Two references worth looking up:

1. "In Search of Clusters", 2nd. Ed., by Gregory F. Pfister. An 
excellent discussion of SMPs, clustering, etc. Available from Amazon 
(or whatever other source you use for technical books).

2. all the papers from the "Locus" project at UCLA, Gerald Popek was 
principal investigator. Locus tried to be a "single-system image" 
system with different hardware in it (vaxen, suns, etc). It was often 
referred to derisively as "slowcus" and the whole thing got 
incorporated (as a company: Locus Computing Corp.), flopped around a 
while, and finally sold to IBM. It became their "transparent 
computing facility" (TCF). I have no idea if anyone is still using 
it; all this stuff happened in the 1980's.

	Erik <fair@clock.org>