You are right. Load test can be complicated because of the various patterns of web applications and usages. The simple tests I mentioned and conducted just give myself some idea about the performance. Given the same set up, some numbers should be comparable and reveal some aspects on web servers' performance. They are by no means formal load tests, but they are helpful to me :-)
"Istvan Albert" <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED] >> >> I will have to install lighttpd or other web servers. > >> If it is a Python web server, it would be nice to extend it by putting >> code >> right into the web server. The performance should be better than FastCGI >> because it removes the cost to send the requests/replies back and forth. > > you'll need to make a distinction between a webserver written in python > (primary purpose to serve data) and a web framework that integrates > python and allows you to generate said data with python > > as for the so called load test that you mention above, those "tests" > are pointless and provide no insight whatsoever to the realistic > behavior of the server ... unless of course all your users are expected > to connect from the same machine while asking for the same 2 byte file > at the maximum speed the system allows them to. > -- http://mail.python.org/mailman/listinfo/python-list
