43 B) > MASK o .STANDARD PARTS STANDARD PARTS PROGRAMMABLE o o 3 •o 0> 3
I l\5
MOS CUSTOM w GATE ARRAYS MOS CUSTOM MOS CUSTOM GATE ARRAYS 'GATE ARRAYS _(D o > ASIC = 26% ASIC = 38% ASIC = 48%
ASIC = Application Specific Integrated Circuits I Source: Dataquest, Incorporated 1982 o(O 018 o o Silicon Compilers Inc.
to u~J D 2. S' n c (D
E
w Barriers to Integration c cr (0 a! ni'
> b 5 • VLSI Design Requires I - Skilled IC Design Craftsmen I - Long Lead Times
N) I Compromise Design Approach
W - Limited Density 0> 3 - Limited Tools o >
i
CO •'- 8 065 o Silicon Compilers Inc. -I SE. (B _!X . ->l to -^l W o S n c (0
^
> w The Missing Link c o- (fl a! B)' •< o^ > p Z (D' o o 3 INTEGRATED SYSTEMS TECHNOLOGY 1^ (O I o 3) kO that allows systems engineers to rapidly d Q. I (D and efficiently design electronic end products —t TJ 0)
CO 0> o >
o 00 o 8 066 Silicon Connpilers Inc. 2. (D to •>! CJ a Si a •O
3 O o •a o
> W Integrated Systems TechnolcDgy Goals c cr en S oj'
> o
(0 (D 3 • Exploit Silicon in Systems o o 3 "O • Obtain Size, Speed, Cost, Power, DJ ;; I Uniqueness Benefits of VLSI Jg I- o o I3] ' • Use by Systems Engineers • Rapid Design
W lU 3. Efficient Silicon o > Process Portability
o*. 3 (O -4
S 052 o Silicon Compilers Inc. !E. (D X
-^l
CO ->4 CO u w ST •O
o o •3 o
> b
m THE KEYSTONE OF INTEGRATED o o 3 •a SYSTEMS TECHNOLOGY DJ
-» I i>j
^3J ^ • Complete Design Methodology and Tools Q. ' (D -o 0) • Matched to Real Engineering Process • Encapsulates IC Design Knowledge CO 0>
C_ o (fl p • Optimum Designs by Systems Engiriiiers o >
.b. O 00
8 067 Silicon Compilers Inc. 5.
> c Silicon Compilation CT (A Q! 0)' -3 o_ > b
(B o o 3 •uo
_fc ' N3 «3 I-" IS *NOT Computer-Aided Design O (vj 3) K I Q. CD IS NOT Computer-Aided Engineering
0) D> 3
o >
oJ^ 00
§ 044 o Silicon Compilers Inc.
•vj CO CO a to c (D (A
O O E
(D a « Silicon Compilation c cr — a »• -1 o > b
o o 3 •o D)
_. I lo (O 1-1 ° w IS Design Automation X £ I a. CD •a 0)
CO D)
o >
o.b. 00 g 045 Silicon Compilers Inc. it (D X -^l
O O •o O
> W Silicon Compilation-Overview a 0)'
• C^mpMe Design Automation
•Io • High Level Micro Architecture Input D> •< H I IVJ • Synthesis of All Required Views to I- I "^ - Layout Geometry I - Functional Simulation Models % - Timing Analysis Models w • Technology Independence 3
o Optimum Combination of > - Design Efficiency - Silicon Efficiency
oCO o o 068 Silicon Compilers Inc.
to -Nl CO o 5J. S^ .a c (D U O O E
> w Design Efficiency I >p o
•D fi>I • Hierarchically Structured
22 I - Complexity Management N) (O I—I ?*" d. I • Rapid Design Feedback Q. Ja> - Exploratory Design
• Interactive and Incremental CO V a> e_ o - Iterative Design a O
I o§ o ''^ Silicon Compilers Inc Ell
CO a Si a r> c m
3 O o •• o
> (n Design Efficiency: implementation Medium Q. a •3 o > p 2 S'
O o 3 •o D> 3 ^ I S t- O ON
CL I Q. (D -, •D 01
CO 0) 3
o >
COMPLEXITY i 025 5. CD Silicon Connpilers Inc.
(O •~J w fil i3
O O •3 o
(D a Design Efficiency: Complexity Management > CO
a D)' •- p z 2. INTEGRATED SYSTEMS (A / 20 \ (D TECHNOLOGY o /MODULESX ••• 3O •o 01 3 / 100 BLOCKS \ LEVERAGE LINE I 2,000 CELLS \ 20,000 GATES
CAE/CAD / 50,000 TRANSISTORS m
O VD 1,000,000 POLYGONS o > CO
o.b. OD i 024 oto Silicon Compilers Inc. o o
2. (D X
(O VI CO o 0)
o o •3 o
(D a. 03 Silicon Efficiency c CT (O a 9?
> O
it CO ID o 3 •o o> Builds Exactly the Functions Needed 3
<0 1-1 °"^ 00 • Optimum Implementation 51 a. I o. CD • Minimum Interconnect Wiring Merchant Market Quality Implementation w
o 00
to "^~*^" 8 055 o Silicon Connpilers Inc. 5.
CO -si w o 01
3 o o •3 o
p.(D > V) Silicon Efficiency Analysis c —
> p JAPX 80286^" MicroVAX'" 1 z 2. c» O Die Area (sq. mils) 119K 97K 3o •o at 3
•~ I # of Transistors 52K 37K S M o VO 3] 5! I a # of Blocks 55 18
W % Area for Routing 38% 43% O (fl _(D O> CO
o 00 * Normalized to the same feature size. ROM not considered on 286.
S 048 o Silicon Compilers Inc. o 5. (D
1
> VLSI Design 05
o. BJ'
> b WISHES
(D AflCHFTECTUHE o (HTL) o o 3 •D 0) SiUJi FUNCTIONS -I I o to NJ O LOGIC 3) u. Q.' a UJ n> T) g GATES 0) w UJ Q TRANSISTORS
POLYGONS I FIXED ARRAY FIXED VARIABLE PARAIÊTRIC HAND qELL LIBRARY CELL LIBRARY FUNCTIONS CRAFTED
SILICON EFFICIENCY i o(O 022 o o Silicon Compilers Inc.
