We've written periodically in the last year about attempts to introduce routing and partitioning into storage area...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
networks (SANs). Logically, it seems that both those capabilities will be as vital to building enterprise-wide storage networks as they have been to large-scale data networks. And we've been encouraged by attempts by Brocade, Cisco and McData to address different aspects of that problem, all the while suggesting that they'll be looking to standardize their approaches in the near future.
So color me surprised when an executive at one of the big three switch manufacturers asked me recently what I could do to convince another vendor to cooperate in advancing switch interoperability. Everybody had been making such kissy-face noises recently that I'd been lulled into thinking they were playing nicely. But in fact, some of the children were pulling wings off flies while others colored within the lines.
As I told the switch exec, I don't buy switches, so there's only so much I can do. And here it is: If you buy switches--especially top-of-the-line ones with future-oriented bells and whistles--make sure you don't get locked-in to one vendor. Right now, most people want to have just one switch vendor, but that could change. You should be the one to determine if it should change, not your switch vendor.
But color me not just surprised, but horrified, when a storage manager told me recently that he considered using two vendors' switches in a limited interoperability situation--connecting two SANs that occasionally needed to swap data--and ran up against a brick wall of a different sort. The two switch vendors swore it would work, but his storage vendor wouldn't certify the combination.
The large storage subsystem vendors are the channel for most of what goes into a SAN? What's wrong with this picture?
For one, pricing isn't transparent. If you're paying $1,000 for an HBA, more than half of that may be going to your vendor so they can test it and agree to troubleshoot. But since the OEM's true cost is hidden, it's not easy to negotiate that part of the deal.
Second, all players aren't equal. Buying storage is a bit like buying a car stereo. All other things being equal, the salesperson is going to encourage you to buy the product that the store makes the best margin on. But all other things are hardly ever equal, and it then becomes a balancing act in how far they can push you without risking your business. Storage vendors have biases, some not so subtle. Getting what you want can be an uphill battle.
In that light, my storage manager friend's complaint makes more sense. Why is it in the interest of storage vendors to qualify any-to-any combinations of SAN components? They have to spend more money to qualify those combinations, yet get a lower average margin. You may want interoperability, but you might not be able to get it, at least officially.
If storage were built to robust standards at every interface point, this would be trivial. Until then, you'll have to push harder for more standardized products, and for storage vendors to earn their margins on SAN components by doing new kinds of testing and support.