IMS System Update - IMS UG May 2012 Stockholm, Helsinki

1

| IMS

© 2010 IBM Corporation© 2011 IBM Corporation1

IMS Technical Update SeminarStockholm, Sweden & Helsinki, Finland

IMS System Update(More Reasons Why It Makes Sense to Migrate to IMS 12)

Jouko JänttiIMS Worldwide Technical Specialist Team [email protected]

2

| IMS


Topics

IMS Roadmap and Strategy

IMS 11 and 12 Highlights– Just to remind you that the later topics are not all inclusive

Review of selected IMS 11 and 12 enhancements– Minor items you may have missed– Not covering the main line items such as DRD for IMS 10, Open DB for IMS 11 or

Repository for IMS 12– Fancy stuff reserved for Kyle

– A mix of IMS 12 Overview and 2-day IMS 12 workshop in 45-minutes with some IMS 11 and service stream updates

IMS Enterprise Suite V2.1– New version

Thanks to Suzie Wendler and the whole ATS team for the original slides

3

| IMS


IMS RoadmapIMS 7 – went out of support in September 2005

– High Availability Large Database – The IMS partitioning solution– IMS Java – 1st Step

IMS 8 – GA in October 2002 – End of Support September 2008– IMS Java – 2nd Step– New architecture for better Parallel Sysplex operation management – CSL 1st Step

IMS 9 – GA in October 2004 – Withdrawn from Marketing September 2009 – End of Support November 2010

– Online Reorganisation without restrictions for HALDB – 1st Step– Storing XML in IMS Databases– IMS Java – 3rd Step– Distributed JDBC access to IMS Databases

IMS 10 – GA in October 2007 – End of support November 12, 2012– Dynamic resource definition– Operation management enhancements– SOA Connectivity enhancements including IMS async and sync callout capability– Extensive use of the Common Service Layer – CSL 2nd Step

IMS 11 – GA in October 2009– IMS Open Database for IMS database access in z/OS IMSPlex or from distributed environments– Quiesce function to reduce the complexity in establishing a recovery point for a database– Online Reorganisation without restrictions for HALDB – 2nd Step– IMS Connect, OTMA enhancements

IMS 12 – GA in October 2011– FP Secondary indices– IMS repository

Supported Supported VersionsVersionsof IMSof IMS

4

| IMS


The IMS Strategy

Do more with less – Reduce CPU utilization– Work in memory (above the bar)– Remove expired workload

Enhance the Interface– Performance is NOT an option!– Capacity - Supporting workload consolidation– And also availability, serviceability, security

Simplification (management of IMS systems as well as IMS application development)

– More intuitive UIs and interfaces to talk to IMS– More autonomic IMS Tools– Greater agility in changing and controlling the environment

Integration of IMS assets with SOA and other Web solutions– Support of SOA standards– Support of Web 2.0 for lightweight integration

and rapid web application assembly

5

| IMS


Strategy - Do More with Less and Enhance the Interface

25% CPU Reduction IMS Connect – V10

FP 64 bit buffers – V11

Transaction Expiration – V11

Up to 25% CPU Reduction for OLR – V11

zAAP offload for IMS Java, SOAP Gateway, XML converter, IMS TM Resource Adapter

zAAP on zIIP with z/OS 1.11

ZIIP offload for request response processing from CF for CQS managed SQ or resource structures – V12

FF database pools in 31- bit virtual, backed by 64-bit real storage – V12

CICS threadsafe support to eliminate TCB switches – V12

Ability to increase logging speed and bandwidth – V12

Bucket for each new release with MIPS reduction items

6

| IMS


Strategy - Simplification

Challenge: help address the IMS skills availability issues– Use industry available Systems Admin and AD skills

– Solution: more intuitive UIs and interfaces to talk to IMS– Eclipse based tooling for Application Development – SQL support from Java with Open Database – IMS 11– More consistency with z/OS, CICS, DB2, Omegamon for IMS for Operational

Interfaces

Challenge: Planned outages and manual tuning– Solution: Greater agility in changing and controlling the IMS environment

– Online Reorganization – IMS 9– Dynamic Resource Definition – IMS 10 – Database Quiesce – IMS 11– Dynamic allocation of ACBLIB data sets – IMS 11– Dynamic Database Buffer Pools – IMS 12

7

| IMS


Strategy – Integration

IMS application modernization– Solution: renovation versus “rip and replace”

– Connectivity– IMS Integration Suite – IMS TM Resource Adapter– IMS Enterprise Suite SOAP Gateway

– Coexistence with modern programming languages– IMS Java

– Enhancements– Rational tooling: RAA, RDz, etc.

