In his latest blog post, Michael Hay of Hitachi Data Systems points out that expensive tiers of storage ought to be reserved for business data that generates positive cash flow. By extension, he suggests, non cash-generating applications should be restricted to lower tiers of storage. The logic here is hard to argue with: Why buy expensive infrastructure to support drains on the business?
Hay's argument that lower-tier storage makes more financial sense is interesting, but savings are limited. Any large enterprise IT manager will tell you that the cost of physical hardware (like disk storage systems) is fairly small compared to other IT costs. In fact, a CIO at a major company once suggested to me that the cost of hardware was flat out insignificant: High operational costs form a "glass floor" limiting the effectiveness of cost savings efforts. Saving 50% on the cost of disk drives is nothing when one considers that this line item makes up less than 10% of the total cost of ownership for enterprise data storage!
With this in mind, one could easily look at Hay's argument and ask an "emperor's clothes" question: Why are businesses paying to operate cash-flow negative systems at all? Why buy storage? Why spend time provisioning and managing systems?
Why not cut all these costs off and outsource these applications? Since non revenue-generating applications are a drag on the overall profitability of the business, the cost and effort spent supporting them ought to be cut to the bone and resources (both staff and infrastructure) should be reassigned to core revenue-generating systems. As Simon Wardley demonstrates in this presentation at OSCON '09, common IT systems don't represent a strategic value to the business.
This is easier said than done, however. IT has always faced challenges identifying data under management and understanding the business value of it. Consider the impact of this difficulty on compliance with encryption laws, for example. If IT can't decide which data supports revenue generation, they can't very well move everything else to lower tiers of storage!
But all is not lost. Even if some data is a metaphorical black box, IT often can identify certain application data. Many non revenue-generating support applications are easily spotted and could be rapidly offloaded to lower tiers of storage or outsourced entirely. Risk-mitigation applications like email and file archives are ideal candidates for this shift: They consist of large amounts of data and are easy to locate in the IT storage environment. Why not move these off to a service provider in the cloud and focus on core systems?