(D ~g CO o s J2 C (D
E
w c Applications Areas IT CO Q! 0)' •< o > 6 ECONOMIC LIMIT
o o 3 •D lU 3
to Sj^ g' UJ (D IE TJ D 0) -J O > « Bl C_ O (0 _(D O > HAND CRAFTED
.b. O OD -COMPLEXITY
CO o CO n c
o o •5 o
CD a
CO c Silicon Compilation: Myths and Reality (T cn a tu' •< o > 6 i • Myths I - Just a Research Topic I - Limited to Fixed Architectures < _ - Someday . . . ro (O to ° to - Spec In - Chip Out 3 CL I - The Panacea a. • Realities - Proven Capabilities CO 0) 3 - Industrial Strength - Available soon from o > Silicon Compilers Incorporated
O OS o 8 056 Silicon Compilers Inc. o
o >. o O o -I—'
o
- 23 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 mm^m WSIHWWWiW WW ij.ij.ft iH'ii "•LvULi VI! ^- 24 ,. Da«,„s, ,„,„,„„„,.... s„»s,»i., „, ..0. N..,„„ 0<,™„.„,,,« „™,, p„, ,„.,, 3,„ ,„^^,, ^,3,, ^^, (408) 971-9000/Telex 171973 o K- BT n c (D
E p O) c IT (0 QI D)"
> P z (D'
o o 3 •o B> 3 Graphics RasterOp Controller
aQ- I (D T3 0) • SUN Microsystems, Inc. tn 0) 3
O >
o 00
o 8 019
•vl CO U f a kuâ^saesBaaeflttaai
- 26
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CL v^ o O -4—' c CD E Q- o '3 O" LJJ "cO "tzf 'CD Q
en o
- 27 -
Dafaquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 - 28 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 D w n c (0
E a. > Silicon Compilers Inc.
Q. 0)'
> o
CO (t> O o 3 U INTEGRATED SYSTEMS TECHNOLOGY ro o VO ol Allowing systems designers a (S TJ rapid and efficient development 0) of electronic end products. w
O >
.o. o 00 g 049 o Silicon Connpilersinc iE. (D
Dataquest
EDA FOR PRINTED CIRCUIT BOARDS
Buck Feltman Vice President of Marketing Cadnetix Corporation
Mr. Feltman is Vice President of Marketing at Cadnetix Corporation. Previously, he spent 17 years at Texas Instruments in Sales and Marketing and Systems Engineering, where his most recent position was Division Manager of Gate Array, Standard Cells, and Customized Cells. His earlier experience at Texas Instruments included a position as head of the New Business Development Center for Market Operations. In this capacity, he planned and implemented a program that resulted in placement of 12 world wide Regional Technology Centers for Texas Instruments. Mr. Feltman received a B.S. degree in Electrical Engineering from Auburn University in Alabama, and an M.B.A. degree from Rollins College in Winter Park, Florida.
Dataquest Incorporated ELECTRONIC DESIGN AUTOMATION FOCUS CONFERENCE May 10-11, 1984 Sunnyvale, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 EDA for Printed Circuit Boards Presented at Dataquest Focus Conference May 10-11, 1984 Buck Feltman Vice President - Marketing Cadnetix Corporation
Although the electronic design automation (EDA) market has been in existence since the early 1970's, the market has entered a new, higJi-growtli phase. While the Dataquest definition of EDA includes both integrated circuit (IC) and printed circuit board (PCB) engineering applications, the primary application of EDA equipment is IC development. The market for EDA workstations and has grown from nothing to $86 million in only" three years. This exceptional rate of growth has tended to overshadow the market for workstations in printed circuit board applications, which grew at a 100% annual rate over the same period to $80 million in 1983. Figure 1 shows the Dataquest projections as they have segmented the market. Although the projections recognize differences between PCB CAD and IC CAD, the segments are not differentiated in the EDA application, often referred to as CAE. A different segmentation of the market for electronic design automation equipment is shown in Figure 2, which also lists the more critical tasks associated with each application. The lists for printed circuit boards are very similar to the ones for integrated circuits, but the differences in equivalent items are not apparent. For instance, fault simulation is vastly different in the two situations as fault modes are entirely different. Likewise, simulation and timing analysis, placement, routing, and design rule checking have entirely different meanings for PC boards and integrated circuits. There are many differences in the two applications of EDA, and it is useful to examine each application segment, as the organizational benefits of electronic design automation must comprehend the entire design process.
Electronic Design Automation for Printed Circuit Boards Printed circuit board EDA tends to be perceived as a drafting tool and many view it as old technology. In reality, the problems associated with printed circuit board design are complex and are not solved adequately by the systems available today. The length of time required to complete a design from the initial concept is one the most critical problems facing electronic equipment manufacturers today. Complex systems may require years to complete. Both simple and complex systems may miss market windows if design and development takes longer than - 1 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ELECTRONIC CAD MARKET FORECAST
3000 CAD 2500 . CZIEEDE A $4087M PCS CAD 2000- S2Ba7Itt » 1500- IMIIXiON 1000 $926M 500 SSeiM |H
1083 1984 1085 1088 1087 1088 SOUnCE: OMaquMt
Figure 1. EDA Marlcet Projections.
