Polygraph Beta Release. By this date, Polygraph is able to successfully complete a PolyMix-3 run with a proxy cache. Any known bugs should be fixed by this time. Polyteam will distribute documentation that is sufficient to run a PolyMix-3 test. This is the participant's cue to take the software and begin running tests in their own labs. During the next four weeks, participants must provide feedback on the code and documentation.

Polygraph version 2.5.0 is available. In our lab tests, the code is able to complete a PolyMix-3 test using a proxy. Please start using the code to test your products. While no significant bugs are known, we are sure there are some that we have not triggered in our environment. Your input is crucial in making this new Polygraph branch stable.

We had enough resources to commit support for the following features.

• A better popularity model with ``hot spots''
• Integrated fill and performance tests (we still may need better ways of specifying the working set size!)
• ``Reload'' requests (using Pragma: no-cache and Cache-Control: no-cache fields)
• IMS requests and a better Object Life Cycle model
• Better global configuration awareness among robots and servers
• Better constant working set size (WSS) support (this improvement was necessary for a better popularity model)

The PolyMix-3 workload specification was adjusted to (a) meet implementation requirements and (b) adjust parameters from the previous release. Both polymix-3.pg and contents.pg files were updated.

We continue to seek your input on workload parameters. PolyMix-3 introduces several new features that probably require tuning. Please be specific and constructive in your feedback. If you believe that a certain parameter setting is wrong, explain what exactly is not working for you and propose a fix/improvement if at all possible. For example, saying that ``memory hit ratio is too high/low'' is not very specific or constructive. To improve the workload, we would need reports like ``memory hit ratio is X% during the top2 phase while we observe Y% on an average live cache; hot set size should be changed to Z which, in our tests, results in memory hit ratio being Y%.''

2.3 TODO list

Besides addressing your bug reports, Polyteam will continue to work on the following items:

• Improved run automation tools; make it easier to start, stop, and monitor a test
• Improved report generation tools
• Documentation
• Run-time verification of how much cachable content was actually cached
• Improved (more sophisticated) downtime test
• Portability fixes, including native Win2K and HP-UX support

The above items should not have significant impact on the released code.

Please make sure you are using the most recent version of the code and workload and watch Polygraph mailing list for updates.

A beta version of the ``howto'' document is available. Please let us know if something is not clear or ambiguous. General familiarity with Polygraph and the ``howto'' page should be sufficient to execute a PolyMix-3 test and produce a report.

We are in the process of synchronizing Polygraph documentation with the latest release. A complete Web site redesign is also on the way.

If you cannot find necessary information, please do not hesitate to ask specific questions!