Silvio's Corner

GSE z/OS Expertforum Switzerland # 70, 31.3.-1.4.2009 ®

Silvio’s Corner z/News, Hints and Tips

Silvio Sasso IBM Switzerland ITS Service Delivery for z/OS sisa@ch.ibm.com

Zürich | March 31, 2009 © 2009 IBM Corporation Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation1 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 1 Thun Objectives

The objective of this news session is to provide you with up-to-date and last minute technical information, hints and tips related to IBM mainframe hardware and software, such as IBM system z9, zSeries, z/Architecture and z/OS.

This corner allows me to present you...

ƒHardware news: IBM hardware announcements, new features and options

ƒSoftware news: z/OS and Parallel Sysplex update

ƒVarious resources containing additional documentation (e.g. product specific information or other interesting Websites etc.)

ƒSystem programmer goodies, tools, hints and tips

ƒUseful technical news and flashes

ƒRecommended readings: new Redbooks and whitepapers etc.

ƒTips for education, workshops and conferences

ƒand much more...

Note: the information contained in this document has not been submitted to any formal IBM test and is distributed on an "as is" bas without any warranty either expressed or implied. The use of this information or the implementation of any of the techniques or tips described is a customer responsibility and depends on the customer’s ability to evaluate and integrate them into the custo operational environment. Customers attempting to adapt these techniques to their own environments do so at their own risk.

hints andis

mer’s Trademarks

The following are trademarks of the International Business Machines Corporation in the United States and/or other countries.

CICS* IBM* RACF* z/OS* DB2* IBM eServer S/390* z/VM* DB2 Universal Database IBM logo* System z9 z/VSE * zSeries* DirMaint IMS Tivoli* ESCON* NetView* TotalStorage* VSE/ESA* FICON* OMEGAMON* VTAM* GDPS* On Demand Business logo Parallel Sysplex* WebSphere* HiperSockets z/Architecture* HyperSwap

* Registered trademarks of IBM Corporation

The following are trademarks or registered trademarks of other companies.

Java and all Java-related trademarks and logos are trademarks of Sun Microsystems, Inc., in the United States and other countries Linux is a trademark of Linus Torvalds in the United States and other countries.. UNIX is a registered trademark of The Open Group in the United States and other countries. Microsoft is a registered trademark of Microsoft Corporation in the United States and other countries.

* All other products may be trademarks or registered trademarks of their respective companies.

Notes: Performance is in Internal Throughput Rate (ITR) ratio based on measurements and projections using standard IBM benchmarks in a controlled environment. The actual throughput that any user will experience will vary depending upon considerations such as the amount of multiprogramming in the user's job stream, the I/O configuration, the storage configuration, and the workload processed. Therefore, no assurance can be given that an individual user will achieve throughput improvements equivalent to the performance ratios stated here. IBM hardware products are manufactured from new parts, or new and serviceable used parts. Regardless, our warranty terms apply. All customer examples cited or described in this presentation are presented as illustrations of the manner in which some customers have used IBM products and the results they may have achieved. Actual environmental costs and performance characteristics will vary depending on individual customer configurations and conditions. This publication was produced in the United States. IBM may not offer the products, services or features discussed in this document in other countries, and the information may be subject to change without notice. Consult your local IBM business contact for information on the product or services available in your area. All statements regarding IBM's future direction and intent are subject to change or withdrawal without notice, and represent goals and objectives only. Information about non-IBM products is obtained from the manufacturers of those products or their published announcements. IBM has not tested those products and cannot confirm the performance, compatibility, or any other claims related to non-IBM products. Questions on the capabilities of non-IBM products should be addressed to the suppliers of those products. Prices subject to change without notice. Contact your IBM representative or Business Partner for the most current pricing in your geography.

General Information Mainframe Executive Magazine Build a smarter Planet The expanding System z Community Hardware News zHPF – High Performance FICON System z10 Design, Testing and Verification System z10 CPU Measurement Facility IBM DS8000 Storage Subsystem Update Software News z/OS Key Dates z/OS Support Summary z/OS Platform latest RSU IBM System z Innovations for a Dynamic Infrastructure Statements of Direction z/OS 1.11 Enhancements Preview z/OS Petabyte Capacity Enablement

Software News… EAV – Extended Address Volumes APAR's of Interest (HIPER's and Red Alerts etc.) Parallel Sysplex Update Possible Differences between old/new Duplex Structure Sizes IXCMIAPU Administrative Data Utility Enhancements for LOGR Policy System Logger Offload and Staging Dataset Sizing (follow-on) Hints and Tips, System Programmer "Goodies", Tools and Resources SMFPRMxx Parmlib Member: Prevention of possible Parsing Errors RSU Settings on System z10 GETMAIN Enhancements in z/OS 1.10 System z and High Availability: Cycle Steering GTF Trace to an external Dataset Redbooks and Redpapers Doc Juke Box

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation5 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 5 Thun GSE z/OS Expertforum Switzerland # 70, 31.3.-1.4.2009

Zürich | March 31, 2009 © 2009 IBM Corporation Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation6 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 6 Thun The Mainframe Executive Magazine

• A new magazine for CIOs and IT managers in enterprises with IBM mainframe systems • Don't miss a single issue!

rt • Sign up for your own free Subscription now at: a P r s ne a or ed C x d s bo http://www.mainframe-exec.com/ vi o' e ro vi k P il Ju S c of o • Some Topics of the January/February 2009 Issue: D z/OS versus Linux on System z: which is the right Choice? The Mainframe and Web 2.0 Mainframe Staffing: Truth and Consequences Straight Talk for Mainframe Executives: Conquering the Mainframe Skills Shortage Software Strategies: the distributed Computing Disaster – Time to roll it back! Re-thinking SOA: a Guide for CIOs

art s P d a er ide orn rov 's C x P ilvio ebo • Some Topics of the March/April 2009 Issue: of S Juk Doc Linux on System z hits the Mainstream Still not dead: More Evidence of healthy Growth in Mainframe Workloads Mobile Access to Mainframe Data How to reduce Mainframe Expenses and stay in Line with your Business Strategy Time to hunker up The green Benefits of Virtualization And much more…

Building A Smarter Planet

We know the world is becoming smaller… and flatter.

Something else is going on that may ultimately have a greater impact on business and society.

The world is about to become smarter.

This is quite literally about how the world works… the world’s infrastructure is becoming intelligent.

In 2001, there were 60 million transistors for every human on the planet ...

… by 2010 there will be 1 billion transistors per human…

… each costing 1/10 millionth of a cent.

In 2005 there were 1.3 billion RFID tags in circulation…

… by 2010 there will be 33 billion.

Worldwide mobile telephone subscriptions reached 3.3 billion in 2007 and expected to reach 4 billion by the end of this year

One billion camera phones were sold in 2007, up from 450 million in 2006 …

3G devices growing 30% annually.

An estimated 2 billion people will be on the Web by 2011 ...

… and a trillion connected objects – cars, appliances, cameras, roadways, pipelines – comprising the "Internet of Things."

The reason we will all begin to transform our systems, operations, enterprises and personal lives to take advantage of a smart planet isn’t just because we can.

It’s because we must.

U.S. CPG companies and retailers lose $40 billion annually due to inefficient supply chains.

In North America, up to 22 percent of total port volume is empty containers.

The Port of Jersey has 100,000 empty containers sitting in storage – worth nearly $200 million.

In the United States alone, 2.2 million dispensing errors are made a year because of handwritten prescriptions.

The U.S. healthcare system loses more than $100 billion a year to fraud.

In a small business district in Los Angeles, driving around for parking in one year generated the equivalent of 38 trips around the world, burned 47,000 gallons of gas, emitted 730 tons of carbon dioxide.

Congested roadways cost $78 billion annually in the form of 4.2 billion lost hours and 2.9 billion gallons of wasted gas.

In the U.S., a typical carrot has traveled 1,600 miles, a potato 1,200 miles, a chuck roast 600 miles…

…grocers and consumers throw away $48 billion worth of food every year.

In distributed computing environments 85% of computing capacity sits idle.

In six years the power consumption of a server has risen from 8 watts to more than 100 watts per $1,000 worth of technology.

On average, for every 100 units of energy piped into a data center, only 3 units are used for actual computing. More than half goes to cooling the servers.

We’ve thought about IT as the world of data centers, software, PCs, routers, bandwidth.

We’ve thought about infrastructure as the world of buildings, factories, hospitals, roads, pipelines.

Those worlds are converging.

We’re confident that the world can become smarter. We’re building it with our clients.

SMART HEALTH CARE

IBM is helping ActiveCare Network monitor over 12,000 clinics and provide over two million patients with the proper delivery network for injections, vaccines and other pharmaceuticals.

ACN is using IBM software to lower the cost of therapy by 90 percent and reduce time and cost required to develop patient and clinic applications by 60 percent.

SMART ENERGY

IBM leads 7 of the 11 smart meter deployments globally, building intelligence into utilities to lower costs for customers and better balance the grid. Pacific Northwest National Laboratory has helped homeowners reduce energy costs up to 10 percent by turning ordinary thermostats into day traders for energy, ensuring the best cost for the customer and better load balancing for the grid.

SMART SUPPLY CHAINS

IBM has helped companies ship smarter, with an innovative tracking system that does more than help shippers manage supply chains; it reduces empty cargo containers and monitors the condition of container contents.

The smart shipping system provides temperature and humidity readings, intrusion alerts, quicker customs clearance and more … all communicated wirelessly and shared over the Internet with partners.

SMART TRACKING

IBM and Matiq are developing a smarter food tracking solution, the first-of-its-kind in the Nordics …

… it uses RFID technology to track and trace meat and poultry from the farm, through the supply chain, to supermarket shelves.

SMART FINANCIAL SYSTEMS

IBM rebuilt the world’s currency markets - - one of the most advanced and complex exchanges -- to help understand risk in real-time.

SMART TRAFFIC

IBM is working with Brisbane, London, Singapore and Stockholm to deploy smarter traffic systems. At least 20 other cities have active bids to do the same. Stockholm has seen approximately 20 percent less traffic, a 12 percent drop in emissions and a reported 40,000 additional daily users of public transportation.

SMART LAW ENFORCEMENT

IBM is helping the city of Chicago fight crime by digitizing their law enforcement practices and deploying smarter surveillance systems …

… even testing a system that uses audio sensors to direct cameras to locate gunshots, determine the caliber of gun fired and pinpoint its exact location – long before 9-1-1 is dialed.

SMART WATER

IBM and the Nature Conservancy are finding smarter ways to manage water supplies … providing analytics and research capabilities on the behavior of watersheds and the impact of human activities on freshwater supplies.

