http://www.microsoft.com/sql/evaluation/casestudies/alphalisting.asp,
especially the NASDAQ one.

www.tpc.org

700'000 tpm fast enough?

It's not a case of if your DB server can handle the load, but a case of if
your software is capable of pushing it there. Ask your software vendor if
they can push SQL, Oracle, DB2 or Sybase it it's limits. They will look at
you stupid and not have an answer.

SAP is capable, and I have seen no other software able to do it.
--------------------------------
Mike Epprecht, Microsoft SQL Server MVP
Zurich, Switzerland

IM: mike@epprecht.net

MVP Program: http://www.microsoft.com/mvp

Blog: http://www.msmvps.com/epprecht/

I have personally worked on systems that did over 1 million trans per minute
on SQL Server and they were not the largest by any means.  There are
hundreds of systems out there with a thousand or more users with no problem.
Just look at the links Mike pointed out and browse the SQL Server web site
for lots of white papers and case studies.  But don't get confused when you
talk of clustering.  Clustering is a hardware failover solution only when it
comes to sql server. These systems I am talking about are mainly single
boxes with anywhere from 4 to 64 processors.

Sounds like you have an Oracle bigot on your hands.  His defining the term
'cluster' to match Oracle marketing's definition is a dead giveaway.  As
Mike and Andrew noted, SQL can match/exceed Oracle on any conceivable
database implementation.  They will do it differently and if your boss
defines "different" as failure, you won't have a chance.  Show him the
numbers and if he is reasonable, you will have a chance. If he insists that
clustering is a bunch of cheap brick computers that scale out, then he has
gone over to the dark side.

Geoff N. Hiten
Microsoft SQL Server MVP