Oracle OpenWorld: Aftershock!

Larry Ellison talks the talk, but can his New Database Machine walk the walk? Given past events, the experts wonder if Oracle’s next big thing will be all that it’s supposed to be.

Let's be David Tennant. Let's jump in the Tardis and go back in time. It's 1998. It's the CeBIT trade show in Germany and the Oracle CEO is making an industry changing announcement. "This all falls under a simple umbrella. We will compete on total cost of ownership. We can deliver a lower purchase price and higher service," he says. "We are not doing this as a marketing stunt."

He's talking about a little venture called Raw Iron. It was pretty big in its day. It was a push to market and sell specialised server appliances from Hewlett Packard and Dell that ran Oracle databases as native. It was a grand idea, it was heavily promoted and it bombed – really, really bombed. No, really, really, really bombed!!!

Set the co-ordinates Doctor, dematerialise and we fast forward to 2008. The self-same Larry Ellison is on stage at San Francisco's Moscone Centre to announce a couple of database server products, arguing that Oracle is moving into the hardware business. His argument might be seen as remarkably familiar: it's about database bandwidth and throughput and simplicity.

So now Oracle is a manufacturer of Database Machines – in the company of HP, Dell having been quietly dropped somewhere in the web of time – and it's on a collision course with the likes of Netezza and – most obviously – Teradata. Plus ca change, plus ce le meme chose? Maybe. It's certainly a bit of hyperbole (from Larry? No!!!) to claim that his is Oracle's first hardware adventuring. OK, some of the previous efforts are more Ellison than Oracle, but nCube, Pilar Data and the Network Computer are all testimony to interest in the hardware market.

But the difference is this is a pureblood Oracle push and it's got a specific market and specific companies in its sights: Netezza and Teradata. Of the two, Netezza seems the most in tune to the potential threat. No sooner had Ellison completed his on-stage sneer-fest at Netezza than the firm had emailed its rebuttal to journalists and analysts still sitting in front of the Oracle CEO. It was as immaculately timed and perfectly targeted propaganda as you could hope to deliver.

"For years the old, incumbent vendors have tried to bolt together their products to try and tackle the data warehouse appliance market - and every one was a failure," said Jim Baum, president and COO, Netezza. "You just can't slap together existing solutions in clever packaging and expect to deliver much faster performance. The power and simplicity of the data warehouse appliance model lies in integration and design from the ground up. Engineers in the same company, the same building, working to integrate a shared vision - not patch it together with glue and spit."

In contrast, Teradata seemed somewhat taken by surprise "We can't comment on the performance claims of Oracle over Teradata on a product which we haven't competed against to the best of our knowledge." said a Teradata statement the next day. "We are unaware of any customer benchmarks where Oracle has beaten Teradata in head-to-head competition. Oracle's announcement confirms the importance of having a specific analytical database and platform to handle true enterprise data warehousing requirements. Teradata has been focused on this for many years. This is why we have been the market leader in data warehousing and analytics."

The notable difference in the robustness of their reactions might explain why Teradata was identified as the more likely loser in this latest Ellison gambit. That said, Teradata saw a 7% jump in its share price as speculators suspected an imminent takeover bid for Teradata by Oracle arch-enemy SAP, which has been hinting about a possible headline grabbing new acquisition.

JMP Securities analyst Pat Walraven thinks it's entirely possible. Redmonk analyst James Governor has slightly more faith. "Teradata is the cockroach of decision support - you can't kill it, you can only change its owner," he quips. "Oracle has made hardware related pitches before – at the end of the day we're talking about a joint go to market with HP, which is pretty much par for the course in the two company's business models."

AMR Research's Bruce Richardson picked up on the high PR prominence of the Database Machine at Oracle OpenWorld. "The first two full days of Oracle OpenWorld generated very little buzz," he said. "As attendees climbed the stairs or rode the escalator in Moscone North, they couldn’t help but notice the large banner proclaiming that 'The X is Coming on September24'.

"People would see it and ask, 'What is X?'," he noted. "While this could imply a buried treasure for Oracle, I see it as marking a target. The new appliance is clearly aimed at displacing rivals Teradata and Netezza. IBM may be another target, and Microsoft just acquired technology in this market.

"The project with HP is said to signal the start of the next phase of a much tighter relationship between Oracle and HP. HP had been trying to crack the very large database (VLDB) market with its Neoview product, but hasn't gained much traction. Perhaps this relationship with Oracle is meant to bolster those efforts, but what now becomes of Neoview isn’t clear."

"A number of companies are in the competitive crosshairs here - Teradata and Netezza in particular," suggested Ovum's David Mitchell. "Oracle and HP are clearly aiming to make a move into the high-performance data warehouse market, pulling it back towards the standard database and standard hardware markets - and away from the specialist database and data warehouse alliance market. If Oracle and HP together can achieve this then there will be structural changes in the market. This is a tough market for the HP and Oracle pairing to destabilise, but they have clearly done substantial engineering work and have put in place the sales channels and customer support foundations needed to support the products, if they gain traction.

"While the HP Oracle Database Machine represents an advance that addresses the database bandwidth problem, there are other hardware and software elements that will have an impact on the future design of databases. Current database architectures were mostly designed with the premise that disk was slow and memory fast, but the volumes of data were relatively low in comparison with the monstrous volumes of data seen today. Pushing elements of the database workload onto the storage server, as the products announced by HP and Oracle do, recognises that I/O limitations have become more of a constraint to scalability and have been architected to address that limit by moving away from a block I/O model."

But there may be a twist in the tale. "This is a new example of Oracle branching out from its traditional DBMS market," reckoned Gartner's Donald Feinberg. Oracle customers now have four data warehouse options — custom warehouses, reference configurations, optimized warehouses or the database machine. Oracle database licences are transferable among the solutions and customers do not have to migrate or upgrade between any of them. The storage management, database and system management intellectual property are all separate, so migration is no different than a hardware upgrade."

Ironically given the evident new closeness between Oracle and HP, Gartner sees the latter as potential target in all this. "Gartner believes the offering competes directly with HP's Neoview, but HP believes the offerings are complementary and that the Exadata products compete with every other appliance except Neoview," noted Feinberg. "Both HP (Neoview) and now Oracle (Exadata family) have sophisticated database appliances that are optimized for data warehousing. We believe Oracle's scalable hardware platform sets the Oracle data warehouse solution in direct competition with HP's Neoview."

Related stories