IBM’s Stream Computing system will provide minute-to-minute deep analysis of New York's Hudson River via an integrated network of sensors, robotics and computational technology distributed throughout its 315 miles.

SYSTEMS SOFTWARE SERVICES FOR A SMARTER WORLD

More Information

Smarter Planet Website:

http://www.ibm.com/ibm/ideasfromibm/us/smartplanet/20081106/index.shtml

Link to additional Information on the Status and Progress of the Initiative:

http://domino.research.ibm.com/comm/research.nsf/pages/d.compsci.smarter.planet.html

Current Infrastructures pose significant Challenges

Web Servers

Security/ IBM System z SSL/XML Directory Appliances Servers Application Servers

Routers Switches File/Print Application & Servers Data Servers DS Servers Caching Firewall Appliances Servers

LAN Servers

• Rising costs of systems and networking operations • Explosion of data and information requirements • Difficulty in deploying new applications and services • Landslide of compliance & security requirements • Systems and applications need to be available • Rising energy costs & demand

Today’s Mainframe addresses these Challenges

The Right Technology Transformed Economics • Highly secure and resilient • Technology dividends • Huge new performance and capacity levels • Specialty Engines • Industry leading virtualization and • Low TCO for consolidation management • Highly responsive and flexible

The Right set of workloads Expanded Community • Ideal platform for consolidation • 100s of partners, integrators and ISVs •Linux®, Java™, WebSphere®, SOA • Over 1000 New applications on System z • Secure data serving in 2008 • New possibilities eg. Business Analytics • 5,500 applications • More than 1,000 new applications in 2008 • 2,800 on Linux on System z™ • More than 50,000 ‘mainframe graduates’

System z: The right Technology… 45 Years of Market Leadership

IBM System z World-Class The world’s most powerful Virtualization enterprise computing platform Secure and Large scale Resilient consolidation for savings Improved price/ Just in Time performance Capacity Mitigate the risk of of up to 80% in total cost security breaches of ownership compared 100s of Capacity choices Permanent capacity for to distributed platforms for the right size server non-disruptive growth Dedicated cryptographic Deploy servers, Business Resilience Temporary capacity for processors networks, and solutions LOW COST OF fluctuating workloads fast Industry leadership OWNERSHIP Interim capacity for Support for multiple continued operation capabilities and Leadership capabilities with certification operating systems IBM Systems software Policy based automation Dynamically optimize capabilities Where mean time The future runs on between failure is resources according to System z and the future Offerings can be business priorities begins today replenished dynamically measured in decades

The expanding System z Community

• Continued growth on the System z platform in 2008 – Over 150 new ISV partners and 1,000 new applications – More than 500 new Linux applications added – 174 new/upgraded WAS, DB2®, CICS® and IMS application/tools on z/OS • ISV Partner loyalty – Over 1,500 ISVs building applications for System z – Over 1,800 applications on z/OS® 1.8 and above (over 3,500 for all z/OS releases) – 2,800+ applications for Linux on System z – 86% of our ISVs maintain OS currency • Academic Initiative delivering mainframe skills: – More than 500 schools participating – More than 50,000 students trained – 30 courses and more to come – Student MF Contests – z Skills Help Desk – Over 200 z IBM Mainframe ambassadors – Assist Professors Simplified Energy efficient Virtualized Secure

Clients are using the Mainframe to address their Key Challenges today

• System z worldwide capacity growth for 2008: – 25% YTY growth – 68% specialty engine growth

• 70% growth in Linux on System z (IFLs shipped)

• More than 50 new accounts in 2008

• Since IBM launched the z900 in 4Q 2000, System z has nearly doubled its share, from 17% to 33% in the enterprise server segment1

• Clients in all sectors, all around the globe are expanding their use of the mainframe

1 IDC server tracker, Nov 08, , $250K+ servers

And the Industry agrees

“…IBM has reinvented the mainframe. The pursuit of new workloads has been a critical aspect of the resurgence, while at the same time Growth and protecting and nurturing the installed base. The introduction of specialty Innovation engines has been a catalyst to the growth but represents only the first step toward a significant change in architecture during the next five years.” Source: Gartner Data Center Conference, Dec. 2008 (Gartner The IBM Mainframe Platform Ongoing Challenges, New Opportunities)

“IBM’s z10 is the computer industry’s pinnacle systems platform. Blow-away performance (fast quad-core processors), increased capacity (1.5 times Cost Effective greater than its predecessor), and expanded memory (three times greater than its predecessor) makes this system the absolute best scale-up architecture — and Modern bar none — in the computing industry."

"For those IT organizations that have seen the light and are moving Simplifying their emphasis from the never-ending challenge of trying to optimize their IT infrastructure to the more important optimizing of the delivery IT of IT services, nothing beats the mainframe," Mike Kahn, The Clipper Group, February 2008

System z remains a growing product line for IBM because, during the past decade, it has morphed historical strengths into new forms aligned with today’s languages, operating environments, and computing standards. And, in fact, System z strengths such as mitigating business risk in many Manage forms have once again shoved their way to the fore of many IT departments’ priorities. This includes protecting against hardware and site failures, managing a company’s security policies and credentials, acting as a focal point for integrating the myriad applications that make up a business process, and Risk being able to adapt to sharp spikes in load. Gordon Haff, Illuminata, January 2009

System z: The Journey continues…

The next evolution of Integrated or IBM Mainframe network attached computers Accelerators Extend mainframe IBM System z10 - Delivering new qualities to a The worlds most functionality heterogeneous powerful enterprise Dynamic Increasing efficiencies Specialty computing platform Infrastructure® to Engines (IFL, Performance Enabling new solutions Support Business Critical Applications zAAP, zIIP) Capacity Superior Resilience Lowering the cost price/performance Responsiveness of deploying new workloads Expanded offload on zIIP LOW COST OF OWNERSHIP

IBM System z: Smart, cool, affordable

• System z10 designed to: – Reduce cost – Manage data and information – Deploy new applications and services – Manage security and compliance needs – Maintain availability and productivity – Reduce energy consumption

Unlike many distributed servers… … System z Delivers extreme business value through industry leading security, availability, scalability, virtualization and management capabilities

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation42 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 42 Thun GSE z/OS Expertforum Switzerland # 70, 31.3.-1.4.2009

Zürich | March 31, 2009 © 2009 IBM Corporation Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation43 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 43 Thun zHPF: High Performance FICON for System z z10 High Performance FICON for System z (zHPF) • FICON architecture for protocol simplification and efficiency, reducing the number of Information Units (IUs) processed. When exploited by the FICON channel, the z/OS operating system, and the control unit, FICON channel overhead is reduced and performance improved – Packages a series of commands in a single packet that looks like an SCSI CDB – Allows streaming of data for multiple commands in a single data transfer • The maximum number of I/Os per second is designed to be improved up to 100%* for small block sizes data transfers that can exploit the zHPF protocol • Requires control unit feature code for exploitation – Feature is ‘High Performance FICON’ – Requires a priced license feature (one time charge) and a monthly maintenance charge – IBM System Storage DS8000 family – Release 4.1(LMC level 5.4.1.xx (bundle version 64.1.x.x), or later.) • Applicable to all FICON Express4 and Express2 features CHPID type FC – Implemented in System z10 Licensed Internal Code – An FC channel can support multiple CUs using both older FICON protocol and zHPF protocol at the same time. Control unit capability is determined automatically – zHPF protocols will not be used unless all the CHPIDs and control unit ports for the device support the new protocols • Exclusive to System z10 • Supported by z/OS 1.8 and higher releases with PTF

*Note: Some complex channel programs can not be converted to zHPF protocol

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation44 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 44 Thun zHPF: High Performance FICON for System z z10 High Performance FICON for System z (zHPF) • Optimization of storage area network (SAN) traffic using zHPF to improve performance – Maximum number of I/Os per second can be increased by up to 100%* – For OLTP workloads (DB2, VSAM, PDSE, and zFS ) that transfer small blocks of fixed size data (4K blocks) • Exclusive to System z10 – FICON Express4 and FICON Express2 • Requires – Control unit exploitation – IBM DS8000 Release 4.1 – z/OS V1.7 with the IBM Lifecycle Extension for z/OS V1.7 (5637-A01), V1.8, V1.9, or V1.10 with PTFs

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation45 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 45 Thun zHPF: High Performance FICON for System z z10 High Performance FICON for System z (zHPF)

Link Protocol Comparison for a 4KB READ

FICON I/O Protocol Hierarchy

• If the I/O is zHPF eligible and to a device that supports zHPF, the I/O will be zHPF

• If the I/O is MIDAW eligible and to a device that supports MIDAW, the I/O will be MIDAW

• If the I/O is neither zHPF or MIDAW eligible, then the I/O will be standard FICON

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation47 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 47 Thun zHPF: High Performance FICON for System z zHPF Summary and Highlights • Maximum Application Benefit for Typical OLTP Workloads – Optimize DB2 Performance. Media Manager builds a new type of Channel Program • Transport Control Word (TCW) instead of Command Control Word (CCW) – Use simpler protocols to encapsulate channel programs while preserving the enterprise class qualities of service of FICON – Complex channel programs continue to use CCW Chains with base FICON protocols – Maximum I/O rate for a channel with a simple 4KB read hit benchmark doubles with zHPF – Realistic production workloads with a mix of data transfer sizes may see up to 30% savings in channel utilization (compared to FICON) – Sequential workloads that transfer up to a single track (for example, 12 x 4KB per I/O) may also benefit – OLTP Workloads that exploit zHPF could see up to 30% improvement in DS8000 throughput • Compatibility between existing CCWs and new TCWs – Bilingual Channel and Control Unit Ports – CCWs continue to use FICON protocols – TCWs use new Transport Mode Protocols

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation48 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 48 Thun zHPF: High Performance FICON for System z zHPF Summary and Highlights (cont.) • DS8000 Code Structure Optimized for Simple I/O Chains – No CKD operation (ECKD only) – Streamlined Internal Communication Protocols (equivalent to FCP Exchanges) • Improved RAS and Workload Management – Additional channel and control unit diagnostics for MIH conditions – I/Os are queued in control unit when a device is reserved by another host • zHPF is Strategic – Watch this Space!

art For more Information refer to: s P d a ner ide or ov 's C x Pr vio bo – zHPF for System z Technical Summary for Customer Planning, ZSW03058-USEN-00 Sil uke of c J Do

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation49 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 49 Thun System z10 Design, Testing and Verification

Interested to read more about the Design of System z10?

