Performance analysis questions

Alan Batie (batie@ibeam.jf.intel.com)
Wed, 11 May 1994 15:07:52 -0700 (PDT)


I'm about to do some server/os/platform benchmarking, and the following
issues come to mind:

1. What makes a "typical" URL mix? One thing I would like to do is ask
webmasters running busy servers to send me copies of say a Monday's log
file (feel free to remove client names/addresses if you like), especially
if you feel your server can't handle the load it's getting.

2. Initial indications are that process startup time is the dominant
factor. This suggests that a benchmark consisting of lots of little
documents will be most telling; I expect that big documents will be
bandwidth limited and not dependent on server performance (unless it's
got a direct T3 interface).

3. How should scripts be handled? Obviously, a script can be arbitrarily
complex, but are there some common things that should be analyzed (for
example a WAIS search, but then that may be more appropriately handled
in an entirely separate analysis)? I'm inclined to ignore them for now.

Any comments and suggestions are welcome, and I do plan to make the test
suite available when finished, as well as a result summary. Thanks!

-- 
Alan Batie                     ------
batie@aahz.jf.intel.com        \    /        X.400: same day delivery
+1 503-696-8844 (voice)         \  /         ...in a nanosecond world.
+1 503-696-6388 (fax)            \/