HP's cluster bomb

The small and beleaguered band of Alpha die-hards suffered another setback yesterday, as HP confirmed rumours that it won't after all be porting its TruCluster technology to HP-UX. That particular project has been underway for years, and was key in persuading the Tru64 hold-outs to move along smartly, please, and get with the great Itanium plan.

HP is showing some sense here, but in a dangerous way. In choosing Veritas' clustering technology for HP-UX, it's acknowledging that TruCluster was taking too long and costing too much to deliver for any commercial advantage that it might have brought with it. Yet many users were already unclear about HP's enterprise server and storage strategy: while this move will cheer the Carlyologists who make their living by double-guessing what on earth's happening now, it leaves more uncertainty on which the competition will feast.

With Tru64, OpenVMS, HP-UX and Linux to manage -- a curious dish of dead duck, spruce goose and penguin -- it's not surprising that HP's existing and potential customers find the results indigestible. HP's faltering enterprise sales, down five percent in the last quarter, are suffering from this. While the company is not alone in having problems combining a strong open-source story with the need to promote its proprietary systems, its task is uniquely complex.

This complexity -- a continuing hangover from the Compaq wedding party -- is deadly. While nobody ever wants to abandon old customers or cut loose from technologies that have been integral to a company's history, such addictions can be costly. When they run counter to the dynamics of the market they appear to the outside world, not always incorrectly, as vacillation and internal conflict. No company can afford this perception when selling enterprise technology. It's hard enough for customers to plan for their future requirements without the added headache of changing vendor roadmaps and a history of incomplete promises.

To change its fortunes, HP must radically simplify its future plans and demonstrate some old-fashioned clarity of vision. This may mean offering its many legacy user bases a more abrupt upgrade path than it or they would like, but it still has the advantage of knowing those people better than do its competitors. That's a solid advantage, independent of technology, which the company must seize.

HP cannot afford to yield to the inevitable in a piecemeal fashion. The bolder it is now in selecting one path forward, one clear story to tell, the sooner it will regain its momentum. We see the future as high performance, highly reliable, highly flexible and highly open: HP's story must reflect this with clarity.