IBM Data Studio ̛ • Faqs / Tutorials • Downloads • Forum / Blogs • Join the IBM Data Studio User Community

Getting The Most Out of DB2 in Your New Applications Curt Cotner IBM DB2 Development, IBM Fellow and VP

Session Code: S03 Tue May 3, 10:15-11:45 | Platform: z/OS, LUW IBM Information Management

Optimizing Your Java and WebSphere Applications with Data Studio and Optim Tools

Data Studio -- pureQuery tooling is in the box! Speed up problem isolation for developers – even when using frameworks

Capture application-SQL-data object correlation (with or without the source code) Trace SQL statements to using code for faster problem isolation Enhance impact analysis identifying application code impacted due to database changes Answer “Where used” questions like “Where is this column used within the application?” Use with modern Java frameworks e.g. Hibernate, Spring, iBatis, OpenJPA

Data Studio -- OPM Performance View

Table Columns Table Actions Golf score Export – Exports the data to file SQL statement Remove All – clears the table of all rows Annotation Inputs for host variables Row Actions Total Server Time Open in SQL Editor – opens SQL editor Average Server Time with selected SQL statements Number of Rows Filter – Hides all but the selected rows Number of Rows Examined Remove – removes selected row(s) Average Number of Row Returned CPU time Number of Sorts Number of RSCANs Number of ISCANs Number of physical IOs Number of logocial IOs

Data Studio -- New Routine Creation Wizard

Data Studio -- Templates Management

Data Studio -- Server Profile Management

Data Studio -- Deployment Management

JDBC – acceleration for any JDBC application ̛ Convert dynamic SQL to static SQL ̛ Replace problem queries without changing the source ̛ Remove literals from SQL to get better statement cache hit ratios Hibernate/OpenJPA/iBatis – acceleration for persistence layers ̛ Improved SQL “batch” peformance ̛ Auto-tuning of Hibernate and OpenJPA peristence options SQL-friendly APIs for OO access to relational ̛ Object to relational mapping ̛ APIs that can be tailored to return XML, JSON, arrays, etc. Improved management, monitoring, problem determination ̛ Tracks SQL back to the Java class file and line number ̛ Enables performance monitors to report by application name Provides the foundation for improved developer tooling ̛ Syntax assist, code generation, performance reporting, etc.

CICS AOR1 z/OS LPAR Txn1 - Pgm1 - Pgm2

CICS AOR2 TxnA DB2PROD - PgmX - PgmY CICS AOR3 Txn1 - Pgm1 - Pgm2

App CPU PLAN DB2 Accounting for CICS apps allows you to study performance data from many perspectives: Txn1 2.1 • By transaction (PLAN name) TN1PLN • By program (package level accounting) TxnA 8.3 • By address space (AOR name) TNAPLN • By end user ID (CICS thread reuse)

This flexibility makes it very easy to isolate performance problems, perform capacity planning exercises, analyze program changes for performance regression, compare one user’s resource usage to another’s, etc.

E DB2 or IDS D J P D B a e

B t r Q USER1 a s 2

A1 i u A s

J e t c

A4 e a r c y n v e

a c

L USER1 s

A3 e a s D

A2 n

L r L g i a v o u y

e USER1 A5 g e a r i c r A6 g e

What is visible to the DBA? - IP address of WAS app server - Connection pooling userid for WAS User CPU - app is running JDBC or CLI PACKAGE USER1 2.1 JDBC What is not known by the DBA? USER1 8.3 JDBC - which app is running? USER1 22.0 JDBC - which developer wrote the app? - what other SQL does this app issue? - when was the app last changed? - how has CPU changed over time? - etc.

z/OS LPAR Unix or Windows

WAS 21.22.3.4 CICS AOR2 TxnA (Set Client App=TxnA) TxnA (PLANA) - ClassX - PgmX - ClassY - PgmY

Data Studio and pureQuery provide the same granularity App CPU for reporting WebSphere’s DB2 resources that we have TxnA 2.1 with CICS: TxnB 8.3 • By transaction (Set Client Application name ) • By class name (program - package level accounting) • By address space (IP address) • By end user ID (DB2 trusted context and DB2 Roles)

