Implementing and Administering WebSphere Business Integration Server V4.2.2

380 Implementing and Administering WebSphere Business Integration Server V4.2.2

11.1 Introduction

As with any Business Process Integration project, several major software

components work together to comprise the complete business integration

processing environment. Clearly, a performance bottleneck in one component is

likely to affect system throughput as a whole. Although overall system throughput

and response time remain the primary measurements of interest, site

administration should possess some level of confidence that all of the major

component parts are functioning at acceptable levels of performance.

In a situation in which end-user performance does not meet customer objectives,

it is usually necessary to perform system-level monitoring in order to isolate the

cause of the performance issue. The tools and techniques that are used for

system-level monitoring differ from platform to platform. To find additional

information about system-level monitoring, consult the system documentation

and the manufacturer's Web site.

As an example, the InterChange Server supports three different database

managers. The documentation for each of these database products contains a

wealth of information regarding performance, capacity planning, and

configuration. This documentation would likely offer the best guidance for

performance considerations in a variety of operational environments. Assuming

that all of these issues have been addressed from the perspective of the actual

product, additional levels of performance implications are introduced at the

interface between these products and the InterChange Server.

Several configuration parameters are available to the InterChange Server

administrator. This chapter identifies specific parameters that have been

observed to affect performance, but not

all available configuration parameters of

the InterChange Server. For a complete list of configuration parameters and

possible settings, see Appendix A of WebSphere InterChange Server V4.2.2

System Installation Guide, which can be downloaded from:

http://www-306.ibm.com/software/integration/wbiserver/ics/library/infocenter/

11.2 General performance checklist

This section begins with a tuning checklist that enumerates the major

components and their associated tuning concepts. The subsections that follow

address each in more detail, describing performance-related design concepts

and discussing the tuning parameters and their suggested settings (where

appropriate), as well as ways to determine potential settings for a particular

configuration. Although there is no guarantee that stepping through this checklist

Chapter 11. Tuning a WebSphere Business Integration Server infrastructure 381

provides immediate, acceptable performance, it is likely that degraded

performance can be expected if any of these parameters is set incorrectly.

The checklist includes:

򐂰 InterChange Server

– Configure threads in collaborations and in adapter controllers

– Use caches for maps and collaborations (instance reuse)

– Configure threading for CORBA / IIOP

– Configure database connection pools

– Setting flow control queue sizes

– Turn off component tracing

– Turn off event sequencing for server access interface

򐂰 WebSphere Business Integration Adapters

– Configure polling frequency and quantity

򐂰 Database (general)

– Place database tablespaces on fast disk subsystem

– Size database cross-referencing tables correctly

– Place logs on separate device from tablespaces

򐂰 Database (DB2-specific)

– Maintain current indexes on tables

– Update catalog statistics

– Set buffer pool size correctly

򐂰 Database (Oracle-specific)

– Set buffer, block, and shared pool area sizes correctly

– Set processes, Open_Cursors, and IO_Slaves

– Use dedicated connection if possible

– Query optimization

򐂰 WebSphere MQ

– Place MQ logs on fast disk subsystem

– Monitor message queue depth

– Configure MQ log files and buffer pages

򐂰 Java Virtual Machine

– Set the heap and nursery size to handle garbage collection efficiently

– Set AIX® threading parameters

– Use HotSpot Server instead of client

– Set thread stack size if using many threads

– Reduce or increase heap size if java.lang.OutofMemory occurs