Message from Ken Kittlitz, received on 11. Oct 2005 at 18:51:
> Mostly because the web server (in this case, a servlet engine) can take
> care of multiple incoming requests and the associated multi-threading
> issues for us automatically. That's basically what the connection server
> does now -- it acts as a "gatekeeper" to the database server, which only
> accepts one connection and command at a time (as you discovered <g>). The
> connection server has had its own bugs in the past, so if we can eliminate
> it from the equation, so much the better.
So would it make sense to connect to the DB server directly? A Perl or
PHP class certainly can handle connection pooling.