Interoperability with remote web services– Solution:

– IMS asynchronous and Synchronous Callout– Support with both IMS TM Resource Adapter and IMS ES Soap Gateway

8

| IMS


Strategy – Integration …

Develop + Debug +

Test

Analyze

Visualize

COBOL Application

IMS Data

DLI Model UtilityOr IMS ES Explorer

Rational Asset Analyzer (RAA)

Rational Developer for System z (RDz)

Modernize IMS SOA Integration Suite + RDzWeb

Application

9

| IMS


Strategy – Integration …

Open integration to IMS databases– Challenge: direct and simple access to IMS database resources

– Solution: IMS Open DB with ODBM (V11)– Provides an industry standard mechanism (Distributed Relational Database

Architecture -DRDA) that can access IMS databases from z/OS and distributed environments

– Solution: The IMS Universal Drivers– Support a JDBC – SQL call interface to IMS DB

– Metadata generation and testing with the IMS Enterprise Suite V2.1 ExplorerIMS Java

– JMPs and JBPs traditionally run in IMS Java Dependent Regions – Supports Java and Cobol interoperability

– Challenge: access from Cobol to Java within an IMS MPR

– Solution: IMS APARs for V10, V11 and base support for IMS V12– Ability for Cobol to call Java in an IMS MPR

– Supports a persistent JVM in MPP/BMP/IFP regions

10

| IMS


And even Beyond

Solutions that take advantage of OPEN DB– Cognos– Optim Data Studio– Data Source Explorer (with RDz, RAD), – IMS Explorer– Hibernate – SQL Squirrel– … and any product that supports a standard JDBC driver interface

Other database solutions– InfoSphere Classic Replication Server

– IMS Replication– InfoSphere Classic Data Event Publisher (CDC)– InfoSphere Classic Federation Server– InfoSphere Optim Data Growth Solutions for z/OS with Classic Federation for IMS – InfoSphere Guardium

– Data encryption for DB2 and IMS– Change Data Capture for IMS

© 2010 IBM Corporation

| IMS

© 2011 IBM Corporation© 2011 IBM CorporationMay 14, 2012

IMS 11 and 12 Highlights

12

| IMS

© 2010 IBM Corporation© 2011 IBM Corporation12 IMS 11 Overview | May 200912

IMS 11 Highlights

IMS Connect Enhancements

User Exit Interface

Dump Formatter

Syntax Checker & IVP

LSQA Storage Reduction

KBLA

/DIAGNOSE Command

System

Type-2 Query TM Commands

OTMA Enhancements

OTMA Type-2 Commands

Transaction Expiration

Shared Queues AffinityRouting

Transaction Manager

BPE-Based DBRC

Security Override for Non-Production RECON

Unconditional deletion of PRILOG Information

DBRC Migration/Coexistence from IMS 9 and 10

DBRC

IMS Open Database

Database Quiesce

ACBLIB Usability

Database RAS

OLR Performance

Fast Path 64 bit BufferManager

Fast Path Usability

Database Manager

13

| IMS


IMS 12 Highlights

- IMS Repository and Usage for DRD Resources

- IMPORT Command Enhancement

- Logger Enhancements

- Syntax Checker Enhancements

Systems Management

- IMS to IMS TCP/IP Communications

- MSC TCP/IP Support

- OTMA TCP/IP Support

- IMS Connect Type-2 Commands Support

- Additional Connect Enhancements

- OTMA Security Enhancements

- APPC/OTMA Synch SQ

- Enhanced CQS Traceability

Transaction Management and Connectivity

IMS Catalog

Full Function Database - Extended Addressability

Support (EAV) Support- FF Dynamic DB Buffers- FF DB Storage Enhancement

Fast Path- FP Buffer Manager 64 bit

Enhancements- FP DEDB Secondary Index

Enablement- Additional FP Enhancements

DBRC- DBRC Enhancements- Migration/Coexistence

CICS Threadsafe Support

Database Management

IMS Enterprise Suite V2.1IMS ES Soap Gateway IMS Explorer IMS ES Connect APIs for Java and C IMS ES Java Message Service (JMS) APIIMS ES DLIModel Utility Plug-In