• Hibernate: http://www.ibm.com/developerworks/data/library/techarticle/dm- 1008hibernateibatispurequery1/index.html?ca=dnw-1133&ca=dth-i • iBatis: http://www.ibm.com/developerworks/data/library/techarticle/dm- 1009hibernateibatispurequery2/index.html • Spring: http://www.ibm.com/developerworks/data/tutorials/dm0806hsing/index.html

• Generate SQL and bind as Static Packages before deploying application • Improve performance with heterogeneous batching & Static Execution • Identify and replace problematic SQL with hand-tuned alternative SQL • Track SQL requests back to the framework query, including java source file/line #

© 2011 IBM Corporation IBM Information Management Object/Relational Mapping pureQuery can monitor your Java application’s object access patterns and class Customer { public String Name; automatically select the optimal eager/lazy public String mailingAddress; fetch setting for each SQL statement!!! public String daytimePhone; public Order[] recentOrders; Table Column Type public Complaint[] complaintHistory CUST NAME CHAR(64) … CUST ADDRESS CHAR(128) } CUST PHONE_NUM CHAR(10)

Table Column Type COMPLAINTS CUST_NAME CHAR(64)

COMPLAINTS COMP_ID CHAR(18) Table Column Type COMPLAINTS DESC VARCHAR(32K) ORDERS CUST_NAME CHAR(64) ORDERS ORDER_NUM CHAR(12)

ORDERS DATE_ORD DATE Table Column Type CREDIT_DATA CUST_NAME CHAR(64)

CREDIT_DATA CARD_NUM CHAR(18) Table Column Type CREDIT_DATA VALID_UNTIL DATE ORDER_ITEMS ORDER_NUM CHAR(12) ORDER_ITEMS ITEM CHAR(128)

ORDER_ITEMS QUANTITY SMALLINT

© 2007 IBM Corporation IBM Information Management Eager vs. Lazy Fetch class Customer { public String Name; “Select object(customer) WHERE…” public String mailingAddress; public String daytimePhone; public Order[] recentOrders; public Complaint[] complaintHistory … } “SELECT CUST.NAME, CUST.ADDRESS,CUST.ADDRESS … … FROM CUST WHERE…” FROM“SELECT CUST, ORDERS.ORDER_NUM COMPLAINTS, … WHERE …” Table Column Type CUST NAME CHAR(64) “SELECTCOMPLAINTS.COMP_IDORDERS, ORDER_ITEMS … WHERE …” CUST ADDRESS CHAR(128) WHERE…”O CUST PHONE_NUM CHAR(10) O Table Column Type O Table Column Type COMPLAINTS CUST_NAME CHAR(64) COMPLAINTS COMP_ID CHAR(18) ORDERS CUST_NAME CHAR(64) COMPLAINTS DESC VARCHAR(32K) ORDERS ORDER_NUM CHAR(12) ORDERS DATE_ORD DATE

Table Column Type Table Column Type CREDIT_DATA CUST_NAME CHAR(64) ORDER_ITEMS ORDER_NUM CHAR(12) CREDIT_DATA CARD_NUM CHAR(18) ORDER_ITEMS ITEM CHAR(128) CREDIT_DATA VALID_UNTIL DATE ORDER_ITEMS QUANTITY SMALLINT

© 2007 IBM Corporation IBM Software Group – Information Management Hibernate AutoTuning • Automatically identify and fix common problems with Java Persistence applications • hundred’s of SQL per transaction • tens of unwanted joins per SQL https://www.ibm.com/services/forms/preLogin.do?source=swg-iopahb

Captures SQL for Java applications ̛ Custom-developed, framework-based, or packaged applications Bind the SQL for static execution without changing a line of code ̛ New bind tooling included Delivers static SQL execution value to existing DB2 applications ̛ Making response time predictable and stable by locking in the SQL access path pre- execution, rather than re-computing at access time ̛ Limiting user access to tables by granting execute privileges on the query packages rather than access privileges on the table ̛ Aiding forecasting accuracy and capacity planning by capturing additional workload information based on package statistics ̛ Drive down CPU cycles to increase overall capability Choose between dynamic or static execution at deployment time, rather than development time

In-house testing shows double-digit reduction in CPU costs over dynamic JDBC % increase/reduction in CPU per transaction compared to JDBC using Type 4 driver Normalized Throughput by API for JDBC Type 4 Driver

