Distibuted Processing (Re: singularity and exi-GIMPS team)

Aaron Davidson (ajd@ualberta.ca)
Fri, 13 Aug 1999 22:54:05 -0600

"COSM" Is distributed.net's spinoff to create a general purpose distributed computing backbone for the internet:

http://cosm.mithral.com/

I am also writing some small time general purpose distributed computation protocols for some my own projects, seeing as I have access to about 200 fast unix systems and a few hundred fast windows systems scattered around campus.
I may as well make use of them!

Java is in a nice position for distributed work, since any java code can be serialized and sent to any other computer regardless of OS and then executed on the fly.

All of the above mentioned systems are only good for distributed processing, and not true parallel computation. Only a small subset of problems can make use of such systems, although there are enough such problems to make these systems useful. We need much less latency before parallel processing across the net becomes possible.

In the mean time, can anyone think of useful distributed applications?

I can imagine a large scale project using Genetic Algorithms could potentially work. Each node could evaluate various specimens, then report to the central server with the top percentile of tested specimens.

Tierra was more difficult because all of the specimens must be tested against each other in a competetive arena, which does not parallelize well. GA projects that can be tested in isolation would be ideal however.

+-------------------------------------------------------------------------+
| Aaron Davidson <ajd@ualberta.ca> http://ugweb.cs.ualberta.ca/~davidson/ |
+-------------------------------------------------------------------------+