While it may no longer be the hot topic of conversation in storage circles that it once was, it's clear that the tiered storage model has taken hold in most large organizations. Given this fact, it's worthwhile to take a step back and assess how things are working out. The question that should be asked is whether your tiered storage strategy has provided the expected results, and if not, why?
Having seen a variety of storage environments, one observation that must be made upfront is that storage tiers often tend to align neatly with the product offerings of an incumbent vendor. This introduces at least the appearance that the tiering strategy may, to some degree, have been influenced by the vendor and was likely founded on a tacit assumption that cheaper storage would invariably result in savings -- "If you build it, they will come."
In establishing such a strategy, invariably some sort of cost justification, often in the form of total cost of ownership (TCO), was provided that included an extrapolation of projected savings. Such cost calculations are often incomplete -- what is presented as a TCO is lacking in the "T," emphasizing technology acquisition data with a very limited (or at best incomplete) accounting of ongoing operational costs.
One factor affecting operational cost that is rarely given proper consideration is the effect of additional complexity. Introducing additional entities for an already overextended staff to deal with can lead to inefficient management of at least of some of these devices. When complexity is factored into a comprehensive TCO model, the cost savings rationale may be significantly reduced; i.e., a carefully managed single tier of storage may turn out to be more cost-effective than several haphazardly managed tiers.
From a practical perspective, avoiding the problem is fairly straightforward: Don't establish too many tiers of storage. The right number will vary depending on the environment, but it is probably somewhere between two and five. How do you know if you have too many tiers? The answer may require a bit of digging, because it depends on understanding the fully-loaded per-unit cost of each of the tiers. If the unit-cost differential from one tier to the next is less than 40% to 50%, then it is likely that the tier may not be worthwhile. Certainly, there can be exceptions based on organizational needs, but the first rule of operational efficiency is understanding cost, and keeping it simple with a few well-managed storage tiers is a step in the right direction.
Jim Damoulakis is chief technology officer of GlassHouse Technologies Inc., a leading provider of independent storage services. He can be reached at jimd@glasshouse.com.