Alright. Pass 1 is done. Here is a link to it. The server is in Python. Clients in PHP and Python are provided. It follows this design document. On a quad Opteron, it gets about 600 short messages a second. It isn’t threaded. Next step on this project is to redo this in Erlang. And then maybe C++ for the heck of it.
This uses WSGI, specifically the wsgi reference server, so in theory it shouldn’t be hard to adopt to other wsgi servers, like mod_wsgi. However, beware of problems of thread safety. Also beware that wsgi servers that use multiple processes will require some sort of external data store instead of process local memory.
(edit: Download link was moved to github, design doc link was also changed)