Published: 12 Jul 2007
Some of the least-sexy storage tools just might be the most useful ones for coping with power problems.
History is littered with products and technologies that were "nice to have," but not "need to have." Because of all the fires we fight daily in our storage shops, we just don't have the time or energy to deal with "nice to haves"--we have too many "need to haves" on our plate. As Bob Brennan at Iron Mountain once told me, "You don't paint the house when it's on fire."
It's because of this phenomenon that technologies languish on the sidelines, even though it seems as if they'll provide obvious and immediate benefits. Storage resource management (SRM) is a classic example. What could possibly be bad about a tool that provides visibility and insight into the deep, dark recesses of your storage? SRM turns the lights on. Who wouldn't want to know that they have 87 copies of the same illicit pornography sitting on their Tier-1 storage? Would you buy new capacity if you knew you had tons of space you weren't even using?
Boring "vision" tools that provide insight and report on things like data types, asset utilization, mission success or failure, root-cause analysis, change management and so on, seem as necessary to me as ensuring that the system you run is getting power from the wall. But somehow these things haven't managed to make it onto the sex-appeal list. It seems the only time we decide to "turn on the lights" is after a big problem. We buy backup reporting tools after we couldn't recover something important that hadn't been backed up successfully in eight years. We buy change management tools only after we spend four days trying to figure out why our production system is running 90% slower even though someone else swears the only thing added over the weekend was a new laser printer.
It's human nature why we do some of the stuff we do. It can be embarrassing letting everyone know your bulletproof security schema has a hole the size of Guatemala in it, or that you insisted on $300,000 for new capacity only to find out that 60% of your current capacity is consumed by photos of Britney Spears. I also know that if we put this technology in place it might jeopardize our "importance" within the organization, which is completely wrong. Being strategic and solving problems that prevent future problems is never considered anything but good by intelligent business folks.
The "green" phenomenon might just change everything. The traditional answer to storage management has been "just buy more." Managing is hard; adding more of what you know isn't. That won't work when you've run out of physical room or can't add a 40-watt lightbulb without dimming all the lights in the building. And it won't work if the next "wafer-thin" server makes the data center four degrees hotter than the sun. These events will require us to do things differently. And once you're forced to do things differently, the floodgates of possibilities open wide.
Money has historically solved all problems, but it won't anymore. It doesn't matter how rich you are if there's no more power to buy or you can't physically put one more pound of gear on the floor without creating another leaning Tower of Pisa. The good news is that the boring tools that "shine the light" are really good, usually cheap and will expose obvious things we can fix almost instantly to buy us some time. As a matter of fact, I contend that the average data center could probably save 25% to 50% of its green costs by simply changing processes, being smarter about how it uses the stuff it has and cleaning up silliness. Put that in your carbon-free emitting pipe and smoke it.