mod_snake is much more than the CGI accelerator described by the previous post. It provides complete access to the Apache internals.
mod_snake is more low-level than mod_python. It's possible to build the mod_python user model with mod_snake, but not vice versa.
Hmmm... I'll hazard a guess.
mod_snake, from what I gathered while experimenting with it (for a little while, I confess, because it seemed to me a bit limited in scope), seems to be a Python CGI accelerator, it keeps the Python interpreter at hand, so when it needs to run a Python CGI, it just passes the script to it's interpreter along with all relevant CGI information, and lets it rip. Also, it has the added convenience of not having to reload the script every time it's run, only on it's first invocation or after a modification (date changes between *.py and *.pyc).
And, on the other hand, mod_python is a handler, which can interface directly with Apache and it's internal module API; so, you can write Python extensions to Apache, use Python to handle many chores like authentication, pre-/post- connection processing, URL translation, etc... it depends on what part of the HTTP handshake you want to handle, and write a Python handler for that chore.
So, it seems to me that mod_python is much more flexible than mod_snake, but with the added disadvantage of being more complicated to develop for. Check them both out, to see which benefits you more.
What are the differences between this and mod_snake? two
parallel projects trying to acquire the same functionality?
A generic synchronous replication engine for transactional applications.
A flexible I/O tester/benchmarker.