DESIGN BNGINEERIMG (EDA) PHYSICAL LAYOUT (CAD)
Schematic Entry Autolayout PRINTED Simulation & Design Rule Checking CIRCUIT Timing Analysis BOARDS Fault Simulation/ Test Design
Schematic Entry Auto-layout INTEGRAND Simulation & Design Rule Checking CIRCUITS Timing Analysis Fault Simylation/ Model Extraction Test Design
Figure 2. Electronic Design Automation Market Segmentation
- 2 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 planned. It is difficult to isolate any specific task in the design process as the cause of all or most of the problems, rather r the design process should be examined for taslcs which may be automated or streamlined. One of the more common problems, and one which spans the entire design process, is documentation. The maintenance of accurate records is a major problem which can be automated and/ as a result, streamlines the design process. Figure 3 depicts a generic design process and lists the documentation products of each step. The process usually starts with a loosely described concept, which is modified as the design progresses, causing design changes at all points in the process. The specification and functional design are often not well documented, and changes, regardless of the source, must be recycled through all design documents, requiring endless re drawing. Automated drafting of new documents saves some initial labor, but the real saving comes when changes are required. The draftsman makes only the change and does not have to re-draw the document. In many organizations, documents are not re-drawn every time a change is made, rather, changes are accumulated for a period, resulting in inconsistent records being used by designers and others involved in design. Automatic PCB design systems also can significantly impact the time required for building breadboards and for developing PCB artwork. The physical layout of a breadboard may be performed on a CAD system for automatic wire-wrapping, eliminating a cumbersome manual task. Physical layout of a simple PC board may be reduced from the four to eight weeks required for manually developing artwork using mylar and tape, to a week or less to perform the same task on a CAD system which produces photoplot artwork on film. Design verification is another area where design cycle time may be improved. This problem has two parts, the easier one involving the consistency between the engineer's schematic and the resulting layout. Automatic PCB design systems can compare the schematic to the layout for consistency, but in most Operations, the layout draftsman also enters the schematic. The solution is to provide the design engineer with an entry device SO that he develops the schematic and keeps it up to date. This has not been practical in the past due to the cost and low functionality of available equipment. Several companies offer low-cost products based upon personal computers, but none of them provide a complete solution. The primary criticism relates to their limited processing power, prohibiting reasonable implementations of functions such as simulation, timing analysis or PCB layout. It is feasible to use personal computer-based systems for front-end schematic entry devices, interfacing to other systems for the complex functions, but this configuration creates a problem. The need for back-annotation of changes made during the layout process requires close coupling between the systems, which is an
- 3 -
Dataquest incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 DOCUMENTATION PRODUCTS CONCEPT 1 SPECIFICATION WRITTEN PRODUCT DESCRIPTION t FUNCTIONAL BLOCK DIAGRAM DESIGN T LOGIC SCHEMATIC DRAWING 1 DESIGN f DESIGN DESIGN FINAL SCHEMATIC DRAWING HOD. EVALUATION PRELIMINARY MFRG. PKG. - LAYOUT DRAWING '' _ ECN PHYSICAL - ASSEMBLY DRAWING DESIGN - DRILL DRAWING - DRILL TAPES - AUTO-INSERTION TAPES ' ' - PHOTOPLOT TAPES ^ ECN PROTOTYPE - BILL OF MATERIALS CONSTRUCTION T ^ ECN ALPHA OPERATING PROCEDURES TEST T ECN PILOT FINAL MANUFACTURING PKG. PRODUCTION T , ECN BETA PRELIMINARY USER'S MANUAL TEST PROBLEM REPORTS * ^ ECN VOLUME FINAL USER'S MiÛAL PRODUCTIO N FIELD SERVICE REPORTS
Figure 3. The Electronic Hardware Design Process
- 4 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 impracticality for competing vendors. The more efficient approach is to offer a low-cost terminal integrated with the full-function CAD system. The more difficult tasic in design verification involves engineering design and development. The task today is performed by building breadboards and using a logic analyzer to test the circuit. It is certainly feasible to use computer simulation to model the circuit and perform the same tests, and integrated circuit design relies very heavily upon computers for design verification, due to the high expense of fabricating prototypes. Breadboarding has been the lower cost alternative for printed circuit board design and, consequently, computer simulation has not been widely accepted for this purpose. The situation is changing, however, as new circuit technologies are raising the cost of breadboarding for design verification. Several trends in equipment design are changing the nature of printed circuit board design and the required design tools. Semicustom VLSI circuits, which are rapidly growing in use, obviously must be incorporated into PC board systems, but have relatively long design cycle times. The semicustom IC must be designed concurrently with the PC board, otherwise, significant delays may be incurred. The breadboarding approach to design and test ensures that design will be dependent upon IC availability, and will be delayed until functioning prototypes can be designed and fabricated. The use of computer simulation eliminates this dependence and allows design and test of both the PCB and semicustom IC to proceed concurrently.
The use of semicustom VLSI circuits also introduces some new problems into the design task, causing the transition from EDA to CAD to become fuzzy and the engineer's role to increase. Due to the complexity of VLSI, large, high pin-count packages must frequently be used. Pin-grid arrays, chip carriers and small- outline IC's address this need but require the use of multilayer PC boards and fine-line geometries for PC board signal traces. This, in turn, introduces problems of cross-talk and interference between signals, which requires engineering interaction in the layout process. The packages required for VLSI also tend to dissipate large amounts of heat relative to their surface area, causing thermal problems when several VLSI devices are used. Again, engineering expertise is required to solve a layout- related problem.
Many electronic systems also tend to use large board sizes and/or dense component placement. Both raise the complexity of PC board design to the extent that manual artwork development is no longer practical. Automation of the layout process is required to perform the task within any reasonable period of time.