© 2010 IBM Corporation

| IMS

© 2011 IBM Corporation© 2011 IBM CorporationMay 14, 2012

Review of selected IMS 11 and 12 enhancements

15

| IMS


IMS 12 DB and DBRC Enhancements

16

| IMS


Extended Address Volume SupportProblem: customers may be running out of z/OS addressable disk storage

– Rapid data growth pushing 4-digit device number limit (65,280 devices)

Solution– DS8000 or DS8700 Storage Subsystem – Extended Address Volumes have more than 65,520 cylinders

– Configured as a 3390 Model A – Up to 262,668 cylinders– Any data set may reside on the first 65,519 cylinders

– Only data sets with EAV support may reside on cyl 65,520 or above– EAV support for VSAM data sets was added in z/OS V1R10 – EAV support for non-VSAM data sets was added in z/OS V1R12

Value– Supports the placement of more data sets on a single volume– Allow users to manage fewer numbers of larger volumes

3390-9

54 GBMax cyls: 65,520

3390-9 3390-93390-3

3 GBMax cyls: 3,339

9 GBMax cyls: 10,017

27 GBMax cyls: 32,760 EAV

cylinder-managedspace

track-managed space

Cylinders Beyond first

65,520Cylinders

First 65,520

Cylinders

3390 - A

17

| IMS


IMS EAV SupportIMS 9, 10 and 11 have EAV Support for VSAM

– Database data sets (KSDS and ESDS)– Fast Path and full function including HALDB – RECONs

– Requires z/OS V1R10 or later

IMS 12 adds EAV support for OSAM and non-VSAM system data sets– IMS OSAM database data sets– Restart Data Set (RDS)– Message Queue blocks data set, Long and Short Message data set– IMS Online Log Data Sets (OLDS), IMS log Write Ahead Data Sets (WADS)– IMS SPOOL data sets– BPE External Trace Data Sets

– Requires z/OS V1R12 or later

IMS 9 APAR PK66704IMS 10 APAR PK72530IMS 11 APAR PK78388

18

| IMS


IMS 12 Full Function Dynamic Database Buffer Pools

Problem: Static definitions limit the ability to tune the environment to address growing workloads

Solution– Provide for dynamic change to an OSAM or VSAM buffer pool without recycling IMS

systems to pick up the change– Type-2 commands are used to add, change, or delete Full Function Database

Buffer Pools– Application activity is internally quiesced

– Increase VSAM buffer pool limit (from 16 to 255)

Value– Eliminates system down time for modifications to buffer pool definitions

– Improves application performance with improved buffer pool specifications

19

| IMS


Full Function Dynamic Database Buffer PoolsDefined in DFSVSMxx, DFSDFxxx and by UPDATE POOL command

– During IMS initialization, buffer pools are created using DFSVSMxx

– New and changed definitions are in the DFSDFxxx proclib member

– UPDATE POOL Commands used to add, change and delete buffer pools

IMS Restart– Committed buffer pool changes are written to Restart Data Set (RDS)

– Emergency Restart will restore buffer pools using RDS

– Normal Restart will initialize buffer pools from DFSVSMxx

QUERY POOL TYPE(DBAS) can be used to query pools

Log Records– UPDATE command changes are logged with x’22’ record

– Log record is for information only– Non-recoverable command

20

| IMS


Full Function DB Storage Enhancement

- DLI PSB pool

- PSB Work pool

- DB Work pool

- DMB pool

- PSB CSA pool

Problem: 31-bit real storage constraints can prevent page-fixing pools that affect performance

Solution– Storage for Full Function database pools now obtained in 31- bit virtual, backed by 64-

bit real storage for the:

–

Value– Could reduce use of 31-bit fixed real frames, relieves 31-bit real storage constraint

– Customers with large database pools who previously could not page fix these pools due to storage constraints should now be able to page fix due to an increase in available real storage

– E.g., Improved performance for managing PSBs in the pool when cast outs occur for large PSBs that are rarely reference and get paged out when the cast out occurs.