• Then have a look on the newest issue of the IBM Journal of Research and Development, Volume 53, Number 1, 2009

• The IBM System z10 EC server is a powerful mainframe system offering increased power efficiency, improved price/performance ratio, and expanded range of business solutions

• It delivers high scalability for growth and large-scale consolidation, high availability to reduce risk and improve flexibility, and enhanced security

• The 17 papers in this issue describe the design, testing, and verification of the z10 microprocessor, the I/O subsystem, and the packaging design of the central electronic complex

• System software and performance are described in papers dealing with reliability, availability, and serviceability, autonomic capability and active resource monitoring, capacity-on-demand advancements, and performance improvements derived through novel software and hardware synergy

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation50 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 50 Thun System z10 Design, Testing and Verification

Interested to read more about the Design of System z10?

• Detailed Topics of JoRD, Volume 53, Number 1, 2009

• To download these 17 Papers go to: • http://www.research.ibm.com/journal/rd53-1.html

System z10 CPU Measurement Facility What is the z10 CPU Measurement Facility? • New hardware instrumentation facility “CPU Measurement Facility” (CPU MF) – Available on System z10 EC GA2 and z10 BC – Supported by a new z/OS component (Instrumentation), Hardware Instrumentation Services (HIS) • Potential Future Uses –for this new “cool” Virtualization Technology – CPU MF provides support built into the processor hardware • So exploiting mechanism allows the observation of performance behavior with nearly no impact to the system being observed – Potential Uses • Future workload characterization • ISV product improvement • Application Tuning

System z10 CPU Measurement Facility… High Level Description • Data Collection done by System z Hardware – Low overhead – Little/No skew in sampling – Access to information which is not available from software • Information about how software and hardware interact • 2 Basic Modes – COUNTERS – SAMPLING • SAMPFREQ=800000 is the default (samples per minute), = 13,333 / sec – 8M samples in 10 minutes is the default » (DURATION=10 is the default, 10 minutes) – Recommendation – Start with a small frequency, e.g. SAMPFREQ=320, and increase after early experiences – e.g. ensure enough disk space for output » Smaller z10 BCs should increase only up to SAMPFREQ=130000 (for DURATION=60) • New IBM Research Article – “IBM System z10 performance improvements with software and hardware synergy” • http://www.research.ibm.com/journal/rd/531/jackson.pdf

System z10 CPU Measurement Facility… Requirements and Steps to utilize z10 CPU MF • Requirements for CPU MF – System z10 machine must be at GA2 Driver 76D – Bundle #11 or higher • Currently restricted use on the z10 BC sub-capacity models – z/OS LPAR being measured must be at z/OS 1.8 or higher with APARs: • OA25755, OA25750, and OA25773 • Not supported for z/OS running as a z/VM guest • Steps to utilize CPU MF 1) Configure the z10 Server to collect CPU MF Data – Update LPAR Security Tabs (See appendix) 2) Configure HIS on z/OS to collect CPU MF Data * - Set up HIS Proc - Set up OMVS Directory - Collect SMF 113s via SMFPRMxx 3) Collect CPU MF Data - Start HIS – Modify with Begin/End – for COUNTERS or SAMPLING 4) Analyze the CPU MF Data - SMF 113s or OMVS files (e.g. *.CNT, etc.) * See MVS Commands SA22-7627-19 “Setting up hardware event data collection” 1-39 Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation54 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 54 Thun System z10 CPU Measurement Facility

System z10 CPU Measurement Facility… How it works

System z10 CPU Measurement Facility… What data is in the CPU MF – per Logical CP • Basic Counters (and Problem) per CPU – (1) – Cycles – Instructions – L1 Cache Sourcing basic information • Crypto Counters per CPU – (1) – Counts and Cycles by Crypto function • Extended Counters – per CPU (Model Dependent) – (2) – Cache Hierarchy Information and more • z10 L1 Sourcing detailed information

System z10 CPU Measurement Facility… CPU MF and HIS provide a z/OS logical View of z10 Resource Usage and Cache Hierarchy Sourcing

System z10 CPU Measurement Facility… What we did • Set up CPU MF on WSC z10 and z/OS 1.10 • Started/Modified HIS and collected SMF 113s and *.CNT Data – Ran “COUNTERS” mode, COUNTERS=ALL (Basic, Problem, Crypto, Extended) via: – “F HIS,B,TT='EncrypCounters2',PATH='/his/',CTRONLY,CTR=ALL” • Ran DASD dumps – DASD dumps sequentially over 20 minute duration – With option: ENCRYPT(CLRTDES) - • Built sample reports with a REXX exec – Used *.CNT output to as input – Validated with SMF 113s – Reports • Basic Counters • Crypto Counters • Basic / Extended Counters - z10 L1 Cache Hierarchy Sourcing Report

System z10 CPU Measurement Facility… SYSHISyyyymmdd,hhmmss.CNT Output unformatted

System z10 CPU Measurement Facility… Sample Report – Basic Counters

System z10 CPU Measurement Facility… Sample Report – Crypto Counters

System z10 CPU Measurement Facility… Sample Report – Basic / Extended Counters z10 L1 Cache Hierarchy Sourcing

System z10 CPU Measurement Facility… CPU MF and HIS provide a z/OS logical view of z10 Resource Usage and Cache Hierarchy Sourcing

System z10 CPU Measurement Facility… z10 CPU Measurement Facility Summary • COUNTERS mode took very low overhead to run – Less than 1/100 of a second for HIS address space in 15 minute interval • CTRONLY option ensures that delta counters are provided in the *.CNT output – Useful for determining rates from the Start and End times • CPU MF provides unique z10 performance measurement capabilities – Future potential uses may include • Workload characterization, ISV product improvement, and application tuning

System z10 CPU Measurement Facility… Summary - Workload Characterization Future Vision • Future Vision is to potentially combine SMF 23s and SMF113s – To help identify workload characteristics and to provide better input for capacity planning and performance

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation65 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 65 Thun IBM DS8000 Storage Subsystem Update IBM DS8000 Storage Subsystem Update

Improve Service, reduce Cost, manage Risk

• IBM System Storage DS8000 R4.2 • z/OS • Full disk encryption for data protection – Tivoli Key Lifecycle Manager – Simplified, highly secure and cost-effective key (planned to be available for z/OS in storage, key serving and key management with March) is an unpriced product that Tivoli® Key Lifecycle Manager leverages z/OS security, • Solid state drive (SSD) option for high management, and reporting priority, time-sensitive applications capabilities (W/ V1.9) – Increased performance for some transactional applications – Define new z/OS SMS policies for – Faster data replication and recovery from the allocation of new data sets on outages volumes backed by SSD – Absence of mechanical moving parts makes technology and to gather usage SSDs significantly more reliable information using SMF that is – Fraction of the energy consumed, fraction of intended to help you manage data heat dissipated placement to take the best advantage of this new feature ® • FlashCopy and Metro Mirror for more (w/V1.9) effective two-site business continuity – Helps improve data synchronization and – Function available with z/OS V1.8 availability

IBM DS8000 Storage Subsystem Update… Enhancing Security • Drive-Level Encryption – New DS8000 feature supports encryption of “data at rest” – Continuous, real-time encryption of the individual drives – Expected to: • Have no performance impact • Require no application changes – Uses Tivoli Key Lifecycle Manager (5608-A91) • Key management via ICSF and RACF® • Auditability via SMF • TKLM availability planned for March 2009 – Supported on z/OS R9 R10 with the PTF for APAR OA27393

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation67 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 67 Thun IBM DS8000 Storage Subsystem Update IBM DS8000 Storage Subsystem Update

Scalability and Performance • Solid-State Devices for DASD – Flash-based “drives” • RAID-based • Dynamic chip sparing – Improved DASD response times – Caching with controller-based prefetching means SSD probably best suited for: • Infrequently written data • Frequently read data • Random access data • Data with high read disconnect times – HDD probably a better choice for: • Sequential access • Frequently rewritten data – SMF records, DATACLAS support to help with data management – Support available on z/OS R9 and R10 with APAR OA25559, planned to be included in z/OS R11 – Power consumption and cooling requirements markedly lower than for hard disk-based volumes art s P – For more information on DS8000 and SDD drive performance refer to: d a ner ide or rov 's C x P ilvio ebo • IBM System z® and System Storage DS8000: Accelerating the SAP® Deposits Management Workload with f S Juk o oc Solid State Drives, WP101442 D

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation68 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 68 Thun GSE z/OS Expertforum Switzerland # 70, 31.3.-1.4.2009

Zürich | March 31, 2009 © 2009 IBM Corporation Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation69 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 69 Thun z/OS Key Dates z/OS Key Dates

September 12, 2008: First date for ordering z/OS V1.10 ServerPac, SystemPac, CBPDO using CFSW configuration support or ShopzSeries, the Internet ordering tool Note that most z/OS media (executable code) is shipped only through Customized Offerings (ServerPac, SystemPac, and CBPDO) September 26, 2008: z/OS V1.10 general availability via ServerPac, CBPDO, and SystemPac October 14, 2008: Recommended last date for submitting z/OS V1.9 orders via the entitled Customized Offerings (ServerPac and CBPDO). This date will allow for adequate order processing time. October 27, 2008: Last date for processing orders for z/OS V1.9 via ServerPac and CBPDO. November 21, 2008: general availability of Cryptographic Support for z/OS V1R8-R10 & z/OS.e V1R8 Web deliverable. This Web deliverable will support z/OS V1.8 through z/OS V1.10 and z/OS.e V1.8. June 29, 2009: Recommended last date for submitting z/OS V1.9 orders via the fee Customized Offering SystemPac. This date will allow for adequate order processing time. July 27, 2009: Last date for processing orders for z/OS V1.9 via SystemPac. September 30, 2009: End of service for z/OS V1.8 (5694-A01) and z/OS.e V1.8 (5655-G52). For z/OS.e (5655-G52), z/OS.e V1.8 is the last release of the z/OS.e product. Refer to following recommendations for placing last orders.

