> First of all, thank you for doing this comparison and sharing the
> results!
Sure thing. And thanks for sharing Polygraph. It's a great tool.
> We want PolyMix workloads to accurately model a typical [corporate]
> proxy environment. As your graph illustrates, most actual environments
> differ among themselves. For any given trace, I can find traces that
> show up above or below on your graph.
Right.
Corporate proxy environment makes sense, and I should probably have
guessed that myself. But I don't think it's explicitly stated on the web
site itself. You might want to add it to the workload description (e.g.
http://www.web-polygraph.org/docs/workloads/polymix-4/ ).
> The best thing to do would be to (a) accurately compare PolyMix-4
> results with whatever proxy traces you have access to, (b) analyse the
> causes of the differences, and (c) suggest how PolyMix parameters
> should be changed for Polygraph traces to become "better". For
> example, should we increase the size of large files, increase the
> portion of large files, or change BHR discrimination, etc.? Moreover,
> how does changing one parameter affect the overall shape of the
> distribution?
All I have access to is what is publicly available, so I'm looking at the
Boeing proxy logs.
I should mention that my measurements are based on a slightly modified
configuration. Rather than running Polygraph live on real hardware, I run
it inside the ns-2 simulator. There are many things (such as aborted
connections) which I don't have supported in this environment. But I don't
believe that these changes significantly affect the generated workload.
The other point is that I'm not using the specified phase schedule.
Instead, I run at full load for 8 hours. I take my measurements from the
latter of 1 hour after the working set size is frozen (meaning 5
hours from the start), or when the cache fills.
> Please keep us posted on your findings and make specific
> recommendations on how the workloads should be adjusted.
Will do.
mukesh
-- public key: finger mukesh@cs.cmu.edu fingerprint: BDAB AB7A ADFB 9229 1BD8 45FD BE21 850C E36C D4AA
This archive was generated by hypermail 2b29 : Mon Feb 06 2006 - 12:00:23 MST