Casey Powell, chief executive of high-end, open systems vendor Sequent Computer Inc, is well aware that his company’s strategy is diametrically opposed to 99% of the rest of the server industry. And he is brazenly confident that his is the right approach. I’m not going to use commodity. You don’t put commodity in the data center … that shit breaks, he says. While the majority of the open server industry obeys the commodity/volume mantra, seeking ways to cut down on unique system development by deploying commodity components, and boosting manufacturing volumes in an effort to achieve personal computer-style economies, Sequent is treading a virtual lone path. It is employing specialized, but high performance technology to create high-price, low-volume products – a move which Powell believes is going to turn Sequent into a $1,000m company over the next few years, up from $600m today. It is undoubtedly a bold strategy but, admits Powell, he had little choice. Sequent, he says, did not have the structure or portfolio to compete in the increasingly commodity low-to-midrange server market. Instead, Powell has decided to emphasise Sequent’s strength. Following the official launch of its NUMA-Q 2000 range of shared memory cluster servers last month, Sequent is attempting to move up the value chain into what Powell calls the true ‘alternative mainframe’ space. There is $20bn of mainframe users out there who don’t want it [a mainframe] anymore. I’d be happy with 10% of that, says Powell. He believes that, armed with new NUMA-Q systems, which provide between two and 12 times the power of rival open systems, and a newly reinforced service and support staff – Sequent doubled its sales staff last year – the company can persuade users to migrate from IBM, Unisys and other mainframe systems.
High-end aspirations
Sequent is not the only supplier to have such aspirations, however. Rival Unix vendors including Digital Equipment Corp, Hewlett-Packard Co and Sun Microsystems Inc, as well as a growing band of Windows NT specialists, are addressing the same market. But Powell is insistent that these suppliers, with business structures aligned to the commodity/volume model, do not have the strength of technology or infrastructure to compete. Sun is going to try and take a fillet mignon, cover it in ketchup and wrap it in paper, he says. They only know how to do volume. He also claims that Microsoft’s NT operating system is not yet ready for the data center. We are concentrating on eight-way and above. Eight-way means Unix… it will be five years before NT can be in the eight-way space. Sequent has already succeeded in persuading an impressive bunch of users that its approach is right. In the fourth quarter, before official availability, Sequent shipped 75 NUMA-Q 2000 systems, accounting for a quarter of its revenues, to customers including Boeing Inc, Ford Motor Co and Telkom South Africa. This quarter, says Powell, NUMA-Q will account for half of all sales. The switch to NUMA and a more mainframe-oriented structure has, however, caused a number of problems for Sequent at the lower-end of its business. NUMA-Q 2000 is three to 10 times faster than Sequent’s outgoing Symmetry server range, leaving a gap at the low-end of its product line. In a bid to solve that problem, Powell has signed an agreement with NCR Corp, one of its key open systems rivals, to take NCR’s two to eight-way World-Mark NT servers as an OEM product. But, he says, Sequent will only sell these systems as surround servers to NUMA-Q. I’m big, big, big – I’m after the $5m to $40m space. If a [Sequent] sales guy goes to try and sell a four-way [WorldMark], not an eight-way [NUMA-Q 2000], he hasn’t got a job. By effectively ruling itself out however, Sequent is taking a considerable risk. If the initial burst of NUMA-Q sales proves to be little more than pent-up Symmetry upgrade demand, and if the company’s mainframe alternative strategy is found lacking, Sequent will find itself woefully ill-equipped to compete lower down the power ladder. But Powell, bullish as ever, says he is confident that NUMA-Q will have nothing but a positive effect on business results. We will see nice growth this quarter and a net income recovery in the second quarter, he says. For the year as a whole, he is forecasting 30% revenue growth.
