Re: DB2 Crushes Oracle RAC on TPC-C benchmark

Up to now, I have not commented on the substance of the IBM claims from their website (just defending myself for posting the link in the first place). But now I will make a few substantive observations.

I posted the original TPC benchmark report because it specifically mentioned Oracle RAC and was interesting since it compares performance and cost of 2 completely different architectures (1-64 processor node vs. 16-4 processor Intel nodes). Daniel Morgan complains about apples vs. oranges comparisons, but would anyone run Oracle RAC on a single 64 processor node? By definition, comparing Oracle RAC to a single large DB2 node is always apples vs. oranges.

From the Intel website: "The Intel® Itanium 2 processor, Intel's highest-performing and most reliable server platform, moves you beyond proprietary RISC platforms to help you meet your business-critical computing needs with proven capability and mainframe-class reliability."

http://www.intel.com/products/server/processors/server/itanium2/

According to the Intel website above, the fastest The Intel® Itanium 2 processor now available has a speed of 1.6 GHz. The Oracle benchmark used 1.5 GHz chips. So is anyone claiming that Oracle RAC would have beaten the DB2 benchmark with the slightly faster chips now available from Intel? Also, since Oracle says RAC is scalable, why don't they just increase the number of cheap Intel nodes to surpass the IBM benchmark results?

Obviously, a DB2 vs. Oracle (non-RAC) would be an interesting comparison on the exact same hardware. But that was not the subject of this thread. Received on Sat Jan 29 2005 - 15:33:55 CST