* Statements regarding IBM future direction and intent are subject to change or withdrawal, and represents goals and objectives only.

z/OS Platform latest RSU Latest recommended Service Upgrade (RSU)

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation72 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 72 Thun IBM System z Innovations for a dynamic Infrastructure

IBM System z Innovations for a dynamic Infrastructure

• … IBM System z® delivers extreme Business Value through Industry leading Security, Availability, Scalability, Virtualization and Management Capabilities

• Announced for February 9, 2009

– Dynamic Infrastructure® initiative

– IBM System Storage™ DS8000™ Turbo

– Geographically Dispersed Parallel Sysplex™ GDPS® V3.6

– z/OS® V1.11 Preview

– IBM System z Solution Edition for ACI Applications

– Beta program for IBM InfoSphere™ Warehouse on System z

– New System z Trends and Directions

Statements of Direction* (February 2009)

• IBM intends to expand support for EAV with larger volume sizes and to allow additional data set types to reside in the cylinders after the first 65,520 cylinders

• IBM intends to update z/OS with a new function designed to generate messages for Server Time Protocol (STP) - related hardware events

• IBM intends to introduce the z/OS Management Facility, a separate product which will be designed to enable system programmers to more easily manage and administer a mainframe system by simplifying day to day operations and administration of a z/OS system

– The initial release is planned to provide a problem data management capability which is intended to facilitate problem data management tasks for new or less skilled system programmers and system administrators

• In a future release of z/OS, the BIND 9.2.0 function will be removed from the z/OS Communications Server component

* Statements regarding IBM future direction and intent are subject to change or withdrawal, and represents goals and objectives only.

z/OS 1.11 Enhancements Preview Previewed: Feb. 24, 2009, planned GA: Sept. 30, 2009 With Designs for a dynamic Infrastructure Trusted Accountable • Enhanced policy-driven network security functions for • Log, track, and audit distributed identities with RACF- consistent, system-wide enforcement of security controlled user IDs while maintaining the users' original policies identity information for audit purposes • New Query Algorithms helps your applications be • WebSphere Application Server V7 introduces an audit smart about what encryption implementation to use infrastructure designed to enable audit records to be • Validation services that can help determine if created for security-related events applications and other program downloads are un- • Tivoli Key Lifecycle Manager (TKLM), 5698-B35, will tampered provide SMF audit records on z/OS • Resilient, cost effective full certificate life cycle management with enhanced key recovery capabilities Smart Responsive • Ability for z/OS system to heuristically learn from its own environment and anticipate and report on potential system • Autonomic monitoring, management, and deployment issues (however rare) before they impact business of On/Off Capacity on Demand (portions of workload eligible for the zIIP processor) • Intelligent load balancing decisions made for multi-tier applications • Performance enhancements for C/C++ and Web- based applications • Heuristics to automatically generate System z10 prefetch instructions for C/C++ applications • Ability to manage fewer, larger storage volumes, as opposed to many small volumes * Statements regarding IBM future direction and intent are subject to change or withdrawal, and represents goals and objectives only.

z/OS 1.11 Enhancements Preview… Scalability and Performance Enhancements • Improvement in storage response times – DFSMS™ support planned for DS8000 R4.2 solid state drives (SSD, also called flash memory) – New SMS policies to gather usage information using SMF that is intended to help manage data placement to take the best advantage the new SSDs • Performance improvements for XL C/C++ applications on System z10 servers – New prefetch capability can heuristically generate System z10 prefetch instructions as appropriate • Reduced memory management with large (1MB) page support – Support for AMODE 64 XL C/C++ Language Environment applications, in addition to current exploitation by the 64-bit SDK for z/OS, Java® Technology Edition, V6 • Performance improvements for large systems with many zIIPs – Faster processors can actually spend more time waiting for memory access! HiperDispatch helps improve cache management and overall system performance – HiperDispatch algorithms to be updated for zIIP processors • Increase the efficiency of batch windows – Use IEFBR14 to delete catalogue reference to unneeded data sets and avoids the lengthy process of recalling the DS just to delete it • Virtual Storage Constraint Relief – Removes constraints within the base z/OS operation system and can allow more work to be processed on a single z/OS system * Statements regarding IBM future direction and intent are subject to change or withdrawal, and represents goals and objectives only.

z/OS 1.11 Enhancements Preview… Enhancements in Networking Performance • Improved throughput in support of disaster recovery or global operations – Dynamic tuning of TCP window for bulk transfers over high-latency, long distance networks • More performance for Web-based applications – System-wide caching of domain name server (DNS) responses • Applications with frequent resolver queries can benefit – Improved Fast Cache Accelerator function • Intelligent sysplex networking – The Sysplex Distributor plans to take into account the capacity, performance and health characteristics of both the tier 1 and the tier 2 z/OS server applications – This new function is intended to improve the quality of the load balancing decisions made by Sysplex Distributor in a multi-tier z/OS server environment • Many other performance improvements – New TCP/IP resolver improvements, Sysplex Distributor routing accelerators and WLM algorithms, socket error detection, QDIO accelerator function, Enterprise Extender and SMB improvements.

* Statements regarding IBM future direction and intent are subject to change or withdrawal, and represents goals and objectives only.

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation77 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 77 Thun z/OS 1.11 Enhancements Preview z/OS 1.11 Enhancements Preview… Networking Security Enhancements • z/OS Communications Server designs for network security enhancements: – z/OS System SSL • Addressing requirements for NIST FIPS 140-2 Level 1 criteria • TLS V1.1 protocol (RFC4346) • TLS Extensions (RFC3546) • Certificate validation at the RFC3280 level with compatibility with RFC2459 Apply TLS security services – Application Transparent – Transport Layer Security (AT-TLS) Apply TLS security services transparentlytransparently to to applications applications • Adding TLS V1.1 protocol support withwith AT-TLS. AT-TLS. No No need need for for • Can now validate certifcates as specified in RFC3280 programprogram updates updates because because TLS TLS • Supports negotiation and use of a truncated HMAC servicesservices are are applied applied in in the the TCP TCP (RFC3546), maximum SSL fragment size (RFC3546), layerlayer of of the the z/OS z/OS Communications Server handshake server name indication, and setting the Communications Server CRL LDAP server access security level BothBoth AT-TLS AT-TLS and and IPSec IPSec can can • Application access to System SSL design changes related to FIPS 140-2 help reduce application level 1 criteria help reduce application enablementenablement requirements requirements for for • Provides improved performance for short-lived connection workload networknetwork security. security. These These functions are policy driven for – IPSec functions are policy driven for consistent,consistent, system-wide system-wide • IPSec performance improvements for Enterprise Extender traffic enforcementenforcement of of security security • Improved Internet Key Exchange reliability policies.policies. • Improved granularity in IPSec management information * Statements regarding IBM future direction and intent are subject to change or withdrawal, and represents goals and objectives only.

z/OS 1.11 Enhancements Preview… Security Enhancements • ICSF – many Encryption Solutions in one – Methodologies to help protect and manage keys, with a single point of control for all z/OS key- management processes – It takes advantage of the mainframe's high availability (including Parallel Sysplex® data sharing across multiple servers, and GDPS for disaster recovery) – Users can integrate key management into the many existing mainframe management processes – Integration with RACF means access to the key store and administration functions can be controlled and audit records are available to help address compliance requirements

• ICSF Designs – New services for major credit card vendors, such as VISA and MasterCard, to generate and verify the verification values -- the authenticity of the cards – A new Key Store Policy is to provide a set of policy controls designed to specify further limits on application control of key material and provide a central point of control – New ICSF Query Algorithms helps middleware determine what cryptographic implementations to use – helps your code be smart about when implementation to use (hardware vs software, etc) – ICSF will support the AES encryption algorithm with secure (encrypted) keys of 128, 192, and 256 bits

* Statements regarding IBM future direction and intent are subject to change or withdrawal, and represents goals and objectives only.

z/OS 1.11 Enhancements Preview… z/OS Availability Enhancements • z/OS V1.11 plans to extend predictive Failure Analysis – z/OS system heuristically learns from its own environment and is able to anticipate and report on potential system issues (however rare) before they are an impact to your business • z/OS UNIX® System Services with System Call (Syscall) Trace – intended to gather more information about program processing history to facilitate application debugging • New Allocation Commands – can help improve system availability by allowing you to change Allocation settings without an IPL • New Latch Identity Service for improved Latch Contention • Improved Serviceability, including IPL Restart Improvements and improved Dump Management • Parallel Sysplex:

• Networking (Sysplex Distributor) • Availability – New WLM routing algorithms for better zIIP and zAAP – New health checks for DAE and STP workload routing – Alternate Sysplex root file system support – Connection routing accelerator for performance – Enhancement to XCF and XEC – Intelligent routing for multitier z/OS applications – Auto IPL (R10)

* Statements regarding IBM future direction and intent are subject to change or withdrawal, and represents goals and objectives only.

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation80 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 80 Thun z/OS 1.11 Enhancements Preview z/OS 1.11 Enhancements Preview… z/OS Simplifying Operations and Programming • A z/OS Management Facility (Statement of Direction)* – More easily manage system – Initial release to facilitate problem data management • IBM Health Checker for z/OS – New health checks for: • Auto IPL best practices and device validation • DFSMS to detect IMBED and REPLICATE • Static resource manager • Dump Analysis and Elimination • SDSF using SAF – New migration checks for: • IPSec filter rules, BIND9 DNS usage, DFSMSrmm, STP/ ETR, Message Flood Automation • Advanced Communications Facility Trace Analysis Program (ACF/TAP) is planned to be made a part of z/OS Communications Server element (no need for use the Advanced Communications Facility Network Control Program (ACF/NCP)) • Faster and easier Report Generation for DFSMSrmm and RMF • Lots of ISPF Updates • Lots of DFSMSrmm Updates

* Statements regarding IBM future direction and intent are subject to change or withdrawal, and represents goals and objectives only.

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation81 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 81 Thun z/OS 1.11 Enhancements Preview z/OS 1.11 Enhancements Preview… Application Integration • C/C++ Applications – Continued adoption of language standards for skill commonality – Improved application portability – Performance improvements – Improved debugging capabilities provide additional productivity • System Applications – METALC improvements – easier integration of Assembler code and C programs – SYSREXX™ improvements • Helps enable rapid development & deployment of system programmer tools and operator assists • Easier to write Health Checks in REXX • Decimal Floating Point Applications – Additional DFP library function support in in XL C/C+ • Global Application Resources – C/C++ Unicode enhancements – Additional codes page support in LE – Unicode System services enhancements

* Statements regarding IBM future direction and intent are subject to change or withdrawal, and represents goals and objectives only.