- 5 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Printed Circuit Board Design Automation Solutions An obvious solution to tlie design documentation problem is the integrated system, which provides the necessary capabilities at each step in the process, from specification to manufacturing. Each technical discipline involved in the process has unique requirements of hardware and software which must all be tightly linked into a cohesive, comprehensive system. The system designer needs block diagram drafting capability, functional-level simulation and text processing. The circuit designer requires the same capabilities in addition to logic simulation, timing analysis and fault simulation. The FOB layout draftsman needs automatic placement, automatic routing and design rule checking. The manufacturing operation requires assembly drawings,' drill drawings, NC drill tapes, auto-insertion tapes, photoplot tapes, bill of materials and automatic test files. In terms of hardware, the engineer's need may be satisfied with a high-resolution, black & white CRT and the substantial processing power required for simulations. Layout draftsmen need a high-resolution color display with significant processing power for automatic operations. Both disciplines require relatively large amounts of mass storage capability and, for efficient communications, a local area network link to all workstations. Peripheral devices are also required by all workstation users for archiving purposes and hardcopy output. A local area network should provide the capability for all users to share the same peripherals, reducing the average system cost per user. An additional benefit of stand-alone workstations used with networking relates to the cost of implementing a complete design automation system. It is wise to convert a layout drafting department gradually, to minimize inconvenience and risk. Full- function workstations may be added at any time, requiring capital investment only for the additional system. Host-dependent systems, in contrast, require the significant investment in a host computer before the first workstation may be used. Additional workstations do not add any processing power, and consequently, degrade system performance as they are added to the host. Further, when the host is inoperative, all workstations become inoperative. Stand-alone systems have full capability without the host so that additional workstations have no effect upon performance, and the loss of one system, or even the network communications link, has no effect upon other workstations. The networking concept also raises an interesting possibility for the processing power requirements of simulation and automatic layout. Special processing engines may be developed, which provide dedicated hardware, optimized for each specific task, at a cost far lower than general purpose hardware. The network allows access to all users, again keeping the average system cost low while providing extremely high performance for every user.
- 6 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 The Cadrietix Approach Cadnetix Corporation currently offers the CDX-5000 Automatic CAD Worlcstation and the CDX-5001 Interactive CAD Workstation for the physical design of printed circuit boards. Both systems provide 32-bit performance with high-resolution graphics display. The user interface is object-oriented and has seven action keys with three keys on an optical mouse to provide 80% of the required functions. Keystrokes are minimized and learning is made easier by keeping all system operations consistent. The CDX-5000 provides automatic signal trace routing and both systems generate complete manufacturing data. Next week, the Ethernet local area network and a dedicated file server will be announced, which provide communications and file sharing between workstations, peripheral device sharing and additional mass storage. This sunmier, an EDA workstation will be introduced, providing a tightly-coupled schematic entry and engineering analysis tool for printed circuit board design. This new product will complement the present CAD products to provide a complete front-to-back solution for printed circuit board design and development.
Conclusion While the functions and operating sequences required to support EDA for integrated circuit applications are very similar to those for printed circuit board applications, there are significant differences. Technical differences and the available alternatives may be distilled to comparisons of functional value. The value of schematic entry is high for both IC and PC applications due to the benefits of automating documentation management.' Simulation and timing analysis have high value to IC development, but only moderate value to PC development, due to the relative cost of available alternatives. Likewise, fault simulation and test program development have high value in IC design, but low value in PCB design. In summary, EDA has a lower functional value for PCB design and is therefore more price sensitive than EDA for IC design. This riequires low-cost engineering workstations to meet cost requirements and tight coupling between EDA and CAD to maximize system functionality. No vendor presently offers a system addressing both the EDA and CAD problems for PC board development, but Cadnetix intends to be the first with product announcements scheduled for the next three months.
- 7 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Dataquest
THE WORLD ACCORDING TO VENTURE CAPITAL
A. Grant Heidrich General Partner Mayfield Fund
Mr. Heidrich is a General Partner at Mayfield Fund, which is collectively involved in the management of nearly $200 million of venture capital. Mr. Heidrich has worked closely with early stage funding of computer and health care companies. He represents Mayfield Fund on Silicon Graphics' Board of Directors. Prior to joining Mayfield, Mr. Heidrich was a Vice President of Wood River Capital Corporation, a New York-based venture capital fund. Mr. Heidrich received an undergraduate degree in Human Biology from Stanford University and an M.B.A. degree from Columbia University.
Dataquest Incorporated ELECTRONIC DESIGN AUTOMATION FOCUS CONFERENCE May 10-11, 1984 Sunnyvale, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, GA 95131 / (408) 971-9000 / Telex 171973 What's All The Excitement About or How Can An Industry Grow So Fast
Gerard H. Langeler MENTOR GRAPHICS CORPORATION 8500 S.W. Creekside Place Beaverton, Oregon 97005-7191 (503) 626-7000
The market for Computer Aided Engineering products didn't exist in 1981, yet many industry observers predict a market size of over $1 billion by 1987! How can an industry grow SO fast? To answer this question we need to look at what is fueling the growth, and just as importantly the limits to growth. By focusing our attention on the limits to growth, we can determine whether all this euphoria is realistic or whether this market is just a high technology flash in the pan. The limits to growth for any industry or company are a direct function of both internal and external elements. Internal limits center on such things as cash position, manufacturing capacity, market coverage, management ability and something the Boston Consulting Group calls the "Maximum Sustainable Growth Rate" which is approximated by return on net assets minus the interest rate on debt times the debt to equity ratio. External limits are focused on such issues as degree of pent-up demand, economic cycles, capital avail ability and most importantly the return on net assets a customer can expect from newly acquired CAE systems. It turns out that in the short term (e.g., next one to two years) the limits to growth for the CAE industry are largely internal, not external! While the significant CAE players are public companies with excellent cash positions, and all have such large capitalizations that return on assets is not yet a meaningful number, the other internal limits are very much in play.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dramatically lengthening semiconductor lead times put an artificial constraint on manufacturing capacity in the short term. The leading CAE companies have relatively small sales and support organizations (though growing fast) so market coverage limits industry growth at present. Finally, some CAE companies have the management experience and depth to handle over $100 million in sales, growing at over 100% per year. Some will struggle under such pressure. In contrast, there are few external limits to growth in the short term. There is a large amount of pent up demand, stemming from years of "shoemaker's children" phenomena among electronic engineers. We are in the middle of a strong upturn in the business cycle, led by capital spend ing.
