Conventional wisdom suggests that a multi-vendor strategy for storage is the way to avoid becoming overly dependent on a single vendor and also to encourage sufficient competitive pressure to secure favourable pricing. But recently, I've begun to wonder if, in some cases, the pendulum may have swung too far in the opposite direction. Have the hoped-for multi-vendor cost savings truly materialized? Are some organizations actually penalizing themselves by attempting to be too vendor agnostic? Can a happy medium exist?
Several years ago, I was helping a large financial institution in the development of a storage strategy. Interviewing the storage group, which was not held in high regard by IT management, I was struck by their utter dependence on the resident storage vendor SE both in terms of day-to-day activities as well as in planning and design. The answer to almost every question seemed to be some variant of "We'll have to check with Joe." Unintentionally, the organization had, in effect, outsourced its institutional storage knowledge to its vendor, or specifically "Joe" (name changed to protect the guilty). More recently, various clients have voiced complaints about "overly scripted" or locked into propriety vendor functionality.
Conversely, we've also analysed the operational costs of other environments that have embraced a rigorous multi vendor strategy where added overhead and duplication of effort essentially ate up the capital cost savings. Despite this, politics dictate continuing along these lines because the Op-ex tends to be less visible, harder to interpret, and spread across several years whereas those big Cap-ex numbers quickly catch the eye of a watchful CXO.
How do we strike an appropriate balance? One thing that would help is realizing that, despite common belief, enterprise storage isn't really a commodity. Storage arrays vary greatly, and due to the lack of comprehensive heterogeneous management tools cannot be fully managed without reliance on at least some portion of a proprietary toolset. Furthermore, most storage managers would be justifiably horrified by the prospect of heterogeneous sets of LUNS (logical unit numbers) being assigned to a given production host. In effect, we have co-existence, not true interoperability today.
Perhaps what is needed is not so much a multi-vendor strategy as an effective vendor relationship strategy. Overdependence on a vendor puts a customer into a subservient position, but being overly neutral can deprive an organization from valuable insight that a solid vendor partnership can provide. Technology refresh is the time to put current and potential vendors on notice of the partnership requirements and preferences that will come from the refresh evaluation of available technologies with the clear understanding that the chosen technology will become the preference for the next x years. It is also very important that, prior to adopting multiple vendors and/or any vendor specific features, that a full understanding of the administrative costs of multiple technologies is understood as well as the back-out costs for any proprietary management tool This will add a key level of scrutiny and risk mitigation to the process.
Jim Damoulakis is chief technology officer of GlassHouse Technologies, a leading provider of independent storage services. He can be reached at firstname.lastname@example.org