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation82 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 82 Thun z/OS 1.11 Enhancements Preview z/OS 1.11 Enhancements Preview… Optimization and Management • Sysplex Distributor support for WLM enhanced workload routing – Considers displaceable capacity, considers zIIP and zAAP capacity • z/OS Workload Manager: – Improved routing algorithms – More granularity in reporting classes (2,048 Reporting Classes in WLM!) • Optimized LDAP (IBM Tivoli Directory Server) – Cross-system load balancing using WLM – “Storm drain” avoidance though ITDS’s use of the WLM server health service • New BCP internal interface (BCPii) component – API for authorized programs to perform HMC functions • List CPCs, Images, Capacity records, and query related attributes • Change profile contents • Activate, deactivate, start, stop, reset, restart systems • initiate capacity changes (up and down) – Can communicate using internal HMC network, isolated from other network traffic – z/OS Capacity Provisioning - BCPii exploitation and logical processor management • Improved Allocation tape load balancing algorithm • Lots of DFSMSrmm™ and DFSMShsm™

* Statements regarding IBM future direction and intent are subject to change or withdrawal, and represents goals and objectives only.

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation83 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 83 Thun z/OS 1.11 Enhancements Preview z/OS 1.11 Enhancements Preview… What is CIM? • The Common Information Model (CIM) is an industry standard for defining and sharing information between devices and programs • Can be used for developing common cross-platform and cross-system monitoring tools and management applications for diverse resources • A standard framework is used so CIM client applications can request CIM servers to return information about the resources – Distributed Management Task Force (DMTF) www.dmtf.org – Web Based Enterprise Management (WBEM) initiative www.dmtf.org/standards/wbem

* Statements regarding IBM future direction and intent are subject to change or withdrawal, and represents goals and objectives only.

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation84 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 84 Thun z/OS 1.11 Enhancements Preview z/OS 1.11 Enhancements Preview… Application CIM on z/OS CIM Client • CIM is used within z/OS to communicate information on CIM XML Processor or to manage resources for system components CIM HTTP Client – The CIM client on z/OS is a programming API that

enables z/OS applications written in Java for local and CIM clients send CIM server responds requests to CIM with data to CIM remote access to CIM servers server client • Java classes and Java libraries • Java-based CIM client applications on z/OS are already CIM HTTP Server eligible to execute on the zAAP CIM XML Processor – Communication is via CIM-XML over HTTP access CIM Server protocol

RMF monitoring z/OS OS management • IBM, ISV, or custom written applications can access providers providers CIM-enabled resources, can monitor or manage z/OS RMF Distributed data resources Server (DDS)

– z/OS Common Information Model User's Guide RMF Monitor III native z/OS data gathers and returns metrics to the DDS • ibm.com/systems/z/os/zos/bkserv/r9pdf/#cim Managed System

Example with z/OS system data and RMF metrics

* Statements regarding IBM future direction and intent are subject to change or withdrawal, and represents goals and objectives only.

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation85 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 85 Thun z/OS 1.11 Enhancements Preview z/OS 1.11 Enhancements Preview… CIM Server Workload eligible for zIIP*, planned Application Java- • z/OS V1.11 is planned to be updated so z/OS CIM Server CIM Client based Processing is eligible to run on the System (zIIP)* CIM XML Processor CIM Client – Eligible CIM server workloads include CIM provider CIM HTTP Client eligible workloads as well for zAAP CIM clients CIM server – Other CIM-related workloads (such as CIM client and send requests responds with data to CIM client CIM-enabled resource systems processing) are not to CIM server eligible for zIIP CIM HTTP Server • Makes the Development and Deployment of z/OS CIM CIM XML Processor Systems Management Applications more attractive Server eligible CIM Server Option for zIIP

• Applications that access CIM-enabled Resources and RMF monitoring z/OS OS providers management Providers can benefit providers RMF Distributed – Information providers include RMF, WLM, DFSMS and data Server (DDS) RMF Monitor III native z/OS data BCP gathers and returns metrics to the DDS – Systems management exploiters include z/OS Capacity Provisioning Manager and z/OS Management Facility* An example with z/OS system data and RMF metrics (planned, statement of direction)

* Statements regarding IBM future direction and intent are subject to change or withdrawal, and represents goals and objectives only.

System z DS8000 Petabyte useable Capacity

• Objective: install/configure 1 PB of disk storage accessible and useable from a single z/OS image or Sysplex

• These are nonreplicated base volumes, any requirements for replication will require additional storage

• Reality is: you can do this today

• 1 Petabyte = 1000 TB

• ~1GB = 1 DS8000 extent • Use 1,000,000 extents as 1 PB for calculation simplicity

• Considerations for Planning

• Hardware / Useable Capacity • 3390 Logical Volume Size selection • Addressing

DS8000 1TB SATA Preview Announcement on 8/12/2008

• Contained in IBM Announcement Letter 108-327

• Serial ATA (SATA) disk drives: IBM intends to enhance the configuration options of the DS8000series with support for:

• 1 TB 7,200 RPM SATA disk drives, doubling the raw capacity of existing models up to 1024TB for DS8300 Turbo Models and up to 384 TB for DS8100 Turbo Models

• This capability is planned to be available on March 6, 2009

– See announcement letter ENUS109-119 of February 10, 2009

System z DS8000 Petabyte useable Capacity (cont.)

• Useable Capacity (CKD)

• Configurations available today will probably need 2-8 DS8000s to provide capacity • Larger disk capacity features in the future will mean smaller storage configurations • Most popular DS8000 disk feature today is 146GB/15K

– Places Useable Capacity at ~125,000 extents for a 5 frame DS8000 (125,664) in Raid5

– Means we need 8 DS8000s to get 1 PB useable

• Largest disk feature will be the 1TB SATA drives – Places Useable Capacity at ~650,000 extents for a 5 frame DS8000 (657,248) in Raid6 – Means we need 2 DS8000s (1.538 actually) for 1 PB useable

DS8000 3390 Volume Sizes (including EAV)

• GUI includes 3390 standard mod 3 and 3390 standard mod 9 Definitions

• GUI/CLI also allow for definition of a 3390 custom volume

• Any number of 3390 cylinders are supported (1-65520) • Best to align to Extent Boundary of 3390-1 (Multiple of 1113)

• 3390-A is an EAV 3390

• Cylinder support increases to 262,688 – Forced alignment to 1113 cylinder multiple when number of cylinders is above 65667

• Use of a multiple of 3390-9 would be recommended

DS8000 3390 Volume Sizes (including EAV)…

• Advantages of Standard 3390 Volume Sizes: • DS8000 copy services to work without needing to locate correct sized volume pairs • Simplified Volume Migration • Simplified Storage Estimation • Creation of larger volume sizes as a multiple

• 3390-27 would have 30,051 cylinders (25.5GB) • 27 extents, Not 32760 cylinders • 3390-54 would have 60,102 cylinders (51GB) • 54 extents, Not 65520 cylinders • 3390-A would have 240,408 cylinders (204GB) • 216 DS8000 extents, not 262688 cylinders • Same as (24) 3390-9s, or (8) 3390-27s or (4) 3390-54s

System z DS8000 Petabyte useable Capacity (cont.)

• Addressing • FICON allows up to 256 Control Units per channel with 256 devices – 16K total devices per FICON channel • Up to 8 path channel groups per CNTLUNIT – Can increase the number of devices/Channel Group with zHPF • OSs support up to 256 FICON channels/LPAR • 65,536 Device limit per Channel SubSet (CSS) – Limit DISK to 20,000 base addresses for replication technologies • Assume 200 base addresses per LCU, leaves 56 addresses for aliases – Static/Dynamic PAV probably not responsive enough, need HyperPAV • With these requirements/assumptions – 1,000,000 extents / 20000 devices = 50 extents per device (average) 1,000,000 / 54 = 18519 devices (3390-54) 18519/200 = 93 LCUs (across 2-8 DS8000s) 1,000,000 / 216 = 4630 devices (3390 EAV) 4630 / 200 = 24 LCUs (across 2-8 DS8000s) – Larger volumes will want to take advantage of DS8000 Multi-Rank Extent Pools and Storage Pool Striping to create logical volumes

System z DS8000 Petabyte useable Capacity Conclusion

• You now know the answer… we could stop here – but we’ve made some unqualified statements that we want to defend

• Specifically: • Extended Address Volumes are a viable method to enablement • Static/Dynamic PAV probably not responsive enough, need HyperPAV • Larger volumes will want to take advantage of DS8000 Multi-Rank Extent Pools and Storage Pool Striping to create logical volumes • Increase the number of IODEVICEs per Channel Group with zHPF

Overview – What is it

EAV: A volume with more than 65,520 cylinders Size limitedSupport to 223 GBin z/OS (262,668 V1R10 Max cylinders) The HyperPAV function complements this design by scaling the I/O rates against a single volume

3390 Model A: a device configured to have

3390-A 1 to 268,434,453DS8000 cylinders “EAV”

3390-9 3390-9 3390-3 3390-9 100s of TBs 3 GB 9 GB 27 GB 54 GB Max cyls: 3,339 Max cyls: 10,017 Max cyls: 32,760 Max cyls: 65,520 Maximum Sizes

Overview – EAV Key Design Points

• Maintains 3390 track format • Track-managed space: the area on an EAV located within the first 65,520 cylinders – Space is allocated in track or cylinder increments cylinder-managed – Storage for “small” data sets space • Cylinder-managed space: the area on an EAV Cylinders located above the first 65,520 cylinders > 65,520 – Space is allocated in multicylinder units (MCU) • A fixed unit of disk space that is larger than a cylinder. Currently, on an EAV, the MCU is 21 cylinders Cylinders track-managed • System may round space requests up <= 65,520 space – Storage for “large” data sets • Track-managed space comparable to same space on non-EAVs EAV

Overview – EAV Key Design Points

• New track address format needed • Base addressing space: the area on an EAV located within the first 65,536 cylinders – 16-bit cylinder addressing (CCHH) Extended addressing – How disks are addressed today space – Comparable to same addressing space on a non- Cylinders EAV > 65,535 • Extended addressing space (EAS): the area on an EAV located above the first 65,536 cylinders – 28-bit cylinder addressing (CCCCcccH) Cylinders Base addressing <= 65,535 – VSAM data sets are EAS eligible space • New DSCB (data set control block) types – Provides a method of protecting existing programs EAV from seeing unexpected track addresses

Data Set Placement Considerations

• An EAV can be either SMS managed or non-SMS managed • An EAV can be indexed or non-indexed – IBM recommends indexing all volumes • VSAM and Non-VSAM data set types may reside in the track-managed space • In the first release of EAV support, it is planned that: – VSAM data sets can be created and extended anywhere on the volume (cylinder or track-managed space) • Preference of cylinder vs. track-managed space allocations typically based on ‘Breakpoint Value’ – Non-VSAM data sets can be created and extended ONLY in the track-managed space • USEEAV(YES|NO) – PARMLIB option (IGDSMSxx member). With USEEAV(NO), no additional creates of new data sets will be allowed on any EAV. This is the default. – May be updated via SETSMS operator command

DS8000 and z/OS 1.10 EAV Features

• Requires z/OS 1.10 • Requires R4 DS8000 Licensed Machine Code (LMC) • Supports 3390 formatted volumes (no 3380 format) • ~4x capacity of previous maximum volume capacity • Future plans include – Larger volumes – More data types – More access methods • Complements HyperPAV

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation98 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 98 Thun Extended Address Volumes (EAV) z/OS 1.10 EAV Layout

How do we accomplish the increase in size?

