This article can also be found in the Premium Editorial Download "Storage magazine: Storage products of the year 2003."
Download it now to read this article plus other related content.
Am I the only one who finds it odd that just as the separation of storage and servers has become mainstream, some of the biggest computer companies are lumping them back together in the same business organization?
Hewlett-Packard (HP) not only completed the hat trick started by IBM and Sun of putting servers and storage in the same group, but it even put both in its services organization.
What does this mean for storage managers? My worry is that we're seeing the start of another round of competition to lock you into one or another architecture. The idea of a single supplier for the data center died about 10 years ago, and there never was a single supplier for the whole distributed environment. Are HP et al. banking on the combination of blade servers and storage area networks (SANs) to reel in some huge sales? Integrating two major technologies--both of which are moving targets--could prove to be daunting. So, there's some logic to the idea that you're better off with both sides of the coin coming from the same mint.
Some logic, maybe, but not enough to reverse decades of experience that suggests that one supplier is too few and many suppliers are, well, too many. Whether two or three or more is your ideal, one probably isn't. And anyway, aren't blade servers supposed to be a commodity technology? And don't we hope SANs will become more like that, too? So, in the long run, I'm just not sure there's much to this whole story besides the vaunted single invoice.
Here's another thing I'd worry about: Have the major computer vendors decided that they can't compete on the basis of the best products, so they're going to compete as the easiest to buy? Is this the beginning of a retreat from product development efforts as the storage function at the large vendors becomes more and more of a marketing function?
I find this trend slightly unnerving. It's not that I worry that there won't be good storage technology available in coming years. The dedicated storage vendors will be happy to continue to push the state of the art. After all, EMC got its start back in the day by being a step ahead and a dollar cheaper than IBM and that's a model that will work for storage-only vendors.
Which brings me to another worry: Storage vendors such as EMC and Veritas are taking the bait and trying to establish a beachhead in systems management software, something of a foreign country for them. Better that they deliver more and better storage software. No one I know thinks they've finished that job.
There's potential for a nightmare if you can't be assured that any blade server complex can talk to any and all storage with equal facility. Integration and services are certainly needed in the data center, but not at any cost. So, before you sign on the dotted line for someone's "data center architecture," make sure that you and your brethren on the server side can interoperate in any way you see fit. As long as that's true, whether or not you buy your storage as a bundle with servers and/or services will be a decision you can make when it suits you--not one that's been made for you.
This was first published in January 2004