Over the long term, the more enduring factors of return on investment come into play. Given the very high software content to the vlaue added portion of the business, internal return on net assets is likely to be quite high. Coupled with the traditional aversion to debt by successful high technology firms, the maximum sustaninable growth rate for leading CAE firms is likely to remain quite high. The real limit to long term growth in this industry will be found in the return on investment customers ex perience when they purchase a CAE system. Most customers focus on measuring productivity gains to determine the value of their investment. While productivity gains are very important in terms of a shorter time to market serving to lower costs and improve market share, there is a larger mostly unrecognized factor - price flexibility. If you give me a choice of getting to market three months early, or getting there on time with product in novation able to support a 5% or 10% higher price, I will normally choose the second option. CAE systems which not only promote productivity but also provide tools for in novation allow for such gains. This kind of thinking can lead to sophisticated an alysis which can prove beyond a shadow of a doubt that CAE is a good investment, often yielding 50% to 100% annual return on investment. However, even this may pale in com parison to the real return on CAE-survival. It is the
- 2
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, GA 95131 / (408) 971-9000 / Telex 171973 What's All The Excitement About or How Can An Industry Grow So Fast:
Gerard H. Langeler MENTOR GRAPHICS CORPORATION 8500 S.W. Creekside Place Beaverton, Oregon 97005-7191 (503) 626-7000
The market for Computer Aided Engineering products didn't exist in 1981, yet many industry observers predict a market size of over $1 billion by 1987! How can an industry grow SO fast? To answer this guestion we need to look at what is fueling the growth, and just as importantly the limits to growth. By focusing our attention on the limits to growth, we can determine whether all this euphoria is realistic or whether this market is just a high technology flash in the pan.
The limits to growth for any industry or company are a direct function of both internal and external elements. Internal limits center on such things as cash position, manufacturing capacity, market coverage, management ability and something the Boston Consulting Group calls the "Maximum Sustainable Growth Rate" which is approximated by return on net assets minus the interest rate on debt times the debt to equity ratio. External limits are focused on such issues as degree of pent-up demand, economic cycles, capital avail ability and most importantly the return on net assets a customer can expect from newly acquired CAE systems.
It turns out that in the short term (e.g., next one to two years) the limits to growth for the CAE industry are largely internal, not external! While the significant CAE players are public companies with excellent cash positions, and all have such large capitalizations that return on assets is not yet a meaningful number, the other internal limits are very much in play.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dramatically lengthening semiconductor lead times put an artificial constraint on manufacturing capacity in the short term. The leading CAE companies have relatively small sales and support organizations (though growing fast) so market coverage limits industry growth at present. Finally, some CAE companies have the management experience and depth to handle over $100 million in sales, growing at over 100% per year. Some will struggle under such pressure. In contrast, there are few external limits to growth in the short term. There is a large amount of pent up demand, stemming from years of "shoemaker's children" phenomena among electronic engineers. We are in the middle of a strong upturn in the business cycle, led by capital spend ing.
Over the long term, the more enduring factors of return on investment come into play. Given the very high software content to the vlaue added portion of the business, internal return on net assets is likely to be quite high. Coupled with the traditional aversion to debt by successful high technology firms, the maximum sustaninable growth rate for leading CAE firms is likely to remain quite high.
The real limit to long term growth in this industry will be found in the return on investment customers ex perience when they purchase a CAE system. Most customers focus on measuring productivity gains to determine the value of their investment. While productivity gains are very important in terms of a shorter time to market serving to lower costs and improve market share, there is a larger mostly unrecognized factor - price flexibility. If you give me a choice of getting to market three months early, or getting there on time with product in novation able to support a 5% or 10% higher price, I will normally choose the second option. CAE systems which not only promote productivity but also provide tools for in novation allow for such gains. This kind of thinking can lead to sophisticated an alysis which can prove beyond a shadow of a doubt that CAE is a good investment, often yielding 50% to 100% annual return on investment. However, even this may pale in com parison to the real return on CAE-survival. It is the
- 2
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company/1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 impression of many people (not just vendors) that failure to adopt CAE tools will spell economical doom for any firm competing in the fast moving electronics industry. With that attitude the growth potential for leading CAE firms is indeed spectacular. So, how fast can the industry really grow? Our best estimates are 200%-300% in 1984. 40%-70% per year in the 1985-1990 time frame and at least 25% to 35% per year during the 1990's. Along the way, you are going to see some interesting changes. First of all, during 1984 and 1985 some of the walking wounded will stop walking. Some will just go away. Others will merge or get acquired. For the long term leaders, look to Mentor Graphics, Daisy, one of the in strument companies (probably Hewlett-Packard) and one of the CAD/CAM companies (probably Computervision).
Rapid growth and rapid change, it makes for a most exciting business.
/tp
- 3
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CD
O CC Q_ o <=C CJ C_)
cr. o
1 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 CO
CC Q_
O C_) oo ce
CO CJ CD
CD z ^ ^ o I— or Q C3 LU C/l O I—I oc I
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o D>
•a o
?b VENTURECAPITALISTS a>
O Organization o 3 •a D) 3 Size of Funds l>0 (O d •f' Geography Q.
o >
.(^ o 00
oCO o o
to MflYFIELD -vl w >- .
U-)
oo CQ 1—1 ZD i—1 ••-H c --i CQ o Ll_ CQ VO "-\
-^y csI 1—1 CO •(rt 1—1 -^ CQ
<^ CO «• 1— _J ON • en CNI '"^ T— r^ CO 1— •o «i«» CO ON -t-J ON 1—I T— i~l CL. o
cC Ii. CJ
- 5 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 ezeut x9|Si / 0006-l.t6 (90t?) / IEI.96 VO 'ssop uBS / eAUQ >)JBd .leppid 062t / AuBdmoQ uesjajN 'OV lo AjBIpIsqns v 'peiBJOdJoouj jsenbBjBQ 1 - 9 .-
—J Ln 5 PENSION FUNDS ° o cz yo rj --J INDIVIDUAL
—I oo o VjJ FOREIGN
1—•• CO II -O la< j (-1- KJ-) CORPORATE CD • -n -O 1= Z3 C2. INSURANCE cn O c oo ro ON ENDOWMENT
VO (30 ao s> S3c (B CO
o o •o
03 c IT (fl QI a' -1 •<
o_ b
I MAYFIELD •o lU
§lo - Private Partnership 5 ' ai -J Q. £? I •D - $200,000,000 D>
Western United States
CO 0> 3 High Technology
o > Early Stage
o CO
oCO o o
ID ^HftYFIELD. CO •>i w o
•oa
> w REPRESENTATIVE INVESTMENTS
I Adept Technology, Inc. Genentech, Inc. ° > Atari, Inc. GRID Systems Corp. o Avantek, Inc. Linear Technology, Corp, o 3 Businessland, Irie* LSI Logic Corp. •o B) Cadnetix Corp. Megatest Corp.