21

| IMS


Fast Path 64-Bit Buffer Manager EnhancementsIMS 11 introduced optional Fast Path 64-bit Buffer Manager

– Database buffers above the bar (in 64-bit storage)

– Multiple subpools with different buffer sizes

– IMS automatically determined size and number of buffers

– Buffer pools dynamically expanded

– Defined in DFSDFxxx PROCLIB member:

<SECTION=FASTPATH>FPBP64=Y,FPBP64M=xxxxxxxx

– FPBP64=Y invoked 64-bit buffer manager– FPBP64M= set maximum storage used

IMS 12 enhancements to the Fast Path 64-bit Buffer Manager– User control over initial buffer pool storage

– Dynamic pre-extension and compression of buffer pools

– Additional buffers moved from ECSA to 64-bit storage

– Enhanced QUERY POOL TYPE(FPBP64) command output

22

| IMS


FP 64-bit User Control of Initial Buffer StorageNew parameter to set total buffer pools to 25% of DBBF specification

– Default is FPBP64D=N– Initial allocation about 1M– Number of buffers of each size is determined by IMS based on the number of areas of

each CI size– FPBP64D=Y

– Initial allocation is 25% of DBBF specification, distributed among subpools based on the number of areas of each CI size

– Example:– DBBF=8000; 100 1K CI areas; 200 2K CI areas; 700 4K CI areas

– Results:– 200 1K buffers, 400 2K buffers, 1400 4K buffers

100+200+700 = 1000 areas; 25% of 8000 = 2000 buffers

1K areas: 100 of 1000 areas is 10%; 10% of 2000 = 200



23

| IMS


FP 64-bit Subpool Pre-extension, Compression and Deletion, ECSA to 64-bit…

Subpools are expanded before buffers are required– Avoids wait-for-buffer conditions

– IMS 11 expanded subpools only when a wait for a buffer occurred– IMS 12 expands a subpool when it is almost out of buffers

– Specified with FPBP64E=Y (default)– Disabled with FPBP64E=N

Subpools are compressed when buffers are unused– Frees unused buffers in the subpool

– IMS 11 did not compress pools– Specified with FPBP64C=Y (default)– Disabled with FPBP64C=N

Subpools will be deleted when not used for 24 hours– Subpools may be recreated if CIs of buffer size are used later– Specified and disabled with FPBP64C= parameter

IMS 12 uses 64-bit buffers for FLD calls – ECSA buffers were used by IMS 11

Emergency restart uses 64-bit buffers for SDEPs– ECSA buffers were used by IMS 11– May be turned off with FPBP64SR=N

24

| IMS


Fast Path Data Entry Database (DEDB) Secondary Index EnablementProblem: Lack of FP Secondary Index support in the base productSolution

– Secondary indexes for DEDBs are maintained by IMS– Secondary indexes are full function (HISAM or SHISAM)– Multiple HISAM or SHISAM databases may be used for one index

– Supports very large indexes– One index may be built on different fields in a segment

– e.g. Multiple telephone number fields– IMS does not build secondary indexes

– Tool or user program must be used to create them Value

– Access to DEDB via an alternate key

– Sequential processing via an alternate key or alternate segment type

25

| IMS


DBRC EnhancementsSolution

– Output for /RMLIST command entered through the OM API can exceed the previous 32K limitation

– Output size is restricted by the DBRC private storage available for buffering the output message or OM limitations

– Enhancements to DBRC commands:• CLEANUP.RECON – now includes CA record data• LIST.HISTORY – increased timestamp precision/new data • INIT.CA, INIT.IC, NOTIFY.CA, NOTIFY.IC – VOLLIST parameter now optional

if data sets cataloged• INIT.CAGRP, CHANGE.CAGRP – retention period added to GRPMAX• GENJCL – userkeys increased from 32 to 64 and new %DBTYPE kwd added

– Add user data fields to the DBDS recovery records (IC, RECOV, REORG, and CA)

Value– Improved reliability, availability, maintainability, serviceability, and usability of

DBRC and the RECON data sets

IMS

RECONIMS

IMS

26

| IMS


IMS 12 DBRC MINVERS

IMS 12 MINVERS valid values– ‘10.1', '11.1', and '12.1‘

Upgrade of RECONs– MINVERS(‘9.1') changed to MINVERS(‘10.1')