500 r e p

U 25% B P ) 400 D C 15% J R

n T i o I

t (

n t 6% d o u i e

p 300 t r c a h

g 524 u p u d

485 m e o

r 446 r o / c h 420 e T

200 s 360 n

a -14% s d e e n r z a i 274 c r l t n a i

m 100 r

% -35% o

N -50%

c 2 c i c A i t ti B P a 0 J J m t a a t E S 2 n S C c c c . i i i A t t y t d B B P m a a D p o J J D t t a O h E J n S S d t t y o e n d h n D z o t M i ie h e l d m t Q i C o t e M p h t p M Q e O p Q t M p n Q e i IRWW – an OLTP workload,p l Type 4 driver C Cache hit ratio between 70 and 85% 15% - 25% reduction on CPU per txn over dynamic JDBC

Table 1, operation 1 Data Server Table 1, operation 2 Table 1, operation 3

Table 2, operation 1 Table 2, operation 2

Table 3, operation 1 Table 1, operation 4

JDBC batching used by Hibernate Batcher is currently limited The advantage of Heterogenous Batching ̛ Cannot batch entities that map to multiple tables • Primary and Secondary tables. 500 • Inheritance Join and Table per class

) 450

strategies c ̛ Cannot batch different operations e 400 against same table S 350 m ( • Field level updates

e 300 • Insert, update m HeteroBatching i 250

̛ Cannot batch different entities T 200 No Batching ̛ Each batch is a message to the d e

database s 150 JDBC Batching p

a 100 l

pureQuery heterogeneous batching E 50 plug-in for Hibernate on the other 0 hand ̛ Can batch entities that map to multiple 0 20 40 60 80 100 tables ̛ Can batch different operations against # of operations per transaction the same table ̛ Can batch different entities into a single batch * Preliminary findings based on validation with a test designed to demonstrate ̛ Combines insert, deletes, updates into heterogeneous batching differences. This is not intended to be a formal single batch benchmark.

JDBC app

INSERT INTO T1 VALUES(‘ABC’,2,’DEF’) pureQuery INSERT INTO T1 Runtime conversion VALUES(:h1,:h2,:h3)

• pureQuery can identify statements that use no parameter markers, and strip the literals out at runtime • significant performance gains: • less CPU cost at PREPARE • better use of dynamic statement cache

WebSphere – a first class OPM citizen

WebSphere support

has a built-in support for OPM (starting with V6.0.21), allowing to ... 47 57.67 47 57.67 • identify problems with 52 0.34WAS connection pool • identify differences in the configuration of nodes in a WAS cluster • check if a node in a WAS cluster has a system or network problem

OPM can tell you where the query came from …

Application source

Capture SQL with pureQuery runtime

Upload collected metadata into OPM

SELECT * FROM DB2ADMIN.I…

SELECT * FROM DB2ADMIN.INVENTORY

TestOPM my.test main 13 blahVer N/P N/P blah capture…

OPM Extended Insight (EI) Overview dashboard

OPM Extended Insight Dashboard – Client Details

Open Source Scripting Languages

Application

Django Ruby on Zend/Drupal Rails

Perl Python PHP Ruby

DB2 CLI

DB2

• Up-to-date with latest Django/Rails/Zend releases. • All open source drivers and adapters are available on the DB2 media • Python: http://code.google.com/p/ibm-db/ • Ruby: http://rubyforge.org/projects/rubyibm/ • PHP: http://pecl.php.net/package/ibm_db2/ , http://pecl.php.net/package/PDO_IBM • Perl: http://search.cpan.org/~ibmtordb2/

• DB2 Drupal Support – A widely used PHP based Web Content Management System – DB2 support for Drupal 6 publicly available shortly – Drupal 7 support to follow

• SQL Generation for Java API based Query Systems before deployment • Complete Accelerator support for Hibernate / JPA Criteria Queries

Current NoSQL Landscape Tabular Stores Key Value Stores 6 + 23 +

Object Stores Others 12 +

Currently there are more than 100+ noSQL systems Not clear which of these will survive