Equatorial Communications Co, Tandem Computers, Inc. wo _
o CO
CoO o
(D MAYFIELD (O u->l § E c a U) rf 3 O O •o o
.^ > CO c I INVESTMENT CRITERIA «> >- ~~J -^ Management 2.~~ ~~s3 - Reasonable Valuation o >~~
S
~~2 _^ ^MAYFIELD. 3 o i c CD I/> ^* 3" go •o5~~
~~> w ac 0) 5! tf' 1 BUSINESS PLAN REVIEW > p z 5' Product Description o •° - strategy and Market Overview 3~~
~~§ , - Operations Plan M O ' - Management and Key Personnel~~
~~Financial Projections~~
~~Proposed Financing o >~~
~~(£> O O O~~
~~(0. CD~~
~~CO MAYFIELD ~J CO >- cC~~
CM
~~CO to OJ OJ~~
~~CO c: CO >> ^H a a. o +-> CT) 1—I tIT)o O) a. 4-> o OJ 1—t O) OJ c: c: a. u .>£ j^ "H o (= i~t u to OJ o CO O) 3 Q- C_) Z 1— CD~~
~~- 13 -~~
~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 .o Q. UJ~~
^
~~CO fsl isi OJ~~
~~c CN~~
~~(/) t/7~~
~~- 15~~
~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o a n c CD CO~~
~~•o o~~
~~CD Q. > CO~~
~~Q. B)'~~
~~> b TRENDS~~
~~CD 3 o - HIGH LEVELS OF ACTIVITY 3 T3 fl) i - LARGE REQUIREMENTS FOR CAPITAL l>0 (£> O~~
~~Q. a\ - MORE CORPORATE VENTURING Q. i? I •D 0) - MORE MERGER ACQUISITIONS~~
~~wU - HIGH TECHNOLOGY INVESTMENT BANKING o >~~
~~.o. o 00 ota o o~~
CD
~~53 . . , ^ . ^MAYFIELD~~
~~Dataquest~~
~~THE EDA IJUDVSTRY UNIQUE REQUIREMENTS MANDATE UNIQUE SUPPORT~~
~~Thomas O. Binder Vice President, Marketing and Sales Silvar-Lisco~~
Mr. Binder is Vice President, Marketing and Sales, at Silvar-Lisco. He is responsible for the company's field-based customer support resources. Prior to joining Silvar-Lisco, he was with Calma Company. Before joining Calma, Mr. Binder held various field sales and customer support responsibilities with AccuRay Corp., a supplier of computer-based process control systems. Mr. Binder received a degree in Mechanical Engineering from the University of Cincinnati and an M.B.A. degree from Indiana University.
~~Dataquest Incorporated ELECTRONIC DESIGN AUTOMATION FOCUS CONFERENCE May 10-11, 1984 Sunnyvale, California~~
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 The EDA Industry Unique Requirements Mandate Unique Support Thomas O. Binder Vice President, Marketing and Sales Silvar-Lisco Menlo Park, California
~~I. Traditional Product Support~~
The products supplied to the electronic industry by Electronic Design Automation (EDA) vendors require similar product support as technology based products which have been supplied to industry for decades. Their support begins with the delivery of quality products, which are the result of both high quality design and development and high quality product testing and production.
~~The quality product must be properly installed and the support provided to insure a rapid and effective start-up and implementation.~~
~~Training support is necessary to assist the user in rapidly becoming productive in utilizing the EDA System and achieving the maximum benefits.~~
In addition to the initial training support, continuing application consulting^-is neees&ary so the initial results can b& maintained and so that changing application requirement and personnel turnover can be accommodated without loss of effective product utilization.
~~- 1 -~~
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Occasionally the product will have some failures which must be corrected rapidly so as to prevent any significant loss of product utilization. Even worse, failures which are not corrected quickly will result in loss of user confidence in the product and a resulting tendency to reduce reliance on a tool which was originally acquired to improve productivity.
But these are traditional requirements which describe the majority of design and production tools. What about the EDA Industry and its characteristics and requirements? Surely this industry has unique requirements which mandate that unique support capabilities be developed and supplied with the EDA Systems in which you all have an interest.
~~II. EDA Industry: Unique Requirements~~
The EDA Industry and the users of EDA tools have many characteristics which make them unique and, thus, cause them to have unique requirements in the areas of product and customer support. The electronics industry has created the need for an EDA System that is truly "state-of-the-art" and this industry is changing faster than any technology in the history of the world. The users of an EDA System comprise a wide variety of manufacturers ranging from relatively small, localized enterprises to huge, multinational companies. -2-
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 And while electronic technology is changing rapidly, all other aspects of the electronic industry are changing just as fast.
~~All of these facets of the EDA user community mandate that the support provided by EDA vendors meet the resulting unique requirements.~~
~~III. Unique Support~~
~~A. Multiple Hardware Environments~~
The application addressed by EDA System cover a wide variety of tasks and a wide range in the magnitude of the tasks. Consequently, the overall EDA application must be partitioned into appropriate levels of computational capability.
In addition, you may want to change those functional assignments based on the size of a specific design. For instance, a logic simulation for a relatively small circuit, or cell, may be adequately accomplished on an engineering workstation. However, the same simulation for a larger circuit, say a large scale integrated circuit, would more efficiently be done on a supexmini, and a VLSI circuit may well require a mainframe for efficient logic level simulation.
