It's What's Inside That CountsObsessed with controlling costs, the storage management field is in danger of missing the point: To manage infrastructure with a focus on data-the lifeblood of enterprise intelligence
by Richard Scannell Continued from Page 2 To accomplish this shift in responsibility, IT has to understand the architecture necessary at a number of discrete levels. For simplicity, let's assume three levels. The lowest might be defined by an architecture using only tape-based backup of cheap disk. The middle could be defined by more expensive disk utilizing a "snap" copy for near instantaneous recovery in a single data center. At the top level, the architecture might involve remote copies of data to a redundant, distant data center. Given an architecture standard for each level, IT can predict, based on the occupancy at each level, the cost of storage. As growth estimates are developed, the cost differential for a 20 percent growth at the lowest level will vary wildly from a 20 percent growth at the highest level. Accordingly, growth at the highest level would only be acceptable if the data were truly critical in nature to the business. This line of reasoning naturally brings us to the concept of governance. Two items about data are certain: it will continue to grow and its value changes along a given life cycle. Therefore, the data placement on a given storage solution should be changed to reflect its changing value. At a micro level, such a change is near impossible since the vast majority of applications have no way to either determine the value of data or affect its placement. However, at a macro level, it's possible to change the manner in which we manage large quantities of data. Based on the architecture defined, advanced features for protection, such as "snap" copies or remote mirrors in distributed data centers, can be turned on and off many times inside the same physical storage device. Many vendors now provide devices capable of storing vast quantities, and they often are flexible enough to adjust the protection features associated with a given data set. These protection features contribute in large part to the shifting cost from one category of storage to another. Management by ValueBy defining a set of architectures that offer discrete levels of performance, protection, and other attributes, IT can develop a cost model for each discrete type of architecture. With that cost model provided, IT can articulate the cost of providing storage for a given data set, which allows for discussions with business leaders that heretofore have rarely occurred. Rather than requiring IT to justify the cost of additional storage, the business leader will have to justify the placement of data on a given architecture level based on the value of the data to the business. IT can rarely control the growth rate of data, but it can now control the cost associated with that growth since placement at one, or a number of levels, can dramatically affect overall costs. With the ability to articulate this impact, focus shifts to the owner of the data to justify its placement at a given level, as appropriate. Transformation TimeThe issue of storage capacity and utilization management, while still critical to control costs, needs to be refocused to the issue of data management. Defined architectures, deliberate placement of data based on value, and ongoing conversations with business leaders to periodically review that placement will drive efficiency and cost reduction in ways business and IT leaders have not yet seen. As for that SRM tool you bought a year or so ago, hang on to it: It remains a tool in your bag to help with this transformation. However, if you thought your job was going to get easier as a result of that — come on — you're in IT, you should know better than that! Richard Scannell [rscannell@glasshousetech.com] is cofounder and vice president of corporate development and strategy for GlassHouse Technologies Inc. (www.glasshousetech.com), a vendor-independent storage strategy and services firm. RESOURCESRelated Article on Application Integration at IntelligentEnterprise.com: "The Next IT Imperative" Aug. 12, 2002
|
 |
Most Popular This Week
IE Weekly Newsletter
Subscribe to the newsletter
| |
| |||||||||||||||||||||||||||||||
