Hi all.

I'm wondering if it would make sense to have all
required server processes (validator, splitter,
scheduler, feeder, etc..) running on each machine,
and split up users and WUs up front.

e.g:
Server1 runs all processes, but handles only
the fraction of the users and the fraction of the
WUs, that were assigned to it.
Server2 handles a different fraction of WUs and users,
and so on.

Server1 users would only get WUs from server1 and send
their results there to be validates against other
results of Server1 users.
There would be no more need to compare everything with
everybody and generate a huge database load this way.


Each server sees only so much users and WUs
as it can handle.


If you want to allow more users, just install another
(identical) server.

There would to be some kind of front end server to
split up the work, but it wouldn't have too much to do,
since it only needs to know where each user and each
WU are residing.

Regards Hans