– MINVERS(‘10.1') remains MINVERS(‘10.1')

– MINVERS('11.1') remains MINVERS('11.1')MINVERS 10.1 or higher forces the use of extended precision timestamps by DBRC

– DBRC commands use timestamps with microsecond valuesMINVERS 11.1 or higher is required for Database QuiesceMINVERS 12.1 is required for XCF use by APPC synchronous conversations and OTMA CM1 (send-then-commit)

IMS

RECONIMS

IMS

27

| IMS


CICS Threadsafe Support

Problem: DBCTL performance

Solution– CICS 4.2 adds support for threadsafe IMS database calls with IMS 12– Eliminates TCB switches for IMS database calls

– Without threadsafe support, IMS call must be done under an IMS TCB– Requires switch from CICS QR TCB to IMS TCB and back to CICS QR TCB– If application is running under an OPEN TCB, it also requires a switch from OPEN TCB

to QR TCB and back from QR TCB to OPEN TCB– With threadsafe support, IMS call may be done under a CICS OPEN TCB

– No TCB switch– CICS has multiple OPEN TCBs

– Multiple DLI calls may be done in parallel under CICS OPEN TCBs– Enhancement applies to both EXEC DLI and CALL DLI– Requires IMS 12 APAR PM31420

Value– Lower CPU use– Increased throughput

28

| IMS


CICS Threadsafe Support …

OPEN TCB

EXEC CICS

process CICS cmd

EXEC SQLprocess DB2 call

EXEC DLI

…

QR TCB IMS TCB

process IMS call

OPEN TCB

EXEC CICS

process CICS cmd

EXEC SQL

process DB2 call

EXEC DLIprocess IMS call

…

switch switch

switchswitch

Threadsafe Support With Threadsafe Support

29

| IMS


IMS 12 Transaction Management and Connectivity Enhancements

30

| IMS


IMS to IMS TCP/IP Messaging

Problem: Inability to send messages from one IMS to another IMS using TCP/IP

Solution 1:– Support for Asynchronous IMS to IMS TCP/IP messaging

– Message is sent with ISRT call using ALT-PCB

IMS1 IMS Connect1 IMS2IMS Connect2

ISRT ALTPCB

DescriptorICON1

OTMA

XCF XCFTCP/IPOTMA

TRANPGM

Value–Leverages the use of TCP/IP between IMS systems

–Eliminates need for customer RYO gateway application to receive output message and send it to the other IMS

31

| IMS


IMS to IMS TCP/IP Messaging …Solution 2:

– Support for MSC communications over TCP/IP– New MSPLINK type (TCPIP)

– Supports operational compatibility with other link types (CTC, MTM, VTAM)– TCP/IP generic name support

– Similar to VTAM Generic Resources

Value– Expect greater bandwidth than VTAM

– Allows installations to standardize on TCP/IP

IMS1 IMS2IMS Connect

IMS Connect

SCI SCITCP/IP

32

| IMS


IMS Connect Type-2 CommandsProblem: IMS Connect command interface does not conform to IMS

standards

Solution– Enhancement to support Type-2 Commands for IMS Connect

– Commands are issued from OM client– TSO SPOC, IMS Control Center, REXX SPOC API, Batch SPOC

– New Type-2 commands for IMS Connect resources– QUERY IMSCON and UPDATE IMSCON– IMS Connect Resources

– Datastore, Link, Port, Alias, Converter, etc.– Can coexist with the previous WTOR and z/OS Modify commands

– All functionality is included in the type-2 commands

Value– IMS and IMS Connect may be controlled from a single point of control

– One command may receive consolidated output from multiple IMS Connects

33

| IMS


Additional IMS Connect Enhancements

Problem: Increased usage of IMS Connect requires additional functionality

Solution– Ability to refresh XML converters without restarting IMS Connect– Provide RACF Userid caching – reduces MIPS– Return actual RACF return codes – more info for security errors– Recorder Trace data capture – new trace points– Commit Mode 0 (CM0) ACK NoWait for RYO clients– Partial read status – new client connection status indicating read/wait– Load modules for IMS-provided exits – no need to assemble/bind