Overview – old Track Address

• Existing track address format with 16-bit cylinder number CCCCHHHH 16-bit track number 16-bit cylinder number

– Today's supported maximum size volume is 65,520 cylinders, near the 16-bit theoretical limit of 65535

• To handle cylinder numbers greater than 65,520, a new format for the track address is required

Overview – new Track Address

• Reformat CCHHR

– Supported Device types use only 15 tracks / Cylinder – Allows 12 more bits for cylinder addressing – CCCCcccHR – CCHHR address changes from 16 bits CC and 16 bits HH to 28 bits CC and 4 bits HH (actually it’s only an H now)

• Addresses non-EAV volumes the same (pre/post z/OS 1.10), zeroes in the high order track address

• By adding the available portion of the addressing for cylinders, 16+12 bits, 2^28 -1102 = 268,434,453 cylinders possible (268435455 - 1002). 1002 cylinders not available due to DS8000 extent alignment

• By adding the available portion of the addressing for cylinders, 16+12 bits, 2^28 -1102 = 268,434,453 cylinders possible (268435455 - 1002)

EAV Volumes cannot come online to pre z/OS 1.10 systems

Additional Cylinders - Eligible Data Sets

• Cylinders Above 65,520: a dataset on an EAV that is eligible to have extents in the additional cylinders (a.k.a cylinder-managed space) – In EAV Release 1, restricted to VSAM datasets with system selected CA Size • KSDS, RRDS, ESDS, and Linear (with system selected CA Size of 1, 3, 5, 7, 9, or 15 tracks)

– This covers DB2, zFS, CICS VSAM, RLS, IMS VSAM, etc. • The following are NOT Eligible for placement in the additonal cylinders in the first release: – Catalogs (BCS and VVDS) – VTOC (continues to be restricted to within first 64K-1 tracks) –VTOC index – Page datasets – VSAM datasets with imbed or keyrange attributes or non-system selected CA sizes – Non-VSAM datasets

In a future release, some of the above datasets may become eligible

Volume Candidates for EAV

• Good volume candidates contain DB2, zFS, CICS VSAM, RLS, IMS VSAM • Poor EAV Candidates (for first release, these can be located on an EAV but not placed in the Cylinders above 65,520)

– DFSMS/hsm ML1, Backup, or ML2 • Only SDSP eligible for Additional Cylinders above 65,520 • Nothing wrong with HSM CDS volumes… – Work Volumes (Public Space, Temp Space) – TSO/GDG/Batch/Load Libraries (largely Sequential or PDS DSNs) – System Volumes •SYSRES •Paging Volumes • Master Catalog • Distribution Libraries –Anything Else?

EAV Device Definition

• IOCP/HCD the same as all other 3390’s attached to 2107/242x – Supported Device types use only 15 tracks / Cylinder • Can start with new definition in DS8000 or expand existing • Option to take an existing definition and expand to EAV – Using CLI or DS8000 GUI – Change the device from 3390 to a 3390 model A – Change the cylinders to some number above 65,520 – Use ICKDSF Reformat RefreshVTOC to realize additional space • Consideration: does not expand VTOC – Does re-allocate VTOC index if previously allocated on volume • Does not allow Storage Pool Striping unless volume created as –EAM ROTATEEXTS • RMF, DEVSERV will report EAV as a 3390 Model A device – Not to be confused with an alias – Similar to 3390 model 9 for all volumes with more than 3339 cyls

EAV Implementation Scenario

• Expect all systems in a SYSPLEX migrated first to z/OS 1.10 • DS8000 has EAV code support • Once these first pre-requisites are met…

• Then implement EAV

– Either by defining new volume(s) – Or expanding existing ones – Or both

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation105 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 105 Thun GSE z/OS Expertforum Switzerland # 70, 31.3.-1.4.2009

Zürich | March 31, 2009 © 2009 IBM Corporation Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation106 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 106 Thun APAR's of Interest

APAR's and Red Alerts

• OA23588 (Hiper) Move of Catalog not reflected on second Systems with shared MCAT • OA25278 (Hiper) Sysplex wide Contention may occur after a Rebuild of Catalog ECS Structure SYSIGGCAS_CAS • OA25868 (Hiper) CNZ4201E Followed by VARY OPERLOG,HARDCPY results in hung System • OA26693 (Hiper) High CSM 32K ECSA Buffer Usage causing Storage Shortage (z/OS 1.10) • OA26841 (Hiper) ABEND0C4-10 after a Structure was rebuilt to another CF as Result of Maintmode / Reallocate

• OA26843 (Hiper) IFA812I IFASMFDL Request incorrout for Logstream.Name possible bad Parameter BLOCKLEN RC08 RSN0804 • OA27013 (Hiper) Duplexed Structure out of Sync when Link returns busy instead of Status pending • OA27032 Output from IWMWSYSQ returns more SU than CEC Capacity • OA27129 (Hiper) IOS500I can not delete Devices Reason=016B Reason 016B System Error – Diagnostic Info 4015 010E 03 • OA27147 (Hiper) Unplanned HyperSwap after CONFIG CHP Command • OA27221 (Hiper) ABEND0C4 in CPOOL Module IGVCPOOL when using SMF Logger

• OA27250 (Hiper) *IRA101E Critical SQA Shortage SQA filled with SSRB and LSSD • OA27291 (Hiper) ABEND0C4,various other ABENDs or Overlays in z/OS V1R10 when GEMAINed Storage is not cleared • OA27293 (Hiper) Operations Log Failure during IPL reported with Messages IEE316I and CNZ4201E • OA27303 ABEND0C4 in IXCT1PCP possibly followed by Spin Loop in IN IXCT1QV ORIXCT1PVP ABEND071 • OA27320 (Hiper) System Wait State 0A3 reason 088 while running in GRS Star Mode after an IXLLOCK RC=10 RSN=0C2D0001

• OA27401 Logrec IXGCONN IXG231I RC8 RSN802 IFB100E

APAR's and Red Alerts...

• OA27455 (Hiper) ABEND5C4 RSN06020007 terminating RRS after receiving an ABEND0C4PIC4 at ATRCMREC +1490 • OA27463 (Hiper) HSM CDS Backup May Fail When Us-ing DSS and Concurrent Copy is not Available Although it Returns with CC=0 • OA27562 (Hiper) Spin Loop in IAXMM in Restart Subroutine, ABEND071 due to improper Mapping Macro CSCOPYPFTE • OA27571 ABEND0C9 in ERB3XPH1 when invoking BPXEKDA to gather Thread Information • OA27685 (Hiper) ABEND9C6 RSN02E BPXINPVT BPXQRIN5 BPXMIPCE • OA27707 Deleted VVR Exposure for ECS Catalog full Volume Backup Restore • OA27869 IRA863E HiperDispatch Mode Algorithm detected an Error RC-002F • OA28005 (Hiper) ABEND0C5 in IAXYD during Dump Processing • OA28441 (Hiper) IRA401E Critical pageable Storage Shortage when using Scroll Hiperspaces and >128Gb Central Storage

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation108 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 108 Thun GSE z/OS Expertforum Switzerland # 70, 31.3.-1.4.2009

Zürich | March 31, 2009 © 2009 IBM Corporation Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation109 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 109 Thun Possible Differences between old/new Duplex Structure Sizes

Possible Differences between old/new Duplex Structure Sizes • When using System-managed CF Structure Duplexing, you may observe some Differences between the Sizes of old and new Structure Instances

• You may see these Differences even when the Coupling Facilities hosting the old/new Duplex Structures have the same Configuration • CF Structure Sizes may increase somewhat when Structures are rebuilt or duplexed, even when the two Structure instances involved are in the exact same CF, or another CF at the same CFLEVEL and Extension of with the same CF Storage Increment etc. capabilities* • This BehaviorToday’s has not been documented yet, but has been addressedCapabilities with Doc APAR OA22860

• Based on APAR OA22860, a new Section describing the Potential for CF Structure Size Growth during a Rebuild Process will be added to "Setting Up A Sysplex" (SA22-7625)

• The new Section will be added to Chapter 4 "Managing Coupling Facility Resources“ immediately following the section on "Structure Allocation when Rebuilding a Structure"

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation110 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 110 Thun Possible Differences between old/new Duplex Structure Sizes

Possible Differences between old/new Duplex Structure Sizes… • Documentation Update provided by APAR OA22860:

Warning About CF Structure Size Growth During Structure Rebuild/Duplexing, and Resulting Potential for Insufficient CF White Space Storage Capacity for Recovery

In order to accommodate possible CF structure size growth when moving CF structures between CFs, z/OS generally allocates an original structure instance "by size and ratio“ in accordance with the structure size specified in the active CFRM policy and the object ratios specified by the structure's exploiter

z/OS then generally allocates a rebuild-new or duplex-secondaryExtension structure of "by counts" so as to try to get the new structure instance to have the exact same object counts that the original structure instance has capabilities* This is doneToday’s to ensure that the new structure instance has a sufficient number of objects (e.g entries, elements)Capabilities to be usable by the exploiter as a rebuild-new or duplex-secondary structure, and in particular, to ensure that the structure is adequately sized and has sufficient objects even when the CFLEVELs of the two CFs involved are different The size of the new structure is allowed to "float" to whatever size is needed to accommodate the required object counts, to ensure that the rebuild or duplexing operation can succeed This avoids a variety of difficult problems that could otherwise occur if the new structure were to be allocated "by size and ratio" and (for example, due to CF storage allocation algorithm changes between CFLEVELs) the new structure got allocated with fewer usable objects, despite being allocated with the same absolute size as the original structure

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation111 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 111 Thun Possible Differences between old/new Duplex Structure Sizes

Possible Differences between old/new Duplex Structure Sizes… • Documentation Update provided by APAR OA22860 (cont.):

