The value of storage management

* Editor’s note: Mike Karp is on vacation this week.

As storage management continues to mature and new technologies like iSCSI impact the storage landscape, it is important to periodically consider the fundamental concept of all storage investments and ask what is the value that the investment delivers. The process of answering that question involves considerable analysis and engineering rigor. It deserves significant energy.

The process of determining value for storage solutions turns up in the almost religious conversations about protocols and service architecture (network-attached storage vs. storage-area network, for example). These conversations rarely focus on the financial impacts of the possible solutions to a given problem, instead choosing to develop along the lines of speeds-and-feeds or feature/benefit conversations. The challenge in developing a value-based rigor is that it requires understanding very complex issues, especially in the modern enterprise.

As an example, consider the requirements for storage in a typical networked enterprise. Applications as varied as the company’s Web site and the inventory management system all require storage. Add to the applications the need for end-user storage and the geographic diversity of many organizations, and the challenge multiplies.

Typically, however, the complexities of the range of storage solutions and the implications of understanding the possibilities make value analysis difficult. Given that most technology staff do not have the opportunity to understand the various options and to stay current with all technology, the task can be arduous. However, it is the role of IT to know and understand how to provide the best value in technology services to the business and this includes storage management.

One possible approach is to develop a value model for each application supported by the infrastructure and a cost model for each technology component that serves each application. However, be sure to consider design elements such as scalability and redundancy, as well. For example, in considering scalability, the best design may be for relatively inexpensive servers scaled by adding new servers sharing the same data on the back-end. Alternatively, there may be a value in using large, multi-processor servers with dedicated storage. But, the design must consider the implications of system failure, scaling, and application requirements.

In the end, the value should be monetary. The process should reduce all of the costs and benefits to dollars and cents, although soft costs may be best communicated in terms of time such as hours lost or labor hours required. In order to develop this model accurately and to make the decision effectively, providers can help designers understand the technologies they don’t understand and use the process to help demonstrate their product’s value. In dollars and cents, please.