in reply to Maintaining a pool of Perl interpreters
You might be able to do the same thing by using apache/mod_perl and your own custom C code. You could write the primary handler in C. If the request meets the requirements that your application has for using the Perl providers then you could perform an internal redirect to the Perl handler, get the results and then return that to the user.
This would, in my mind, simply your code greatly. You would no longer have to worry about handling multiple connections, logging, etc. You would also not have to worry about a pool of persistent Perl interpreters since your pool of apache children would be a pool of Perl interpreters.