While customers are warned to expect structure size growth when going from one CFLEVEL to another, and that they must plan for it by providing sufficient CF storage in the new CF to accommodate such growth (and also by re-sizing their structures in the CFRM policy), the following effect is more subtle and must also be planned for Sometimes, even when rebuilding or duplexing between two CFs at the SAME CFLEVEL, the rebuild- new or duplex-secondary structure will be allocated somewhat larger than the original structure, EVEN THOUGH it has identical object counts to the original structure The reason for the size discrepancy has to do with difficultiesExtension in ascertaining of and reproducing the exact structure geometry (including that of various internalcapabilities* structure controls) of the original structure so that it can be replicated in the new structure, even though the object counts of the original structure areToday’s known with certainty and can be reproduced exactly Capabilities Sometimes, unavoidably, more internal structure controls are allocated for the new structure than were actually present in the original structure, resulting in a larger structure size This growth does not always occur for every structure, and when it occurs, it typically is on the order of one or two CFCC storage increments (the absolute size of a CFCC storage increment is displayed in the output of the DISPLAY CF command and is documented on a per-CFLEVEL basis in the PR/SM Planning Guide)

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation112 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 112 Thun Possible Differences between old/new Duplex Structure Sizes

Possible Differences between old/new Duplex Structure Sizes… • Documentation Update provided by APAR OA22860 (cont.):

Note that this growth can be exacerbated in cases where the structure is initially allocated with a given ratio (and its structure controls are allocated based on that ratio), and then the original structure's ratio is significantly modified (via Alter and/or AutoAlter processing) prior to the structure getting rebuilt/duplexed In such a case, the allocation of structure controls in the new structure, which is based on a ratio that is very different from that which was used in allocating the original structure, will potentially use more storage for controls than the original structure used However, it is still possible for some structure size growthExtension to occur even of in structures for which no Alter or AutoAlter processing has ever occurred (and for whichcapabilities* such processing may not even be supported) Today’s While no CFRMCapabilities policy size changes need to be made because of this effect, customers need to plan their CF storage to be sure to provide sufficient CF "white space" (unused CF storage capacity) to accommodate failure or planned reconfiguration of any given CF, in order to avoid a single point of failure for the sysplex Since, in such scenarios, structures will need to be rebuilt or undergo duplexing failover, it is imperative that your CF images be sized to have sufficient memory not only to accommodate all of your structures as they are CURRENTLY allocated, but also how they might BECOME allocated following a rebuild or duplexing rebuild process (including the additional structure size growth described above)

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation113 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 113 Thun Possible Differences between old/new Duplex Structure Sizes

Possible Differences between old/new Duplex Structure Sizes… • Documentation Update provided by APAR OA22860 (cont.):

For example, suppose that an installation has two CFs, each at the same CFLEVEL, and each of which has 4 GB of usable space that can be used for structure allocation purposes (after subtracting out the storage that is used for the CF image itself, and is thus unavailable for allocation of structures) Now suppose that a number of CF structures (all simplex) are allocated across the two CF images, and that the total size of all of the allocated structures is exactly 4 GB This configuration, most likely, is inadequate, despite having 4 GB of allocated structures and 4 GB of CF white space available for recovery of those structuresExtension of capabilities* If one of the CFs fails, the structures in that CF will need to be rebuilt in the surviving CF, and because ofToday’s the effect described above, the total size of the rebuilt structures will likely be somewhat larger thanCapabilities 4 GB If only 4 GB of space is available, some of the structures might not be able to be rebuilt, leading to sysplex problems and, potentially, outages The System Programmer Response description for MSGIXC582I is modified to read as follows in the "System Messages Vol 10 (IXC-IZP)" (SA22-7640) publication: System Programmer Response: If the allocation size of the structure exceeds CFRM policy definitions, consider whether or not it is necessary to make updates to the CFRM policy so as to bring the requested structure size definitions in line with the actual allocated structure size (as indicated by this message).

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation114 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 114 Thun Possible Differences between old/new Duplex Structure Sizes

Possible Differences between old/new Duplex Structure Sizes… • Documentation Update provided by APAR OA22860 (cont.):

Generally, if the structure has increased in size as a result of rebuilding or duplexing the structure into a CF at a different (higher) CFLEVEL than the one in which the original structure is allocated, and you intend to continue to use the higher CFLEVEL in the future, you should update the CFRM policy size definitions to be in line with the allocated structure size indicated by this message Failure to do so may result in the structure being allocated with an inadequate structure size, or possibly not being allocatable at all, when the structure is initially allocated in the future This may result in system, subsystem, or application outages Extension of Generally, if the structure has increased in size when rebuildingcapabilities* or duplexing the structure into a CF at the same CFLEVEL in which the original structure is allocated, there is no need to update the CFRM policyToday’s definitions to reflect the larger allocated structure size indicated by this message Capabilities Similarly, when the structure has been allocated larger than the policy INITSIZE value because the structure exploiter has requested a larger IXLCONN STRSIZE value, there is no need to update the CFRM policy definitions to reflect any size-related changes You should ensure that the increased CF storage allocation for the structure can be accommodated not only within the current CF, but also in any other CF images in which the structure is eligible to be allocated

Recommendation: to prevent possible rounding of CF structure sizes during allocation, in the CFRM policy always define your CF structures to be a muliple of the CF storage increment size (512K on z9 CFCC level 15, 1024K on z10 CFCC level 16)

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation115 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 115 Thun Administrative Data Utility Enhancements for LOGR Policy

IXCMIAPU Utility Enhancements for LOGR Policy • When defining a LOGR policy using IXCMIAPU, a new keyword, CONTINUE, can be specified, telling the program that if it encounters an error with a statement that it should continue on and process the remaining statements

• Before this enhancement, IXCMIAPU stopped at the first error • Then, after fixing that statement in error you had to resubmit the update job again, and it stopped at the next error, and so on…

• Example of the new CONTINUE statement:Extension of capabilities* //STEP1 EXEC PGM=IXGMIAPU //SYSIN DD * DATA TYPE(LOGR) REPORT(YES) CONTINUE Note that the position of the CONTINUE keyword DELETE LOGSTREAM NAME(stream1) is significant DELETE LOGSTREAM NAME(stream2) DELETE LOGSTREAM NAME(stream3) It only comes into effect at the point it is issued DELETE LOGSTREAM NAME(stream4) If an error is encountered before the CONTINUE DELETE LOGSTREAM NAME(stream5) keyword, the CONTINUE and all other statements DELETE LOGSTREAM NAME(stream6) will be ignored (although they will be syntax- checked)

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation116 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 116 Thun System Logger Offload and Staging Dataset Sizing

System Logger Offload and Staging Dataset Sizing

• In z/OS MVS Setting Up a Sysplex, SA22-7625, there is currently no description on the minimum and maximum size for logger offload and staging datasets • In the log stream definition in the logger policy this is specified with the LS_SIZEactually 4Kand less STG_SIZE than 2GB) keywords • Note: the values for LS_SIZE and STG_SIZE are specified in 4K units • APAR OW44389 documents that the minimumExtension size offor offload and staging datasets is 64Kb and the maximumcapabilities* size is 2Gb Today’s • As stated in APAR OW44389, the minimum size that can be specified for Capabilities LS_SIZE and STG_SIZE is 16, and the maximum size is 524287 (which is

• If a dataset size falls outside this range, message IXG256I will be issued and the dataset will be reallocated to the valid CI boundary that is closest to the requested dataset size

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation117 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 117 Thun System Logger Offload and Staging Dataset Sizing

System Logger Offload and Staging Dataset Sizing…

• Example for message IXG256I: IGD101I SMS ALLOCATED TO DDNAME (SYS11679) DSN (IXGLOGR.IFASMF.CHHCD001.SEC.A0000208 ) STORCLAS (SC01) MGMTCLAS (MC05) DATACLAS (IXLINEAR) VOL SER NOS FOR DATA COMPONENT= SLG001 IXG256I AN INCORRECT DATA SET SIZE WAS DETECTED, 2GB IS USED. DATA SET: IXGLOGR.IFASMF.CHHCD001.SEC.A0000208 IGD101I SMS ALLOCATED TO DDNAME (SYS11680) Extension of DSN (IXGLOGR.IFASMF.CHHCD001.SEC.A0000208 ) STORCLAS (SC01) MGMTCLAS (MC05) DATACLAScapabilities* (IXLINEAR) VOLToday’s SER NOS FOR DATA COMPONENT= SLG008 • Unfortunately,value greaterCapabilities thanpractical 524160 experience for LS_SIZE has shown and STG_SIZEthat message IXG256I is even issued with a (max.) LS_SIZE of 524287 (see above) • In fact, further research in old PMRs then led to the perception that the maximum value that can be specified for LS_SIZE and STG_SIZE is 524160 (which is 2Gb-128K) • As a conclusion, to prevent message IXG256I you must not specify a

System Logger Offload Performance Considerations

• Optimizing the Control Interval (CI) Size

• Increasing the size of the control interval for DASD log datasets can provide performance improvements in the areas of record processing speed and storage requirements • You can modify the CI size attribute of DFSMS Data Classes defined for DASD log data sets to gain performance improvements in I/O operations associated with the offloading of log stream data as well as during browse processing when retrieving log stream data from DASD log Extensiondata sets of capabilities* • A VSAM linear data set requires a CI size of from 4096 to 32768 bytes, and Today’s can be expressed in increments of 4096 bytes (the VSAM default is 4096 bytes) Capabilities • System logger is designed to obtain optimal I/O performance when DASD log datasets reside on devices with 3390 formats

• IBM recommends a CI size of 24576 bytes to realize optimal I/O performance if system logger DASD log datasets reside on devices with 3390 formats • Specify a DFSMS data class that is defined with a CI size of 24576 on the LS_DATACLAS parameter of a log stream definition to have DASD log datasets allocated with control interval sizes of 24576

System Logger Offload Performance Considerations…

• Optimizing the Control Interval (CI) Size…

• Important note: do not change the CI size attributes for staging datasets to be anything other than 4096

System logger requires that the CI size for staging datasets be set to 4096

If staging datasets are defined with characteristics that result in a CI size other than 4096, system logger will not use staging datasets to keep a duplicate copy of log data Extension of capabilities* Operations involving staging datasets defined with a CI size other than 4096 will fail toToday’s complete successfully Capabilities • System logger is designed to obtain optimal I/O performance when DASD log datasets reside on devices with 3390 formats

• If you intend to use control interval sizes other than 4096, you might need to modify the STG_DATACLAS specifications or change your DFSMS ACS routine • For further information refer to: • z/OS MVS Setting Up a Sysplex, SA22-7625

