web clustering

[MJ Ray]

Jeff Davies <jeff at llandre.freeserve.co.uk> writes:

> I have an idea for very simple web clustering. Anyone who's
> interested can see my notes (and diagram in PNG) at:
> http://www.llandre.freeserve.co.uk/webcluster.htm

This has already been done by using rewriting filters on the output
which search and replace http://yourhostnames with suitable host
names.  You can do this in various ways to gain performance etc.  That
said, you have problems in keeping multiple databases in step and
avoiding single point of failure.

As I'm just learning forming connections to relational databases is a
relatively expensive operation, so I'm not sure SQLRANDOM is a good
way forward.

What has this to do with FSFE, by the way?