Network Computing is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Pulling Together

To me, one of the more notable things about Apple's announcement of its upcoming Mac OS X Server 10.4 -- a/k/a Tiger Server -- was its emphasis on the inclusion of Xgrid, its clustering software that allows installations to pile up Xserve G5 servers into a supercomputer. This obviously has some good implications for academia, scientific laboratories, and other places where compute-intensive tasks need to be run on a 24/7 basis. It's not often that Apple gets to claim a hardware pricing advantage over other computers, but even a thousand Xserve G5s are way cheaper than the average supercomputer.

The real question for us here, though, is: Does this have any application to the corporate enterprise? Sure. Why not? Understand, businesses are engaged in selling the latest widget to consumers or other businesses, not mapping the human DNA genome. But that does not mean that they don't have some serious computing tasks on their hands. Larger enterprises may find some value in having a clustered server solution handling a welter of operations as simple as back-office applications or as complex as the data warehouse. And one of the nice thing about Xgrid servers is that they can quickly be reprovisioned at the administrative level to take on another task, passing off whatever number crunching they're engaged in to another machine -- even over the Internet. I'm guessing that universities that have followed the Virginia Tech model for clustering Xserves may eventually hook their machines into a wide-area grid to enable a monster supercomputer; the infrastructure is already there. There's no reason that a widely dispersed enterprise couldn't do the same.

The problem for Apple is its old hardware bugaboo -- large businesses mostly don't use their servers because they can get cheaper ones. As Unix-based boxes go, Apple's aren't bargain-bin. And Apple's advantage in high-performance clustering may evaporate, of course. Microsoft has delayed its high-end clustering version of Windows Server 2003 until next year, but it is eventually coming. At that point, Apple has its usual conundrum: Great ideas that work, but some real competition soon in an enterprise arena where they already face a vast Microsoft market share lead. If I was Apple's sales force, I'd get to identifying some likely large-enterprise targets and trying to sell them Xserves quickly, like maybe yesterday.