Dynamic Cache Statistics Collector and Visualizer is part of the IBM solution for improving performance of Java™ 2 Platform, Enterprise edition (J2EE) applications running within WebSphere® Application Server. The WebSphere Dynamic Cache (DynaCache) component supports caching of Java servlets, Java Server Pages (JSP), WebSphere command objects, Web services, JSR 168 Portlets, and plain old Java objects. The Dynamic Cache Statistics Collector and Visualizer provides for extensive monitoring of the Dynamic Cache.

Starting with WebSphere Application Server versions 5.0.2.18, 5.1.1.13, 6.0.2.17, and 6.1.0.0, the DynaCache component introduced a JMX Mbean for exposing a number of cache statistics ranging from CacheHits to ObjectsReadFromDisk400K. These statistics provide insight into the state of the cache, health of the cache, performance of the cache, composition of the cache, and efficiency of the cache. These statistics enable an application developer to understand cache usage and to tweak strategies for optimal operation. Effective use of Dynamic Cache Statistics Collector can result in ten times better performance.

The Dynamic Cache Collector and Visualizer enables a Dynamic Cache user to collect and visualize the cache statistics exposed by the DynaCache Mbean. The charts generated by the Visualizer can be customized to show only the relevant and necessary statistics. The line graphs of the cache statistics enable introspection and examination of long-term caching trends, which in turn help tune caching. The Dynamic Cache Visualizer is designed with the principle "A picture says more than 1000 words."

This technology runs on Windows® and Linux®.

How does it work?

The Dynamic Cache Visualizer uses the IBM SVG Chart Generator for generating line graphs for each cache statistic. The SVG Chart Generator is a two-dimensional charting package for SVG (Scalable Vector Graphics). The Dynamic Cache visualizer file reads in a flat file that contains the statistics generated by the DynaCacheStatistics.py python script. The script takes a snapshot of the cache statistics at regular intervals and compiles a flat file of the statistic name value pairs. The resulting file is parsed by the Dynamic Cache Visualizer and charted using options provided in a configuration file called setup.txt. The SVG file generated by the Dynamic Cache Visualizer can be looked at in Internet Explorer or Firefox after installing the SVG plug-in, which is freely available from Adobe. Using the Dynamic Cache Visualizer is therefore a two-step operation:

1. Collect statistics using DynaCacheStatistics.py.
2. Prime the configuration file for the Dynamic Cache Visualizer and run DynaCacheStatistics.bat, which runs the program that generates the chart in the form of a SVG file.

Added April 21, 2009:

An alternate Jython script for collection of cache statistics has been added. DynaCacheStatisticsCSV.py outputs statistics in a CSV-separated format ready to be sucked into your favorite charting package like Microsoft Excel or OpenOffice Spreadsheet. These statistics cannot be visualized with the Dynamic Cache Visualizer.