System Logger Offload Performance Considerations…

• The previous slides on this topic have already been included in Silvio’s Corner #69, 10/2008 • They have just been included again in this presentation as a reminder • The following information will provide some clarification of the factors which influence the maximum LS_SIZE that can be specified for offload dataset allocation Extension of 1) Size calculuation for offload datasets assumes that the recommended capabilities* 24Kb CISIZE is used Today’s 2) If Capabilitiesthe size calculation results in more than 2Gb because of CISIZE, message IXG256I will be issued

3) Size calculation is repeated and, if the resulting size is still >2Gb, twice the difference between the actual size needed and what was previously allocated will be subtracted from the offload dataset size 4) Step 3) will be repeated until the allocation size gets below 2Gb

System Logger Offload Performance Considerations…

• Summary and Conclusions

• The initial allocation using an LS_SIZE of 524287 failed with message IXG256I because the CISIZE used was 4K instead of the recommended 24K

• Had a CISIZE of 24K been used, the allocation would have worked

• The allocation after message IXG256I was slightly less than that what has been achieved using LS_SIZE 524160 because of the recalculation described previously Extension of • With a CISIZE of 4Kb, offload dataset allocationcapabilities* using an LS_SIZE of 524160 is theToday’s absolute maximum that can be achieve and, indeed, slightly more than that calculatedCapabilities

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation122 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 122 Thun GSE z/OS Expertforum Switzerland # 70, 31.3.-1.4.2009

Zürich | March 31, 2009 © 2009 IBM Corporation Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation123 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 123 Thun SMFPRMxx Parmlib Member: Prevention of possible Parsing Errors

Preventing possible SMFPRMxx Parsing Errors

• Using system symbols in a keyword statement of SMFPRMxx, and placing a comment (e.g. /*…*/) on the same line, may result in parsing errors during SMF initialization followed by unpredictable errors

• This errors may happen when the end (*/) of the comment statement end in column 72 of the appropriate SMFPRMxx line

• Currently, in “MVS Initialization and Tuning Reference” (SA22-7592) the relating syntax rules for SMFPRMxx are documented as follows:

Support for system symbols You can specify system symbols in SMFPRMxx. In addition, you can specify the &SID symbol when naming SMF data sets (on the DSNAME parameter). Syntax rules for SMFPRMxx The following rules apply to the creation of SMFPRMxx: Use columns 1 through 72. Do not use columns 73-80, since these columns are ignored. Comments may appear in columns 1-72 and must begin with ″/*″ and end with ″*/″.

• Unfortunately, the routines parsing the SMFPRMxx parmlib member during SMF initialization only consider columns 1-71 on each line

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation124 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 124 Thun SMFPRMxx Parmlib Member: Prevention of possible Parsing Errors

Preventing possible SMFPRMxx Parsing Errors…

• Possible errors that may happen when using system symbols in a statement and exploiting comments up to column 72 on the same line, are:

• Follow-on statements may be ignored and (unwanted) defaults being taken instead • For example, this may result in following errors:

– A wrong SMF recording interval (INTVAL) is being used

– A wrong job wait time (JWT) is being used

– A wrong SMF system id (SID) is being assigned to a z/OS LPAR

– Some SMF log streams (LSNAME) used to group SMF records into separate categories may be ignored and written to the default log stream (DEFAULTLSNAME) • The root cause for this possible problem is a doc change, which was introduced in 2001 • Before this doc change, “MVS Initialization and Tuning Reference” (SA22-7592) stated that comments may appear in columns 1-71 only • To prevent future parsing errors, this doc changed will now be reversed

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation125 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 125 Thun SMFPRMxx Parmlib Member: Prevention of possible Parsing Errors

Preventing possible SMFPRMxx Parsing Errors..

• Example of a parsing error reproduced with the Symbolic Parmlib Parser Tool provided in SYS1.SAMPLIB:

Original line as coded in SMFPRMxx: Specified keyword contains a system symbol and is followed by a comment ending in column 72

Interpreted line after parsing: Based on the length of the substitution text for the system symbol used, the remaining characters on the right are shifted to the left, and the ending slash of the comment is being ingnored As a consequence, the next statement (INTVAL) is being ingnored

RSU Settings on System z10

• The storage increment size on the z10 has increased from 64M to 256M • LPARs with RSU values should ensure the change in storage increment size does not adversely impact system performance • Example:

• z9 had an 8GB LPAR with an RSU value of 64M meant 4GB of the storage was considered reconfigurable • z10 with the same 64M RSU value meant 16GB would be reconfigurable

– Entire Central Storage allocation – Impact is a lot of high impact page movement – Very low capture ratios

• APAR OA27801

• New system health check to detect RSU over-specification conditions and issue a non-scrolling message warning of the potential impact

GETMAIN Enhancements in z/OS 1.10

• As of z/OS 1.10, small GETMAINs in user region private are satisfied from the beginning of the page instead of from the end of the page

• Vendor and legacy code may wrongly assumed the GETMAINed area would contain zeroes which may cause overlays or other incorrect processing • GETMAIN change was undocumented in z/OS 1.10

– There was an undocumented DIAGxx setting 'NUCLABEL ENABLE(IGVGPVTN)' to allow GETMAIN to work as it did previously • The change was made to improve performance of DB2 open processing for a large number of datasets • OA27291 - Behavior of GETMAIN for low private storage pools • Default behavior will remain the same as previous releases • DIAGxx keyword provided as a documented method to switch to new GETMAIN behavior

– VSM USEZOSV1R9RULES(YES|NO) – Provide new dispay command and IPCS support to determine the setting

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation128 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 128 Thun System z and High Availability: Cycle Steering

System z and High Availability: Cycle Steering

• Cycle steering mode is entered whenever the Mulit-Chip Module (MCM) has a cooling issue • Often due to a Modular Refrigeration Unit (MRU) failure • Processor will reduce cycle speed and notifies operating system – Reduction in Speed is typically from 4-6% – Higher reported SU/SEC impact are most likely due to not having MCL patch supplied in DR67L MCL211.G40965 in Bundle37 • z/OS will adjust to the lower speed – Issues ENF 20, Changes SU/SEC value (R72MADJ), internal reactivate the WLM policy, stop/restart intervals, issues a SMF 99 subtype 10 record to audit the speed change, issues IWM063I at each change in processor speed • The MSU rating of the processor is not changed – Reduction in processor speed seen as temporary error situation – For sub-capacity users they should follow the "unusual conditions" directions in the SCRT Users Guide • WSC Flash will be published soon describing functionality and system processing – Reduction in processor speed seen as temporary error situation

GTF Traceusing the toSPACE an parameter!external Dataset

• A note of Caution • When you use the Generalized Trace Facility (GTF) and specify an external Dataset in the GTF procedure to hold the trace data, do not specify the RLSE option while

• Output datasets are opened and closed more than once while GTF runs

• If you don’t follow this restriction and use RLSE, the GTF external trace dataset will onlycontain the data which has been written between the last open and close! • For information on using the Generalized Trace facility (GTF) refer to:

– z/OS MVS Diagnosis: Tools and Service Aids, GA22-7589

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation130 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 130 Thun GSE z/OS Expertforum Switzerland # 70, 31.3.-1.4.2009

Zürich | March 31, 2009 © 2009 IBM Corporation Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation131 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 131 Thun Recommended Readings

Redbooks and Redpapers DFSMS V1.10 and EAV Technical Guide, SG24-7617-00 IBM System z10 Capacity on Demand, SG24-7504-01 z/OS Metro Global/Mirror Incremental Resync, REDP-4429-00 Java Security on z/OS - The complete View, SG24-7610-00

IBM DS8000 and z/OS Basic HyperSwap, REDP-4441-00 IBM System Storage DS8000 Series Architecture and Implementation, SG24-6786-05

DS8000 Copy Services for IBM System z, SG24-6787-04 IBM System Storage DS8000 Series: IBM FlashCopy SE, REDP-4368-00 Parallel Sysplex Operational Scenarios, SG24-2079-01 Exploiting IBM System z in a Service-Oriented Architecture, SG24-7651-00

z/OS Distributed File Service zSeries File System Implementation z/OS V1R10, SG24-6580-03

Getting Started with InfiniBand on System z10 and System z9, SG24-7539-01

Redbooks News IBM Redguides - A new type of IBM Redbooks publication IBM Redguides focus on the business view of technology that: - Solves business issues - Provides business value - Or enables competitive advantage by applying existing technologies or exploring a roadmap for emerging technologies To have these new publications included in your weekly Redbooks newsletter please update your subscription to include "IT Business Perspectives"

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation133 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 133 Thun GSE z/OS Expertforum Switzerland # 70, 31.3.-1.4.2009

Zürich | March 31, 2009 © 2009 IBM Corporation Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation134 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 134 Thun Silvio’s Corner Doc Jukebox

The following Documentations will also be supplied on the GSE z/OS Expert Forum CH Homepage at http://www.thebrainhouse.ch/gse/silvio

z/OS Hot Topics Newsletter #20 (February 2009), GA22-7501-15 Coupling Facility Configuration Options, ZSW01971-USEN-12 System-managed CF Structure Duplexing, ZSW01975-USEN-07 z/OS V1R10 AutoIPL for Sysplex Failure Management Users - Restriction lifted, FLASH10655 z/OS V1R10 and z/OS Global Mirror, FLASH10652 zHPF for System z Technical Summary for Customer Planning, ZSW03058-USEN-00 Coupling Facility Disruptions and the DB2 Reorg Utility, FLASH10669 Withdrawal of zOS Function beginning with z/OS 1.5 through current z/OS Release, FLASH10451 System z and DS8000: Accelerating SAP Deposits Management Workload with SSD Drives, WP101442 GDPS 3.6 Update Mainframe Executive Issue January/February 2008 Mainframe Executive Issue March/April 2008

How to get the Handouts for this Session

• The Handouts for this Session and the Documentations listed in the “Doc Jukebox” will be provided on the GSE z/OS Expertforum CH Homepage

http://thebrainhouse.ch/gse/silvio To have immediate access to the handouts for this session and the doc‘s provided in the „Doc Jukebox“ the day after this conference, please write down this link!

Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation136 TheExpertforum Future runsCH #69, on 21.-22.10.2008, System z Page 136 Thun GSE z/OS Expertforum Switzerland # 70, 31.3.-1.4.2008

The End...

Zürich | 26. Oktober 2004 © 2004 IBM Corporation Zürich | 26. Oktober 2004 Silvio Sasso's Corner, GSE z/OS © 2004 IBM Corporation137 TheExpertforum Future runsCHZürich, #69, on March21.-22.10.2008, System 31, 2009 z © 2009 IBMPage Corporation 137 Thun