Value– Enable customers to use IMS Connect more easily while providing better performance

and diagnostics

33

34

| IMS


Enhanced OTMA Security

Problem: Ongoing need for OTMA efficiency

Solution– One RACF ACEE for the same userid is created and cached for all of the OTMA clients

– New maximum ACEE aging value of 99,999 seconds

Value– Reduce the system storage for RACF ACEEs while providing better security and

performance

– More efficient usage of storage for caching RACF ACEEs

– Earlier notification of OTMA clients when IMS is terminating

– Higher IMS availability for applications

35

| IMS


Enhanced APPC/OTMA Synchronous Shared Queues

Problem: Requirement for RRS for APPC/OTMA SQ is unacceptable

Solution– Provide option to use MVS Cross Coupling Facility (XCF) for communication instead of

RRS between Shared Queues Front-End and Back-End systems – IMS is the sync point manager rather than RRS – AOS= parameter specifies new options to use XCF– New AOSLOG= parameter to request 6701 log record

– Remove RRS dependency for APPC’s synchronous conversation and OTMA’s Commit Mode 1 (send-then-commit) with sync level of NONE or CONFIRM

Value– Improve performance and simplify the syncpoint process by using XCF

36

| IMS


IMS 10/11 Persistent JVM in MPP, BMP, and IFP

Provides the ability to use Java in IMS Dependent Regions – MPP,BMP,IFP

– Add ENVIRON= and JVMOPMAS= parms to launch JVM– ENVIRON= path to z/OS Java JVM, IMS Java classes– JVMOPMAS= path to IMS Java subroutine

DB2 z/OS PK93123 Java SQL calls via IMS ESAF– RRS is not required

Benefits– Leverage Java code from existing COBOL applications

– Resolves create/destroy JVM performance problem IMS 10 PTF UK58284

APAR PK82214

IMS 11 PTF UK58285

APAR PM00360

37

| IMS


IMS TM Resource Adapter Version 12 New Features

Support for WebSphereApplication Server Version 8 and its resource workload routing function

– new resource workload routing function that offers failover and subsequent failbackSupport for multiple data stores per IMS activation specification for callout messages

– This enhancement enables a single message-driven bean (MDB) to pull callout messages from more than one IMS data store.

Data store connection failure recovery for callout messages– When the backend IMS is not available, in addition to the attempt to reconnect to IMS Connect, the

resource adapter will also attempt to reconnect to the IMS data store.

Support for IMS V12 OTMA DFS2082 messages for commit-then-send CM0 transactions

– if the IMS application does not reply to the IOPCB or complete a message switch to another transaction, OTMA issues a DFS2082 message to the client to indicate the transaction terminated with no reply.

– This enhancement enables you to convert send-then-commit (CM1) transactions into CM0 transactions without having to modify your applications.

Support for IMS Version 12 RACF return codes– In IMS V12, if a RACF security failure occurs, IMS Connect includes a 2-byte return code from the

RACF RACROUTE REQUEST=VERIFY command.

38

| IMS


IMS TM Resource Adapter Transaction Expiration Support

Invoke message level transaction expiration – New IMSInteractionSpec boolean property added along with getter and setter

– transExpiration– getTransExpiration()– setTransExpiration()

Execution times out OTMA discard or dequeue the transactionBenefit

– reduce the CPU cycles for processing the unwanted transactions

Apar IMS 10 PM01727 IMS 11 PM01822

39

| IMS


IMS 12 Systems Management Enhancements

40

| IMS


System Enhancements…Dynamic resource definition (DRD) enhancementszAAP/zIIP times in accounting log records, zIIP support for some CQS processingMember online change NAMEONLY optionEAV support for non-VSAM data sets (including OSAM database data sets)IMS logging enhancements

– Extended-format data set support for OLDS/SLDS– IMS log buffers in 64-bit virtual storage– Simplified WADS management

System pools in 64-bit real storageSysplex Serial Program Management (SSPM) DisablingSPE for log archiving of non-recoverable databasesCommand enhancementsSystem usability enhancementsRAS enhancements

– CQS traceability enhancements– IMS Dump Formatter enhancements– Reduce IMS Module Aliases in RESLIB– End-Of-Memory / End-Of-Task (EOT) Tracing Facility– IMS 11 SPE BPE Trace Selective Initialization– /DIAGNOSE SNAP command enhancements