Some apps need very few database features, but need high scale

Desire to avoid data/schema pre-design altogether for simple apps

Need for a low-latency, low-overhead API to access data

Simplicity – I don’t need fancy indexing - just fast lookup by primary key

Focus on Give up

Commodity servers, networking, disks Relational data model Easy elasticity and scalability to SQL APIs multiple racks (10s to 100s of servers) Complex queries (joins, secondary Fault-tolerance and high availability indexes, ACID transactions)

Two Worlds

Transactional Analytics Custom high-end OLTP for financial Managing updates applications Support for random access and indexing Scaleout datastores for Cloud/Web 2.0 Scaleout content store

Examples Examples ¾ MemcacheDB, Cassandra, Dynamo, Voldemort, ¾ Bigtable, HBase, Hypertable SimpleDB, Gigaspaces, Websphere eXtreme Scale

© 2011 IBM Corporation Graph Stores Datastores which can handle Large but sparse predicate set for a subject Can do link following Optionally support other graph like operations Transitive closure computation, Shortest paths Common Model : RDF (Resource Description Framework)

A way of expressing linked data : Subject – Property – Object (Value) Natural graph structure. Ideally suited for representing information extracted from unstructured sources or graph-based information Example : IBM hasLocation Hawthorne Hawthorne LocatedIn New York Applications include semantic web, data integration, metadata access, inferencing applications in healthcare, government, telecom,oil

Open source examples : Jena, Sesame

Allows one to store and query values by key

Quick response time, distributed, with optional transaction support

Often used to cache data and objects in RAM for Web 2.0 applications

Open source examples : MemcacheD, Redis

Common Interface : Get Put Remove

Scalable, high-performance, schema-free, document-oriented databases.

Manages collection of JSON like documents.

Used in Web 2.0 scenarios

Open source examples : CouchDB, MongoDB

Details of CouchDB : A document could be one or more field/value pairs expressed as JSON. The field value pairs might be very sparse across documents Every document in a CouchDB database has a unique id.

Example of a document: "Subject": “Cricket World Cup“ "Author": "Rusty“ "PostedDate": “2/23/2011“ "Tags": [“cricket", “world cup", “overs“] "Body": “The Cricket World Cup is currently ongoing. It is played over 50 overs”

Data repository for XML data. Storage scheme could be in native form

Useful when there is variable schema in the data or transmission in XML

Transaction support . Support for indexing and query processing

Applications include creating business reports, SOA, webservices, forms

Open source examples : Sedna, BaseX

Interface : XQuery/XPath

Scalable, distributed repository for structured data

Various levels of transaction support

Richer support for indexing and query processing

Applications include analytics, storage of user metadata information

Open source examples : HBase, Cassandra

Interface : APIs vary from implementation to implementation Thrift Java

Object databases vs. relational ¾ Gemstone, ObjectStore, Objectivity, etc. In-memory databases vs. relational ¾ TimesTen, SolidDB, etc. Persistence frameworks vs. relational ¾ Hibernate, OpenJPA, etc. XML databases vs. relational ¾ Tamino, BaseX, etc. Open Source vs. commercial relational ¾ MySQL, PostgreSQL, etc. Column-store databases vs. relational ¾ Sybase IQ, Vertica, etc.

A very broad set of attributes must be available: ¾ Ability to scale up to large SMPs ¾ Ability to scale out to multiple boxes ¾ Ability to exploit evolving microprocessor and memory architectures (SMP, multi-core, instruction pipeline, FGPA, SMT, etc.) ¾ Parallelism at many levels in the architecture ¾ High availability solutions (replication, failover, high speed restart/recovery, online schema change) ¾ Security/Auditing/Compliance ¾ Compression, archiving, and overall management of storage ¾ Performance/Monitoring/Tuning tool suite ¾ Modeling/Design/Development tool suite ¾ Programming interfaces to the constantly changing application space (Cobol, C/C++, Java, .NET, Ruby, PHP, Python, etc.) ¾ Adoption by packaged ISV applications ¾ Wide range of end-user reporting tools ¾ Enterprise-level problem determination, service, and support This massive infrastructure is needed regardless of the APIs you choose to expose to the end users and applications, or the logical/physical data model of the data. History has shown that we can add a new set of APIs or a new data model to DB2 much faster than the alternative solutions can build out a full solution with all the above attributes.