~~-3-~~
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 To accoinmodate this unique requirement, support for your EDA System must be available at each of these levels of computational capability. This requires that the EDA tools can be implemented at each level, on a-variety of CPU types, and that full file compatibility and consistent user interface is available at each level.
~~B. Multiparty Relationships~~
Many users of an EDA System look to outside resource for the production of the circuits which they design and layout. The EDA System being used must be compatible with the fabrication processes and standard building blocks used by these outside production resources, typically referred to as "foundries".
~~This compatibility is assured through a multiparty relationship which is supported, and even developed, by the supplier of the EDA System with the user and the foundary.~~
The continuing support of this relationship by the supplier of the EDA System is necessary to maintain current compatibility with the foundary as the building block designs are characterized, expanded and modified and/or the fabrication processes are enhanced to take advantage of the latest technology advances.
~~-4-~~
~~Dataquest Incorporated, A Subsidiary of A.G. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 C. Multiple Design Methodologies~~
Because of the multiplicity of design and implementation methodologies, an EDA System must be able to support a wide range of applications. Since different parts of an entire system will be partitioned into different physical implementations and EDA System must support a full range cf physical device designs.This can include printed circuit boards, gate arrays, standard cells, standard integrated circuits, as well as others.
~~D. Changing Technology~~
Perhaps the most critical area of support is that necessary to accommodate the rapidly changing technology exhibited by the electronic industry. A supplier of an EDA System must provide the support necessary to allow the user to continually take advantage of the technological changes being experienced in the electronics industry.
As circuits in general become larger and more dense through advances in manufacturing technology, the EDA System must grow to allow design, verification, and automatic layout of the circuit. And as the new technologies impose new functional demands on the EDA System, ongoing enhancements must be provided through product development support to meet these demands.
~~-5-~~
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 For example, gate arrays are increasing in size from a few hundred gates up to 10,000 gates and beyond. An aggressive product support program which includes planned enhancements to satisfy there requirements is a critical element in the support program.
Another example would be standard cell technology which has seen the advent of irregular cells, two-layer metal processes, and two-layer metal/one-layer polysilicon processes to improve circuit performance and density. Support of the EDA System used for the standard cell layout must include enhancem.ents which accoimnodate these technological advances and allow the EDA System to have a long., productive life.
~~E. Multinational Relationships~~
The electronic industry is international in scope and is getting more so every day. Large systems companies have established facilities worldwide and need to have consistency of EDA tools throughout these widespread resources. Second-source agreements between merchant integrated circuit vendors have taken on a true multinational flavor with many combinations of U.S., Japanese, and European suppliers in cooperative ventures.
~~And integrated circuit merchants themselves are addressing the international markets by establishing design centers in other countries.~~
~~-6-~~
Dataquest incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 The EDA System selected to implement these strategies must be implemented in these different geographic areas; and consequently, the EDA System must be supported in their different geographic areas.
~~Support at the local level, whether it be for installation or training or maintenance is mandatory to achieve the necessary level of system utilization.~~
An example is a merchant integrated circuit supplier located in California, with a design center in Europe and a design center in the Far East. All three locations require local support for the common EDA Systems used at each site.
~~IV. Summary~~
~~The EDA Industry exhibits certain characteristics which result in a unique set of support requirements. These include:~~
~~A. Multiple Hardware Environments B. Multiparty Relationships C. Multiple Design Methodologies D. Changing Technology E. Multinational Relationships~~
These unique requirements apply equa'lly well to users of EDA Systems in a wide variety of situations. And the underlying industry characteristics which mandate these unique support requirements will be with us for a long time. -7-
~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973~~
~~Dataquest~~
~~COPING WITH A CHANGING INDUSTRY AND CHANGING TECHNOLOGIES~~
~~Harvey Jones Senior Vice President of Marketing Daisy Systems Corporation~~
Mr. Jones is Senior Vice President of Marketing and a founder of Daisy Systems Corporation. He joined the company in June 1981 as Vice President of Marketing. Previously, Mr. Jones spent seven years at Calma Company, where his last position was Vice President of Business Development. He received a B.S. degree from Georgetown University, and an M.S. degree in Management from Massachusetts Institute of Technology.