MIPS reduction enhancements

41

| IMS


zAAP/zIIP Times in Accounting Log Records and zIIP support for some CQS processing

zAAP/zIIPtime field is added to the x’07’, x’0A07’and x’56FA’log records– X’07’ program termination– X’0A07’ CPIC program termination– X’56FA’ optional transaction level statistics record

CPU time field is changed to include only the standard CP (central processor) time, not zAAP/zIIP time

– Sum of CP and zAAP/zIIP times is the total CPU time– Benefits

– Users can distinguish between CP and zAAP/zIIP times– Could be used for accounting or charge out purposes– Significant for software licensing

– Most significant for JMP and JBP regions

Request response processing for authorized CQS clients is eligible to run on a zIIP

– When the IMS control region is running with IMS Shared Message Queues or Shared IMS Fast Path Message Queues enabled

– When the IMS Resource Manager address space is using a resource structure

42

| IMS


Logger Enhancements

Problem: Logging bottlenecks that affect performance

Solution – Extended Format support for OLDS and SLDS

– Allows OLDS and SLDS to be striped– OLDS buffers may be in 64-bit virtual storage

– WADS management changed to be more efficient– Track groups no longer used– WADS written in simple wrap-around fashion

Value– Increased logging speed and bandwidth

– Reduced ECSA usage

– Smaller WADS

– Better use of storage system cache by WADS

43

| IMS


Striping of OLDS and SLDS

Use of extended format data sets allows striping– Striping allows multiple concurrent I/Os for sequential processing

– Data set is spread across multiple volumes– Increased logging IMS 12 allows OLDS and SLDS to be defined as extended format

data setsExtended format is specified with data set type of ‘EXT’JCL allocation requires DATACLAS and STORCLAS parameters on DD statement

IMS 12 allows OLDS and SLDS to be defined as extended format data sets – Extended format is specified with data set type of ‘EXT’

– JCL allocation requires DATACLAS and STORCLAS parameters on DD statement– Striping is invoked for extended format data sets when the storage class has Sustained

Data Rate (SDR) value of 5 or higher

44

| IMS


OLDS Buffers Specifications

OLDS buffers above 2-gigabyte boundary – Specified with BUFSTOR=64 on OLDSDEF statement in DFSVSMxx

– Puts buffers in 64-bit virtual storage– Requires that block size is 4K multiple

– OLDS must be extended format– Frees substantial amount of ECSA

OLDS block size – Optionally specified with BLKSZ parameter on OLDSDEF statement in DFSVSMxx

member– Previously, it was set only from the data set characteristics

– BLKSZ= is recommendedIf BLKSZ and BUFSTOR=64 are both specified

– IMS will round up number of buffers (BUFNO=) to a megabyte boundary– 64-bit storage is acquired in megabyte increments– Maximizes number of buffers in acquired storage

45

| IMS


Migration to Buffers above the 2G Bar

All OLDS must be extended format for the buffers to be above the 2G bar

If Logger Exit (DFSFLGX0) or RSR Log Filter Exit (DFSFTFX0) is used– Ensure they are capable of handling buffers above 2G

– The IMS supplied DFSFTFX0 exit is capable

BUFSTOR=64 must be specified on the OLDSDEF statement

OLDS block size is multiple of 4K

IMS must be restarted– Restart may be cold or warm

46

| IMS


WADS Management

The concept of WADS track groups is not used by IMS 12– WADS should be sized to provide enough space for any OLDS buffers not yet written at

any time plus one track

– WADS use 4K block size

– WADS writes are changed from previous IMS versions– WADS writes are sequential– WADS written in wrap-around fashion

Performance– WADS sequential writes generally improve usage of cache in storage systems

– WADS should be kept in cache in storage subsystem

47

| IMS


SPE for Log Archiving of Non-recoverable Databases

Archive (DFSUARC0) is enhanced to write “undo”records for non-recoverable full function databases– Previously, “undo” records were not archived

– New control statement option not to archive “undo” records– Required to write these records as was done previously

Non-recoverable database logging:– “After image” log records are not written for non-recoverable databases

– “Undo” log records are written to OLDS for non-recoverable full function databases

– This allows uncommitted updates to be backed out– “Undo” log records are not archived from OLDS before this enhancement

