In hot water with only a dodgy database for company

Gene Kligerman must have got himself into a spot of hot water in recent days. He's the IBM database product marketing manager responsible for posting benchmark results on IBM's website.

Obviously not one to cloud the truth, Gene recently reported results that proved Oracle databases run faster on IBM hardware than IBM's own DB2.

Sitting in his Toronto office Gene followed this up with benchmark results showing that IBM database clustering technology lacks the scalability of Oracle's offering.

Larry Ellison is not a man to miss an opportunity to irritate the opposition, and Gene's benchmark results got great mileage at Oracle's OpenWorld event this week.

If you believe the marketing hype Oracle's 9i clustering technology scales, IBM's DB2 doesn't. Oracle 9i runs real applications, DB2 - at least the Unix and Windows version - requires customised applications. And Oracle 9i will cost you 50 to 70 per cent less than IBM's offering.

In fact, IBM got so much attention at Oracle's OpenWorld it should have been charged a marketing fee. But underneath the publicity spiel Oracle omitted one or two minor financial details.

Customers who want to upgrade to 9i - and Oracle estimates 70 per cent of its customer base will upgrade within six months - will more than likely need to replace their hardware. Oracle 9i will not run on processors more than 12 months old.

And that won't be cheap. Oracle reckons an upgrade to 9i's enterprise database on a four processor system will cost $160,000. Customers can expect to pay up to $20,000 for each extra processor as they increase scalability.

It's a situation reminiscent of the Wintel alliance where Microsoft and Intel worked together to support each other's upgrades. Ultimately customers ended up forking out for a new processor to handle the improved features on the new software.

Ellison admits that business had slowed by 15 to 20 per cent in the past year. If one fifth of Oracle customers decided not to spend money on a database that would run on their current mainframe hardware, how many will be willing to spend money on a database that could require new hardware?

In the current conditions 70 per cent uptake seems an optimistic outlook. Analysts warn that beta testing of 9i was rushed, problems could occur, and take up will be slower than expected.

Ellison's retort to that is that 9i is ten times quicker and more reliable than 8i. Until last week Oracle was touting 8i as an amazing database offering. But that's a whole different story.