Looking back at 2005, storage networking has come a long way. If you're typical of many storage managers, some subtle, yet significant, changes have occurred in your organization. Your SANs may still be confined to the data center, but they're increasingly connected to other locations, be they server rooms in the same building, other data centers or that branch office in Des Moines.
And they're not just replicas of your old DAS stovepipes anymore. Fewer, larger SANs seem to be the order of the day. Along with that comes a diversity of hosts and attached devices. Once there were just a few big Unix application servers and a big tape library or two; now there's Unix, mainframe, Windows, Linux and who knows what. The applications now probably include e-mail and, increasingly, a layer of important but not necessarily revenue-generating applications. OK, maybe they're not on top-tier storage or Fibre Channel either, but that's part of the story. Whether it's IP SAN, a NAS head or full-blown NAS, that's part of today's picture, too.
Add it all up and we've moved from a specialist technology that filled a niche for a few big data hog servers to … well, to the building blocks for a storage network that could conceivably span the enterprise.
See how that snuck up on us? We were once talking about how awful HBA firmware upgrades are, and now we're looking at the possibility that storage networks will be the primary medium to carry data between servers and storage.
When will we really have enterprise storage networks? Who knows? I don't think it's a slam dunk. Data networking was at this point maybe 15 years ago, when there were islands of LANs that IT managers began to stitch together and then extend beyond the building. What were the magic ingredients then? A very active commitment to standards from vendors both major and minor, falling bandwidth prices and, oh yes, the Internet.
What will allow you to merge your ever-larger fabrics into a single, cohesive, enterprise-wide storage network? Surely a firm commitment to standards for many currently quasi-standard or proprietary parts of the network, not the least of which is management software. (Yes, I know about SNIA, but where are the products?) And if 99% of the fiber-optic cable in the U.S. is empty (those are the kind of estimates people throw around), why does it still cost tens of thousands of dollars a month for a link? And is there a killer app, like the Internet, to power this movement?
The answer to that last question might be data protection. Disasters, lawsuits, lost business—it all spells one thing. Businesses can no longer close their eyes and pray that once they write data to disk it will always be there. It'll have to be copied to more places and be retrievable much more quickly. Disaster recovery, business continuance, compliance, litigation discovery—it's all those things.
There might be stronger arguments (getting storage spending under control?) and perhaps more killer apps will come along. Until then, at least, you have the building blocks in place. But don't take your eye off the ultimate goal.
Speaking of next phases, it's with mixed emotions that I announce that this will be my last regular editorial column for Storage. I'm passing on the pen to Rich Castagna, who'll be taking the reins as Editor-in-Chief. Most of you have read many of Rich's insightful cover stories over the last two years, but you may not realize that he's also guided the daily operations of the magazine during this time. Storage will remain a major focus for me at TechTarget, but I think you'll find that the magazine is in good hands with Rich.