A recent reader response to my State of Storage 2011 report got me thinking about the difference between data and information, or, more generally between technology itself and the ways we creatively use it to solve problems. State of Storage, as these annual reviews are wont to do, focused on trends in storage technology. My correspondent, however, called me for not discussing information management-one way we apply tech to improve how we do business. "I was underwhelmed with the article in that it only focused on the technology platform components of storage and was silent on how we should organize the data we store," he wrote. "In an old-world analogy, it would be like the Library of Congress worrying about the number of shelves they build and the type of wood to build the shelves."While we always like getting feedback, this is akin to complaining that a Motor Trend Car of the Year report is silent on the topic of alternative fuels and national energy policy: "All you talked about was engines and handling. There was nothing on moving away from petroleum-based vehicles to battery cars and fuel cells." The State of Storage report really was tailored to shelf-builders, not library curators; it's the nature of the beast. Similarly, our upcoming State of the Data Center report will explore PUE, cooling system design and cable management, and not make value judgements about the type of data stored, or applications running on the servers overheating those racks.
Still, our correspondent's larger point is valid: "Prior to buying more shelves, we need to decide what we will store directly ... and how I expect users to find the right information at the right time. Those decisions will drive the platform architecture decisions and make for a more effective solution." Indeed, when faced with increasing demands, IT has a tendency to just add "shelves" (read: disk arrays or servers) rather than addressing the underlying factors fueling the demand. The reason is simple: It's historically been easier (and cheaper) to just throw hardware at the problem rather than peel the onion to examine larger information or application architecture deficiencies.
However, the days of easy technology fixes really are ending. Managing vast pools of data is consuming an increasing share of the IT storage budget and, as my report points out, adding more spindles is unwise. While my recommendations concerned technical ways to make more efficient use of existing capacity, our correspondent highlights another strategy for tackling out-of-control storage demands; information management. You can maximize storage with techs like deduplication, or you can just store less data.
The concept of information lifecycle management (ILM) isn't new, but if you're going broke building new shelves, it's worth a second look as part of an effective enterprise storage strategy. ILM isn't a new topic, yet it has been tried (and often abandoned) by many enterprises over the years, but not because it's a bad idea, but rather because it's hard and doesn't lend itself to simple technological fixes. Products that equated the frequency of data access with information importance were easy for vendors to implement, but woefully incomplete as information management tools. Certainly the technology for categorizing, indexing and archiving data has greatly improved, as evidenced by automatic tiering features mentioned in the State of Storage, but technology is just a tool. ILM is really more about retention standards, archive processes and IT governance. ILM is undoubtedly an important element of an effective enterprise storage strategy and one worthy of its own treatment. I thank our writer for pointing this out and look forward to addressing his concerns in a future report.