this is what is stupid about big servers, if you were a cio and you had a choice which would you purchase
option a) 1 of these @ $4,140,830.00 (72 processers)
option b) 4140 rack mount comps @ $1000 each (4140 processers)
q: what if (a) breaks?
a: you pay a million to fix it?
q: what if (b) breaks?
a: throw the bad machine away
Microsoft clustering services make this soo much of a better option.
this is exactly why sun and ibm fail at what they do.