For more than a decade, Teradata, NCR’s data warehousing division, has been sitting proudly at the pinnacle of the database market. It has not been quite alone at the summit: many of Oracle and IBM’s biggest customers preferred to apply those vendors’ arguably less sophisticated databases to data warehousing in order to keep costs down. But the conventional view has long been that, assuming an organisation could build a business case for a Teradata engine, they would find it had no equal, particularly when it came to getting (comparatively) rapid answers to complex queries based on vast amounts of data.
This fact still holds true today, although Oracle and IBM continue to snap at Teradata’s heels. What has always set Teradata apart from the rest is its architecture: multiple Intel processors working in parallel, which gives it extraordinary scale. So much so, in fact, that Teradata believes it still has the market for databases containing at least 1 terabyte to itself, and is even moving onwards and upwards to build databases in the petabyte (1,000 terabyte) range.
But it is not just trying to find new peaks to scale. At Teradata’s annual European use conference in Paris, a lot of the talk, admittedly, was of speed and scalability. But the ‘i’ word – integration – seemed more prominent than before. That suggests the company believes it must do a better job linking its technology with front-end tools: and, just as importantly, this should be seamless enough to keep integration costs down. That way, Teradata may even be able to meet growing demand from smaller large enterprises and, perhaps, even medium-sized companies, which are at least beginning to explore the possibility of data warehousing.
It might have its own applications, but Rob Armstrong, the director of technical marketing, says one of the company’s biggest challenges is establishing partnerships with third-party vendors. “We are trying to bring Teradata into a broader environment, so that they integrate and optimise on different platforms,” he says. Indeed, in the last few months, Teradata has announced two significant deals – with SAP and Siebel Systems – and more seem set to follow. In the case of the SAP initiative, engineers are blending Teradata software with SAP’s NetWeaver web services integration framework and other products, while in the Siebel case, the companies will work together to provide enhanced integration and optimisation between Teradata and Siebel’s analytics products. Siebel will also make Teradata one of its supported databases.
This integration activity has been well received by analysts, many of which are concerned about Teradata’s long-term prospects unless it can adapt and broaden its appeal. In the present day, the company goes from strength to strength: it now has more than 700 customers, including Vodafone and Lloyds TSB in the UK, and sales are growing steadily, totalling $1.2 billion in 2003 and jumping another 10% in the first quarter of 2004.
But Armstrong’s insistence in Paris that the company will continue to focus on the global 3,000 – “large companies, with lots of data and big requirements – did not go down well with everyone. “Oracle, IBM and Microsoft will certainly put upward pressure on,” says Gartner analyst Andy Butler. “Teradata is the Rolls Royce of the data warehousing market, but to successfully compete it may need to cascade the offering down to the Robin Reliants out there.”
Its inability to integrate with open source could prove to be another drawback, says Butler. But for now, at least, Teradata seems well placed to benefit from the data explosion.