Immediately takes advantage of: • DB2 storage infrastructure • DB2 backup/recovery • DB2 pureScale technology • DB2 performance monitoring Data Analyzer RDF queries • DB2 security and auditing (SPARQL) • DB2 high-volume utilities SchemaCreator • etc.

We’ve found that Jena using DB2 Query translator RDF Loader out-performs the open source SPARQL-to-SQL Jena implementation by up to 300%.

RDF query Relational store uses Relational store Processing Query processor Functions Indexing

Put (Key, Value)

Get (Key) Remove(Key)

Value

Data Store

Key/value access is very well optimized with the recently GA support for hash data access in DB2 11 for z/OS. Range partitioning and DPSIs also help optimize for key/value access patterns.

Applications include creating business reports, SOA, webservices, forms etc

Product vs -SKU, price, … -Color, web ratings, green ratings, …

Merchant -StoreID, address, volume sales, profit… Warehousing -Employees, Neighboorhood, … domain Events -Person x bought product y at merchant z -Financial trace (for fraud), Clickstream, Hadoop domain Complaints, Location Based Services (smart phone customer tracking), hospital records

-non traditional, semi-/un-structured data, with no heavy DBA work up front -Algorithmic analytics that is not just SQL programming– requires high-scale analytics server

Microsoft .Net 4.0 and VS 2010 Integration

Single Offering of VSAI for Visual Studio 2005, 2008 and 2010 Install – Single VSAI installer for all installed versions of Visual Studio – Integrated VSAI Help with VS 2010 using the new VS 2010 help integration processes – New program group enabling registration of VSAI for versions of VS installed after VSAI install. The program group also has links to online documentation and the .Net Development with DB2 and Informix forums. Seamless integration of VSAI functionality into VS 2010 development environment – Support for Microsoft Windows® Presentation Foundation (WPF) and Extensible Application Markup Language (XAML) with the ability to generate sample XAML application from an IBM database table, view, or procedure – Ability to access and manage DB2 and Informix data connections – Create and modify stored procedures, functions, tables, views, indexes, triggers, generate web services etc. – Support for debugging SQL and SPL procedures. – XML support including support for XML data type, XML schema repository and annotated XML schemas. – Support for creation of IBM Database Projects, Windows or ASP.NET Web applications with DB2 and Informix Data Sources – Support for SQL capture and running SQL statically

A new .IBM Database Provider for .NET Framework 4.0 – Includes all previously supported functionality for .NET framework 1.1 and 2.0 including but not limited to XA transactions, sysplex and XML support – Ability to configure .NET framework through client installer for automatic re- direction of 1.1 and 2.0 .NET applications to run against the IBM 4.0 provider – Works with full version of .NET Framework 4.0 as well as its smaller- footprint version known as the ‘Client Profile” A new Entity Provider for Entity Framework 4.0. – Support for optimized and seamless mapping of .NET canonical functions in LINQ queries into SQL functions on IBM DB2 and IDS servers. – Support for translation of LINQ queries into SQL queries that exploit the database features (e.g. indexes) on IBM DB2 and IDS servers.

• A data client and server configuration management solution providing centralized view and control of data client and server relationships

What is its value to DB2 customers?

• Gain insight into IT environments through automated discovery of data client and data server configurations and relationships

• Reduce database outages and enforce compliance with centralized control and analysis of distributed database access

• Help maintain SLAs of database applications with configurable client connection properties and connection-level load balancing

Functions and features

• Registry of server connection properties for the enterprise

• Links to the correct IDA model, pureQuery metadata, and OPM performance repository for each database server

• Historical view of data client and server configuration changes

• Push down client configuration changes to data clients

• Isolate or roll-out apps to specific data server(s) through connection-level load balancing

• Workload Manager classification can be updated and validated against expected SLAs

51 © 2010 IBM Corporation SYSTEM 2 Discovery IOCM WEB UI Applications SYSTEM 1 DB2 A IOCM Server DB2 B DB2 DB2 C DB2 JCC DRIVER Repository Catalogs M1 M2 M3 CMX Controller Discovery Tasks •DB2 z/OS DataSharing DB2 z •DB2 z/OS 3-way •DB2 Distributed Data Sharing •DB2 DPF •DB2 pureScale • Automatic discovery of clients connected to database servers • IP address, connection properties, client driver software level, etc. • Automatic discovery of database server changes • Server name, system parameters, database schema, etc. • Repository contains a history of how these things change over time: • Which clients are using client software older than DB2 9.7.1? • What changes have occurred on DB2A since Sept 1, 2010?

