HTTP is stateless, on the one side, and there can be any number of (unrelated... isolated...) service processes on the other side. This is what makes HTTP so easy to scale-up. If you need for the processes to share information, you must use something like "memcached" (memory-cache daemon) or some other persistence mechanism. If you happen to be using Perl Goodness, such as Moose, that might just be a "role."
in reply to mod_perl variable persistence across request and/or session