Storage Management: It’s All About the Price Tag
Customers say management features for heterogeneous storage should be a "given"—built-in (and free). Vendors say it's an add-on worth paying for. When will vendors wake up?
Opinions, they say, are like elbows, and everyone has a couple. I received many such elbows in response to my recent column covering the Storage Management Initiative Specification (SMI-S), which we described as an important move toward the improved management of heterogeneous storage infrastructure—but one that has been badly handled from a promotional standpoint by the Storage Networking Industry Association (SNIA).
My column last week concluded that, absent grassroots support for the burgeoning technology, there would be little incentive for vendors to create SMI-S “providers” for their storage products. A recent Webcast notwithstanding, SNIA has still done a pretty poor job of communicating the potential value of the technology to consumers, with the predictable result that no one is lining up around their local computer store to buy SMI-S enabled devices.
Of course, the general lack of enthusiasm for management technology like SMI-S may not be entirely SNIA’s fault. There seems to be a widely held view of storage management among consumers that it ought to be “free”—delivered as part of any storage product they buy.
This was underscored by a bit of unscientific polling conducted at the Networld+Interop Storage Track event that we chaired several weeks ago in Las Vegas, NV. I asked for a show of hands among the 100 or so attendees to identify how many made “management” or “manageability” a criterion in storage product selection. None of the participants raised their hands. I also asked how many had heard of SMI-S. One or two hands went up … and only after some hesitation, suggesting to me that they may have heard the moniker but didn’t really understand what it meant.
A follow-up question sought to discover what the primary criteria actually were for selecting storage products. A few answers were offered: capacity, performance, resiliency, security, price, and connectivity. When I queried about the relative importance of each criterion, price clearly won the day. Everyone in the room—a mixture of IT guys from larger and smaller shops—was concerned about price and held it up as the primary product selection criteria.
While hardly definitive, the poll went to the heart of the matter. Storage consumers are focused primarily on capital budgets and equipment cost. While they agree that manageability is important and generally agree that, without effective management, labor costs associated with storage administration are enormous and growing, most believed that management capabilities should be a “given,” built into the products they buy. The customers (at least in my audience) believe that he or she shouldn’t have to pay anything extra for management capabilities in the storage products they buy.
Of all the vendors out there, Microsoft seems to get this. Their notable absence from SMI-S meetings has been accompanied by a steady improvement in their server-based, heterogeneous storage management capabilities. The company's comments smack of hubris from time to time. For example, when I recently brought to the attention of a storage guy at Microsoft that his company had problems with some third-party virtual device drivers, I was summarily dismissed with a comment that Redmond’s products are for consumers, not for geeks who seek to “enhance” Microsoft platforms with third party components. Even so, the fact is that Microsoft believes that the management components should be integral to their product—part of its basic value proposition. Storage services in Windows Server 2003 do not cost the consumer any extra.
By contrast, most storage vendors seem to think that support for heterogeneous equipment management is an add-on and something for which they should be compensated either in the form of plaudits or pesos. Larger vendors who are implementing SMI-S in their products today are touting the fact as testimony to their “good-natured generosity.” Conversely, smaller vendors who are reluctant or unable to allocate resources to building SMI-S providers are stating simply that, until customers are willing to eat the cost, support for the heterogeneous management scheme will not be provided on their gear. Management isn’t free.
Needless to say, I prefer Microsoft’s position. Manageability should be a core value of any product that is sold in business technology space. Any product that doesn’t have it isn’t qualified for business use.
I liken it to buying a car. What automobile manufacturer would reasonably expect a customer to pay extra for the dashboard display—gas gauges, thermometer, speedometer, odometer?
Some storage vendors would argue that their products do have basic management features: command line interfaces or self-articulating Web pages that enable their product, like your car, to be administered on a one-off basis. They would claim that they are not responsible for providing the means to manage the car within the larger context of “traffic-heavy driving.” Just as features such as collision avoidance systems and GPS navigation systems cost extra in the automotive industry, support for heterogeneous storage administration is a “value-add,” not core value.
While the argument may have some merit, it also ignores a few other points—about cars, that is. Automobiles are designed to operate in traffic-laden byways. They have seatbelts and airbags, hoods with built-in crumple points and steel reinforced cabins—all security provisions that reflect contemporary driving conditions—as standard features. They adhere to wheel-base definitions to ensure that they fit into driving lanes, and they have mirrors to enable the driver to see what other cars are doing in relation to his own vehicle. They have air horns, blinkers, and other components to advise others of what the driver intends to do. These are all givens and they go to the management of the vehicle “within the context” of the environment in which they operate.
Automobile historians might note that it took a federal mandate to get most of these “management in context” capabilities added to contemporary cars. Makes you wonder what will be required to get heterogeneous management features added to storage gear. Send me your thoughts: firstname.lastname@example.org.