52 © 2010 IBM Corporation Enterprise Connection Management SYSTEM 2 IOCM WEB UI Applications SYSTEM 1 DB2 A IOCM Server DB2 B DB2 DB2 C DB2 JCC DRIVER Repository Catalogs M1 M2 M3 •DB2 z/OS DataSharing CMX Controller Discovery Tasks •DB2 z/OS DB2 z •DB2 Distributed 3-way •DB2 DPF Data Sharing •DB2 pureScale • Centrally manage inventory of database connection properties • Automates connection definition for clients • Tailored list of database connection “folders” for Web and Eclipse users • Centrally manage the relationships between systems and IDA models: • Record with IDA model file is current for each of the database servers • Identify which OPM repository should be used for each database server • pureQuery metadata for each database server • Centrally manage client connection properties: • Provide defaults or overrides for IP address, port, isolation level, trace on/off, etc. • Penalty boxing – force misbehaving app to use exclusively member X of system

© 2009 IBM Corporation Capture Replay Validation report enables drill- down on failed replays, like Different Return Codes Open | d Welcome X Capture / Replay X Move Diff Rows Returned Create Test Database SQL Workloads Validation Report X Adjustable >= 5% to 10% Validate that the replay matches the original capture. Remove failed SQL and related transactions.

Overview

Replay Success Response Time

Successful SQL Replays 9000 / 10000 90% JJJJJJJJJJJJ

Failed SQL Replays 1000 / 10000 10% JJ PeakOrders y Different Return Codes 300 / 10000 3% J PeakOrders[5]

y Different # Rows Returned 200 / 10000 2% J

J y Different # Rows Updated 300 / 10000 3% 0 1 2 3 4

y Missing SQL 0 / 10000 0% Elapsed Time (Hours)

Successful Transaction Replays 500 / 800 63% JJJJJJJJ PeakOrders[0] Total 240:35 JJJJJJJJJJJJJJJJJJJJ

Failed Transaction Replays 300 / 800 27% JJJJ PeakOrders[5] Total 220:25 JJJ JJJ JJJ JJJ JJJ J

y Different Return Codes 100 / 800 12% JJ Total Improvements 25:30 10% JJJ

y Different # Rows Returned 60 / 800 7% J Total Regressions 5:20 2% J

y Different # Rows Updated 70 / 800 8% J SQL with >= 5% Improvement 300 / 10000 3% J

y Missing Transactions 0 / 800 0% SQL with >= 5% Regression 200 / 10000 2% J

New SQL 50 Trans with >= 5% Improvement 10 / 250 3% J

New Transactions 2 Trans with >= 5% Regression 2 / 250 1% J

SQL Execution (1000 / second) Rows Returned (10,000 / second)

Open | d Welcome X Capture / Replay X

Create Test Database SQL Workloads Performance Report X

Top ‘N’ SQL Statements Comparison

Sort by: Total Response Time Change | d Number of Statements: 5 | d Show: Both Regressions and Improvements | d

SQL Regressions

Total Response Time Average Response Time Rows Rows Return Baseline Change in Baseline Change Change Baseline Change Change Updated Returned Code Statement Text Executions Executions (sec) (sec) d (%) (sec) (sec) (%) (changes) (changes) (Changes)

UPDATE DBPARTITION… 10050 0 200.849 +100.427 +50% 0.059 +0.027 +50% 0 0 0 JJJJJJJJJJJJJJJ JJJJJJJJJ JJJJ JJJJJJJJJ

INSERT T1.AGENT_ID … 25 0 896.433 +90.708 +10% 12.433 +1.208 +10% 0 0 0 JJJJJJJJJJJJJ JJJ JJJJJJJJJJJJJ JJJ UPDATE DBPARTITION… 2234 0 1765.623 +85.676 +5% 1.223 +0.176 +5% 0 0 0 JJJJJJJJJJJJ J JJJ J INSERT T2.AGENT_ID … 307 0 248.321 +78.786 +32% 0.821 +0.286 +32% 0 0 0 JJJJJJJJJJJ JJJJJJJ JJJJJJJJJJJ JJJJJJJ SELECT * FROM T3 … 529 0 215.765 +75.653 +27% 0.565 +0.133 +27% 0 0 0 JJJJJJJJJJ JJJJJJ JJJJJJJJJJ JJJJJJ

