Senior Vice President of Editorial
Published: 11 Apr 2004
Who thinks their storage network should be stupider?
Right, just that wise guy in the back corner. You can put your hand down now.
Who knows how it should become smarter?
Ah, now we're into some murky waters. Except if you're a major vendor: Many of them have a very definite idea how they will provide you with that intelligence (and the rest are busy getting their stories together). Listen to Brocade, Cisco, IBM, Network Appliance (NetApp), Veritas and a growing list of other vendors, and your storage network will be smarter than Albert Einstein. And more expensive than Howard Stern's FCC fines.
I think the idea of smarter networks is great. What I think is not so great is this headlong race by every vendor to get you to buy their smart widget without any rhyme or reason as to what intelligence goes where and why.
Let's just take the matter of files. NetApp, among others, has a compelling vision of a clustered network storage controller that will sit between the switch and any kind of storage and provide file and block services. IBM has some (but not all) of the same functionality in StorageWorks Virtual Controller. Veritas wants you to buy their volume management and file system software to sit on a smart switch to do some (but not all) of the same things. EMC, just to tick off Veritas, has some pretty cheap software that will do some of what Veritas' stuff does on the same switches. And so on.
And what's the benefit, anyway? Veritas says it will charge exactly the same amount of money for the network versions of its software as for the host ones. Given that the per-port cost of the switch it will run on will more than twice as high, who will bother?
Right now, the evolution of storage intelligence appears to be driven by venture capitalists and storage executives eager to prove to shareholders that "they're not just a hardware company" or "not just a storage company" in hopes that share prices will rise (to the great enrichment of said executives). What do you think are the chances that those two groups of people will get it right and give you what you need?
If you plan the overall architectures of your storage, drive the long-term data center decisions or are just living with what the vendor community produces, get involved in this evolution. However you have to, start talking to your peers about what they want to see where at the network level, and come to your own conclusions. Then, start pushing hard on vendors to produce it. As always, we offer our pages for letters and articles from you on a particular topic, the SearchStorage.com forums for live interaction and the Storage Decisions conference for face-to-face discussion. Use us or use other venues, but please, make your voice known.
I know of at least one instance where storage managers from some large companies had a major impact on SAN/NAS convergence, and I believe you can have that same kind of influence on this whole migration to network intelligence. The transition to network intelligence will take years, but you'll have to live with it even longer.