POST OF THE DAY
Oracle Corp.
ORCL's Latest DB Losses Continue

Format for Printing

Format for printing

Request Reprints

Reuse/Reprint

By robert1050
December 27, 2000

Posts selected for this feature rarely stand alone. They are usually a part of an ongoing thread, and are out of context when presented here. The material should be read in that light. How are these posts selected? Click here to find out and nominate a post yourself!

For all the non-technical Fools out there let me, just one more time, explain what you're seeing. Some months back Microsoft re-wrote the TPC-C [Transaction Processing Performance Council] so you could run the test over multiple databases and all at the same time. Neat trick. Worked so well that IBM did the same thing; thus far, Oracle has refused to follow because doing so would be antithetical to the message Ellison and Co. are trying to get across. But first, more detail-

MS ran the TPC-C against 12 systems running 12 separate databases. Data was physically partitioned across each system. So, your orders were split across multiple systems as are debits, credits, etc.

Oracle says this is bad because:

1. It costs more to manage MORE systems than FEWER. Need fewer people to worry about care and feeding of a single database; as opposed to managing care, feeding and backup of twelve databases.

2. Twelve systems each running their own database each are less reliable than an Oracle Cluster running a single database. If one or more MS servers goes down you lose ALL access to the data being handled by that server. Which part of your orders are you willing to give up, availability-wise, at quarter-close?

Oracle further notes that SAP/Seibel/Peoplesoft are incapable of being run across separate systems as MS proposes. Same goes for Oracles' App software too. So the TPC-C numbers can be viewed as a one-trick pony. Nobody is building Enterprise software this way and the economics of running this kind of configuration coupled with the inherent reliability issues make it unlikely anybody ever will.

Oracles' cluster solution is to run multiple servers against a SINGLE database (all servers share a single drive-array). If any given server (or more than one-) goes down the ENTIRE database is still accessible. You can still close your books. And you have a single database to backup and maintain.

Speaking of tpg.org go check out http://www.tpc.org/new_result/tpcc_perf_results.asp?resulttype=noncluster and see that Oracle holds 8 of the top ten results for TPC-C run against a single database. Cheaper and more reliable. Will the 9i database, to be released first half of '01, show well even against the MS benchmark-special? Stay tuned ;-).

Go to www.oracle.com/ebusiness and watch Ellison explain the MS TPC numbers at this years Comdex. He also provides a demo if you need further assurance.

Well. Know also that in the last quarter Oracles' database revenue grew 26% (ignoring the currency hit) OVER the previous years (record-breaking at the time) quarter. If Oracle is losing market-share to MS and IBM while still growing at 26% the other two companies must be growing at gangbuster rates. But that's not what I've been reading over the last couple weeks.

And finally, I absolutely refuse to believe that Analysts and Mutual Fund managers either understand the intricacies of database benchmarks or make buy/sell decisions based on same.

Have you noticed there have NOT been any press articles documenting how multiple MS/IBM database customers have beat Oracle out of a million dollars (The Challenge) because Oracle couldn't triple performance? That's a real-world indicator. You don't think MS would forgo the opportunity to make Larry write a check? If they could?

For all Fools I would suggest you consider spending time worrying about ORCLs PE (a la Corplaw) and growth prospects going forward. That's what I think about- And I'd be more worried IF the entire NASDAQ wasn't down 50%; but it's not like the market is up 14% and Oracle is down 40% at the same time. Near as I can tell the market is down close to 50% and Oracle is down about 35% split-adjusted.

I believe the market will come back. If 'buying quality' means something then Oracle is well positioned, though still expensive. Oracle will make a comeback. The question is 'how soon?'

Oracle is in no way working to leverage or cooperate with the MS SS2K database. There's no need to.

Happy holidays to all.

Robert