– They are archived with this enhancement unless overridden by control statement

48

| IMS


SPE for Log Archiving of Non-recoverable Databases

Problem addressed by this enhancement– Online backout using archived log stops when it encounters a “missing” log record

– Backout is incomplete, other database updates may not be backed out and will be stopped

– Batch backout using archived log does not backout updates to non- recoverable database but sets no flags in DBRC to prevent its use

Control statement not to archive non-recoverable database log records– Option applies to specified databases

SLDS CMPRSNR((dbd1,ddn1)(dbd2,ddn2),...|ALL)

Benefit– Eliminates potential database integrity and operational problems

SPEs: IMS 10 PM18093; IMS 11 PM19363; IMS 12 PM54945

49

| IMS


Command Enhancements

Enhancements are focused on type-2 commands for the Operations Manager (OM) environment for supporting new functions and improving manageability

Enhancements to existing commands and new commands– CQS trace command enhancements

– DBRC command enhancements

– Dynamic database buffer pool command enhancements

– Dynamic resource definition (DRD) command enhancements

– Fast Path secondary index command enhancements

– HALDB command enhancements

– IMS Connect command enhancements

– MSC command enhancements

– IMS repository function command enhancements

– OTMA command enhancements

50

| IMS


IMS 11 SPE BPE Trace Selective Initialization

IMS Connect Recorder Trace facility delivered in IMS V11 uses a BPE trace table

Inadvertent activation of BPE tracing is possible

– If BPE trace initialization encounters a generic type (*) on the TRCLEV statement in the HWS BPE configuration member:

TRCLEV=(*,HIGH,HWS)

– Performance problems could result from running Recorder Trace inadvertently

Different than activating the original IMS Connect Recorder TraceExplicit IMS Connect “OPEN” command

51

| IMS


IMS 11 SPE BPE Trace Selective Initialization

Creates a new BPE trace selective initialization feature that allows a trace table to be defined as requiring explicit commandactivation

– IMS Connect Recorder Trace, RCTR, trace table definition updated to enable the new explicit commands feature

– No inadvertent starting of IMS Connect Recorder Trace– Configure the RCTR trace table differently:

TRCLEV=(RCTR,MEDIUM,HWS) vs. TRCLEV=(*,HIGH,HWS)

Adds explicit command logic to BPE trace table update command– Trace table types requiring explicit commands are processed only when the NAME

parameter explicitly specifies the trace table name– UPDATE TRACETABLE NAME(RCTR) OWNER(HWS) LEVEL(MEDIUM)– UPDATE TRACETABLE NAME(RCTR) OWNER(HWS) EXTERNAL(NO)

Benefits– No inadvertent activation of BPE supported IMS Connect Recorder Trace

IMS 11 APAR PK98125 / PTF UK53814

52

| IMS


IMS Enterprise Suite V 2.1…

IMS Enterprise Suite Connect APIs for C/C++ and Java – Simplifies application development for stand-alone, user-written IMS Connect clients

IMS Enterprise Suite SOAP Gateway– Provides a solution for integrating IMS assets in an SOA environment that supports:

– Exposing an IMS application as a web service– Calling out to a web service from an IMS application– Emitting business event data to an event processing engine

IMS Enterprise Suite Java Message Service (JMS) API – Supports synchronous callout request to external services from within an IMS Java message processing

(JMP) or Java batch processing (JBP) application

IMS Enterprise Suite DLIModel utility plug-in (eclipse-based) – Translates IMS source files into reliable, application-independent metadata that can be used for Java

application development

IMS Enterprise Suite Explorer for Development – New capability that enhances the functionality previously provided only in the DLIModel utility plug-in

53

| IMS


Summary

Modernize Application Interoperation/Integration by – Standardizing tools/interfaces to speed deployment

Streamline Installation/Management by– Simplifying Interfaces to ease operations– Heightening availability and increasing Productivity

Enable Efficient Growth by– Alleviating bottlenecks– Reducing costs – Optimizing performance and resilience

IMS continues to invest in innovation and

54

| IMS


IMS Technical Update SeminarStockholm, Sweden & Helsinki, Finland

IMS System Update - IMS UG May 2012 Stockholm, Helsinki

Technology

Transcript of IMS System Update - IMS UG May 2012 Stockholm, Helsinki