SQL Improvements

SELECT T2.AGENT_ID … 100 0 1874.321 -195.427 -12% 10.874 -22.337 -12% 0 0 0 JJJJJJJJJJJJJJJ JJJ JJJJJJJJJJJJJ JJJ

SELECT T1.AGENT_ID … 345 0 135.987 -120.7083 -95% 0.421 -0.398 -95% 0 0 0 JJJJJJJJJJJ JJJJJJJJJJ J JJJJJJJJJJ SELECT DBPARTITION… 15454 0 1201.787 -55.676 -5% 0.123 -0.059 -5% 0 0 0 JJJJJ J J J SELECT T2.AGENT_ID … 4443 0 86.874 -20.786 -23% 0.013 -0.007 -23% 0 0 0 JJJ JJJJJ JJJ JJJJJ SELECT DBPARTITION… 56 0 753.765 -15.653 -2% 15.345 -1.334 -2% 0 0 0 JJ J JJ J

Optim Performance Manager Compare performance details of this statement across the two workload runs SQL Statement Comparison Report SQL Statement

SELECT B.COL1, B.COL3, B.COL5, B.COL6, B.COL12 FROM T1.SETLMNT, BRANCH B, ADDR A WHERE S.TRANS_NO = ?, AND S.TRANS_PROC_DT < '9999-12-31‘ AND YEAR (S.TRANS_TARGET_DT) = ‘2002’ AND S.TRANS_TYPE IN (‘A1', Tune SQL ‘A2', ‘A3', ‘Z9') AND S.TRANS_CD IN ('EOD', 'IMD', ‘UGT') AND S.TRANS_SETL_DT = ? AND B.BRANCH_EFF_DT <= ? AND B.BRANCH_INACTIVE_DT > ?

Metric Test Replay 1 Test Replay 2 % Change Average Elapsed Time (seconds) Executions 508 508 0% 2 Average Elapsed Time (sec) 0.567 0.876 +45% 1.8 JJJJJ 1.6 1.4 Test Replay 2 Total Elapsed Time (sec) 254.453 367.463 +45% JJJJJ 1.2 1 Test Replay1 Average CPU Time (sec) 0.0567 0.1376 +275% 0.8 JJJJJJJJJJJ 0.6 Total CPU Time (sec) 25.4567 69.876 +275% 0.4 JJJJJJJJJJJ 0.2 Average System CPU Time (sec) 0.0062 0.0121 +175% 0 JJJJJJJJJJJ 0 1 2 3 4 Execution Time (Hours) Total System CPU Time (sec) 2.3445 6.6503 +175% JJJJJJJJJJJ Average User CPU Time (sec) 0.0434 0.1221 +275% Average CPU Time (seconds)

JJJJJJJJJJJ Total User CPU Time (sec) 20.432 57.876 +275% 0.35 JJJJJJJJJJJ 0.3 Average Get Pages 4.01 4.40 +15% 0.25 Test Replay 2 JJ Total Get Pages 2000 2300 +15% 0.2 Test Replay1 JJ 0.15 Sorts 0 0 0% 0.1 0.05 Table Scans 0 0 0% 0 0 1 2 3 4 Execution Time (Hours)

Open | d Welcome X Capture / Replay X

Create Test Database SQL Workloads Performance Report X

Top ‘N’ Transaction Comparison

Sort by: Total Response Time Change | d Number of Statements: 5 | d Show: Both Regressions and Improvements | d

Transaction Regressions

Total Response Time Average Response Time Rows Rows Return SQL Baseline Change Change Baseline Change Change Updated Returned Code Transactions Type Statements (sec) (sec) d (%) (sec) (sec) (%) (changes) (changes) (Changes)

