enhancement should allow customers to run with smaller, more manageable, receivers with less concern about the checkpoint taken following a receiver roll-over during business hours.

12.3 Test Description and Results

Version 5.3 of WebSphere MQ for iSeries includes several performance enhancements designed to significantly improve queue manager throughput and application response time, as well as improve the overall throughput capacity of MQ Series. Measurements were done in the IBM Rochester laboratory with assistance from IBM Hursley to help show how Version 5.3 compares to Version 5.2 of MQ Series for iSeries.

The workload used for these tests is the standard CSIM workload provided by Hursley to measure performance for all MQ Series platforms. Measurements were done using both client-server and distributed queuing processing. Results of these tests, along with test descriptions, conclusions, recommendations and tips and techniques are available in support pacs at the following URL: http://v06dbl07.hursley.ibm.com/hursley/hiumqweb.nsf/pages/WMQPerformanceTeamHome

From this page, you can select to view all performance support pacs. The most current support pac document at this URL is the “WebSphere MQ for iSeriesV5.3- Performance Evaluations”. This document contains performance highlights for V5.3 of this product, and includes measurement data, performance recommendations, and performance tips and techniques.

12.4 Conclusions, Recommendations and Tips

Following are some basic performance conclusions, recommendations and tips/techniques to consider for WebSphere MQ for iSeries. More details are available in the previously mentioned support pacs.

yMQ V5.3 shows an improvement in peak throughput over MQ V5.2 for persistent and nonpersistent messaging, both in client-server and distributed messaging environments. The peak throughput for persistent messaging improved by 15-20%, while for nonpersistent messaging, the peak increased by about 5-10%.

yTests were also done to determine how many driving applications could be run with a reduced rate of messages per second. The purpose of these tests was not to measure peak throughput, but instead how many of these applications could be running and still achieve response times under 1 second.

Compared to MQ Series V5.2, WebSphere MQ for iSeries V5.3 shows an improvement of 40-70% in the number of client-server applications that can be driven in this manner, and an improvement of about 10% in the number of distributed applications.

yUse of a trusted listener process generally results in a reduction in CPU utilization of 5-10% versus using the standard default listener. In addition, the use of trusted applications can result in reductions in CPU of 15-40%. However, there are other considerations to take into account prior to using a trusted listener or applications. Refer to the “Other Sources of Information” section below to find other references on this subject.

yMQ performance can be sensitive to the amount of memory that is available for use by this product. If you are seeing a significant amount of faulting and paging occurring in the memory pools where

IBM i 6.1 Performance Capabilities Reference - January/April/October 2008

 

© Copyright IBM Corp. 2008

Chapter 12 - MQ Series

175

Page 175
Image 175
Intel AS/400 RISC Server, 170 Servers, 7xx Servers manual Test Description and Results, Conclusions, Recommendations and Tips