~~Dataquest Incorporated ELECTRONIC DESIGN AUTOMATION FOCUS CONFERENCE May 10-11, 1984 Sunnyvale, California~~
~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o 0)~~
~~05 C cr a oj' •3 o > b o o 3 •D U 3 to (O o Q. Q. 01~~
~~T3 lU~~
~~W 0)~~
~~O >~~
~~o.1^ 00~~
~~CD oO o~~
~~•oa OJMi a*^Tiris ccf* > W c cr HOW BROAD IS THE U) MARKET? a 0)' -3 > b • All Electronics Technologies: o Printed Circuits, Gate Arrays, o 3 Standard Cells, VLSI I IO~~
~~Q. Q. (D~~
~~Tl B • All Electronics Market Segments: Computers, Communications, CO 01 Military and Aerospace, Consumer o > > • Worldwide Geographic Distribution: i~~
~~CO U.S., Europe, Japan o o o~~
~~ID ->4 OJ I 3 /^ "\ O o-^ •a O nESH S^TtTTTbCCW~~
~~0) c zr a> a CAE VALUE ADDED 5' •5 > 1982 b~~
~~ro U) n>~~
~~O o 3 •o D) •< IV)~~
~~a. CL Desfgn VMffleafkxi Layout (a~~
~~TJ lU , . ...V*' 'I' ii<'»<>Ul'~~
~~w 01~~
~~o > (D tn~~
~~o 00~~
~~oCO o o~~
~~m_ (D X~~
~~•>( CO - 4 ~~~
~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o » c~~
~~to s~~
~~I I I !5 o 00~~
CO w-J Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o aI-*- X) a r o •o o Serena ufw > cn VALUE ADDED IN SILICON LIBRARIES Q. tu' (Partial List)
~~> b • AMCC • Motorola o o 3 •D AMI National HI~~
ro
~~CL CDI NCR Q.~~
~~-a 01 EXAR NEC~~
~~C/5 OJ 3 Fujitsu RCA o > IMP Siliconix~~
~~o 00 LSI Logic CD O O O~~
1
~~> CO~~
~~lU >< o_ > b~~
~~o 3~~
~~to U3 O~~
~~Q. Q. to •0 0)~~
~~en 0) 3~~
~~O >~~
~~o 00~~
~~oCO o~~
~~CO w--J o E r "\ E~~
~~9«THT16aUW m c CT CA g! SIMULATION: 2ND GENERATION 0)" > O 1981 -Present O o 3 CAE Workstations~~
~~INJ (O I « Highly Interactive o ô 3 d I - Knowledge-Based OL (D TO lU 5^ o Hardware Simulators w Bl c_ «Increased Performance O w _(D o > - Increased Capacity~~
~~O 00 oto o o 01~~
~~(D O. > W~~
~~Q. 0)'~~
~~> b~~
~~o o 3~~
~~5 oi o. CD T3 01~~
~~w a~~
~~O >~~
~~o(C o o~~
(D
~~to -J CJ~~
~~# - 11 -~~
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dataquest Incprgorated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o (U c a> CO /" 5" "N o •3 .o . JWIMB 9i«TBT)5C0W ^ MEGALOGICIAN I .'•'•••If' > '','',,''_".-.i':>J;.' b 2. 0) (D O 1 Million Gate Capacity o 3 •o lU s lOx State of the Art 3
~~a. a. 100K Evaluations per Second (D -1 -D (U = lOOx State of the Art~~
~~(n a> •z Fully Integrated a> a> O > (O CJI IS ! S~~
~~g o iV.~~
(O
~~CO o a~~
E
~~(D a. > Elaisi3 CO c cr W Q.' B)' •3 THE NEED: o > in SYSTEM LEVEL VERIFiCATION o o~~
~~to o THE PROBLEM:~~
~~Q. Q.~~
~~W SJ • VLSI MODELS o > CD~~
~~o 00 oCO o o t^ .ffti-ft'î-'dtfîri** ^wnyw^Mt. # •»•/,~~
~~CD -vi u a S ffi •o~~
(O
CO Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 Dataquest Incorporaied, JT'SuBsidtary of k.li. Nielsen company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 - 18 -
~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o~~
~~r •^~~
~~n> ?•' > w dnisui c cr « • oJ 0)' -5 > PMX BENEFITS: o~~
~~o o 3 • COMPLETE SYSTEM VERIFICATION 5 Q! Q. I 0) ^ FUNCTIONAL, INTERACTIVE TESTER •D Dl~~
~~w 01 RESULT: o > • NO NEED FOR PROTOTYPE, o.1^ (RIQHT THE FIRST TIME) 00 o(O o o~~
~~-vi~~
~~CO Ci v D~~
o
~~> w c «CT Q: 0)' >< O > b~~
~~o o 3 •a 01 3~~
~~So 3P & CD Tl 0)~~
~~w 0)~~
~~o >~~
~~o•c. 00~~
~~CD O O O~~
~~•~J CD -J CO D~~
E
~~> w c a- Ul 5! S' w*TirTii:c«* o^ > o STRATEGIC CAE CONSIDERATIONS o o 3 •o u I ro Functionality ro (O o go d Q. (D Price/Performance -\ •D 01 Sarvica and Support W OJ D Product Migration o >~~
~~oCO o o tD^ CD X~~
~~CO ~J CJ - 22 - Dataquest incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 I3 O o -1 13 O~~
~~(D a. > w~~
~~Q. 0)'~~
> b (••MsuaAJ •~-^^~'^^—~^——^~—~^ FUNCTIONALITY o o 3 •o 0) • From Desiî Entry ii I O K) 31 W to Implemeatation i;. (D • Simulation of T3 ai Entire Systems w 3 • Design for Test o >
~~o 3~~
~~to to o a I n .t c ' 2. I oE'S •5 o~~
~~CD Q.~~
~~(/> C CT Q! B)' •< > b~~
~~o o 3~~
ro
~~Q. Q. CD~~
~~T3 0)~~
~~CO B)~~
O>
§
~~•vj~~
~~CO -s] CO D aSi. n c (D a o~~
~~W c cr (A OL 01~~
~~> b o o 3 •a 0)~~
~~I i a w 01 o~~
~~O > to Ol~~
~~o 00 i~~
~~CO -Nl CO - 26 - Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 o~~
~~> en~~
~~Q. D)' -2 o > p z 0)'~~
~~O o~~
~~w 0) 3~~
~~o >~~
~~o 00~~
i
~~CO CJ - 28 -~~
~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973 I ^J^~~
~~:^^~~
~~Dataquest~~
~~EDA INDUSTRY OVERVIEW~~
~~Beth W. Tucker Industry Research Manager CAD/CAM Industry Service Dataquest Incorporated~~
Ms. Tucker is Industry Research Manager for DATAQUEST's CAD/CAM Industry Service. In this position, she is responsible for analyzing and reporting on the electronics segment of the CAD/CAM industry. Prior to joining DATAQUEST, Ms. Tucker was Market Planning Manager of Calma Company's Microelectronics Division, where she was responsible for identifying market opportunities and developing marketing strategies for the company's electronic CAD product line. She later contributed to the development of marketing information systems at Calma's corporate level. Prior to that, she was a Senior Market Research Analyst with Monroe Systems for Business, where she was responsible for market research and development of small business computer products. Ms, Tucker holds an A.A. degree in Computer Technology from Morris County College in New Jersey and is doing course work in Business Administration at San Jose State University.
~~Dataquest Incorporated ELECTRONIC DESIGN AUTOMATION FOCUS CONFERENCE May 10-11, 1984 Sunnyvale, California~~
~~Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973~~