On Wed, 12 Dec 2001, Eric Smith wrote:

> >> I believe that they have about 6000 Linux servers now. That's what I
> >> call a reasonable sized server farm!
> >
> > And I'll bet they could have done it more cheaply with one mainframe.
> > LOL.
>
> Not if the tasks are compute intensive, which I suspect they are.
> Modern mainframes have only slightly more CPU cycles than high-end
> PC-class servers.

But sorts and searches (DB2, 650 million records, randomly generated) on
my G5 running z/OS on z/VM easily beat the living shit out of the same
operation on my login server (eight processor 700MHz Pentium III Xeon,
Linux).

> Yes, you *can* run thousands of partitions on a z-series or system/390
> machine, but they don't get very many compute cycles each.

But the thing is that with a single mainframe, you'll be able to service
tens of thousands of concurrent requests well.

> Those guys at Google are pretty sharp.  If a mainframe would be a more
> cost-effective solution, they'd probably be doing that.

It all depends on their attitude regarding mainframes.  They have a
totally different "feel" from PC's running Linux.

Peace...  Sridhar