Industrial-Strength Analysis
By Phil Stevens
It's imperative for savvy site administrators to keep up-to-date on the vital statistics of their online enterprises. Ordinary Web-server access logs contain a wealth of information, but sifting through them is a task unto itself.
There are many ways to mine your HTTP log data to uncover facts about your operation. If you run a small to medium-sized site, you can write and run simple scripts on the log files, or pipe the information into a database. You can accumulate an entire suite of routines and output them to tables and graphs, which can scale up to meet the needs of an enterprise site. But if you need to get a large, complex set of Web ventures moving and shaking right now, your timeline may not permit much development. This is where an already proven statistics program can help you.
IBM has made WebSphere Site Analyzer v3.5, originally part of the WebSphere business-application platform, available as a separate component. Certain Site Analyzer features are explicitly integrated into the WebSphere Commerce Suite, but the program can function with any standard Web server that generates logs in combined, W3C extended, or NCSA common log formats, and uses cookies to track user sessions. Site Analyzer is built on IBM's mainframe-capable DB2 database for major scalability, and its client-server architecture supports distributed computing resources.