APPNAME23 App Name 25 200.849 +100.427 +50% 0.059 +0.027 +50% 0 0 0 JJJJJJJJJJJJJJJ JJJJJJJJJ JJJJ JJJJJJJJJ

ACCTSTR456 App Name 5 896.433 +90.708 +10% 12.433 +1.208 +10% 0 0 0 JJJJJJJJJJJJJ JJJ JJJJJJJJJJJJJ JJJ ACCTSTR789 Acnt Str 73 1765.623 +85.676 +5% 1.223 +0.176 +5% 0 0 0 JJJJJJJJJJJJ J JJJ J PKGNM123 Package 15 248.321 +78.786 +32% 0.821 +0.286 +32% 0 0 0 JJJJJJJJJJJ JJJJJJJ JJJJJJJJJJJ JJJJJJJ SQL_SEQ_567 SQL Seq 75 215.765 +75.653 +27% 0.565 +0.133 +27% 0 0 0 JJJJJJJJJJ JJJJJJ JJJJJJJJJJ JJJJJJ

Transaction Improvements

SQL_SEQ_765 SQL Seq 15 1874.321 -195.427 -12% 10.874 -22.337 -12% 0 0 0 JJJJJJJJJJJJJJJ JJJ JJJJJJJJJJJJJ JJJ

SQL_SEQ_988 SQL Seq 43 135.987 -120.7083 -95% 0.421 -0.398 -95% 0 0 0 JJJJJJJJJJJ JJJJJJJJJJ J JJJJJJJJJJ ACCTSTR333 Acnt Str 20 1201.787 -55.676 -5% 0.123 -0.059 -5% 0 0 0 JJJJJ J J J ACCTSTR555 Acnt Str 1 86.874 -20.786 -23% 0.013 -0.007 -23% 0 0 0 JJJ JJJJJ JJJ JJJJJ APPNAME767 App Name 56 753.765 -15.653 -2% 15.345 -1.334 -2% 0 0 0 JJ J JJ J

SQL list for selected transaction. Open | d Welcome X Capture / Replay X

Create Test Database SQL Workloads Performance Report X

Top N Transactions Report > SQL List for Transaction APPNAME23

SQL List for Transaction APPNAME23

Total Response Time Average Response Time Rows Rows Return Updated Returned Code Baseline Change in Baseline Change Change Baseline Change Change (changes) (changes) (Changes) Statement Text Executions Executions (sec) (sec) d (%) (sec) (sec) (%)

UPDATE DBPARTITION… 10050 0 200.849 +100.427 +50% 0.059 +0.027 +50% 0 0 0 JJJJJJJJJJJJJJJ JJJJJJJJJ JJJJ JJJJJJJJJ

JJJJJJJJJJJJJJJ JJJ JJJJJJJJJJJJJ JJJ SELECT T1.AGENT_ID … 345 0 135.987 -120.7083 -95% 0.421 -0.398 -95% 0 0 0

JJJJJJJJJJJ JJJJJJJJJJ J JJJJJJJJJJ SELECT DBPARTITION… 15454 0 1201.787 -55.676 -5% 0.123 -0.059 -5% 0 0 0

JJJJJ J J J SELECT T2.AGENT_ID … 4443 0 86.874 -20.786 -23% 0.013 -0.007 -23% 0 0 0

JJJ JJJJJ JJJ JJJJJ SELECT DBPARTITION… 56 0 753.765 -15.653 -2% 15.345 -1.334 -2% 0 0 0

JJ J JJ J SELECT T2.AGENT_ID … 100 0 1874.321 -195.427 -12% 10.874 -22.337 -12% 0 0 0

JJJJJJJJJJJJJJJ JJJ JJJJJJJJJJJJJ JJJ SELECT T1.AGENT_ID … 345 0 135.987 -120.7083 -95% 0.421 -0.398 -95% 0 0 0

JJJJJJJJJJJ JJJJJJJJJJ J JJJJJJJJJJ SELECT DBPARTITION… 15454 0 1201.787 -55.676 -5% 0.123 -0.059 -5% 0 0 0

JJJJJ J J J

IBM Data Studio ̛www.ibm.com/software/data/studio • FAQs / Tutorials • Downloads • Forum / Blogs • Join the IBM Data Studio user community

Data Studio Book ̛http://bit.ly/dstudiobook