The concept of Information Lifecycle Management (ILM) is widely bandied about in the storage world. Most major storage vendors are touting their ILM strategies, and when new products are released, vendors describe those products' roles within the overall ILM framework. Companies are even acquiring other companies with an eye to bolstering their ILM portfolios.
As is often the case when a concept sweeps through the industry, the term ILM can have very different meanings to different communities of users. Moreover, different vendors might push the idea for very different reasons. Consequently, the term ILM leaves a lot of room for confusion, which could lead to poor customer choices and implementations.
ILM isn't an emerging product category in the same sense that Storage Resource Management (SRM) is. Rather, ILM rests on the common-sense notion that as data diminishes in value over time, it should be stored using a less expensive storage technology. ILM is an exercise in data stratification. Todd Rief, director of corporate strategy for StorageTek, said that ILM differs from the older idea of Hierarchical Storage Management (HSM) because HSM focuses solely on data movement, whereas ILM also encompasses data protection. Like the airline industry has discovered about its passengers, Rief said, storage professionals have determined that some data is more valuable than other data and thus deserves different treatment.
The industry is enamored with ILM now for at least two reasons. First, because storage technology has become more sophisticated, faster, and cheaper, companies are building more complex storage infrastructures. From a directly attached high-performance hard disk or a Storage Area Network (SAN), for example, companies can move data to lower-cost near-online disk storage, then to tape; move data directly to tape; or archive some data while preserving other data on the production hard disk.
Second, the number of data protection schemes is also exploding. For example, companies can use mirroring techniques, point-in-time snapshots, and continuous backup. At the same time, customers have a sense that the most expensive, high-performance production storage is poorly utilized. Industry observers have noted that high-performance disks are often filled at only 30 to 35 percent of their capacity. Although SRM is intended to rectify that situation, the implementation of solutions has been slow.
But improved technology is just one reason that the idea of ILM resonates so clearly in the enterprise. Regulatory-compliance and business-continuity concerns require an improved data management layer and a better archive layer in storage implementations. In addition to being able to retrieve specific files expeditiously, companies must be able to document that those files weren't tampered with in any way.
If successfully implemented, ILM could deliver lower-cost, more flexible, more efficient storage infrastructures. But the road to realizing those benefits isn't at all clear. No single company offers all the pieces to complete the ILM puzzle. James Lee, vice president of product marketing at Princeton Softech, a vendor of archiving technology that recently cut a deal with Network Appliance (NetApp), argues that ILM will require a best-of-breed approach to selecting specific pieces of the solutions.
However, selecting best-of-breed technology requires people to make difficult choices. Moreover, best-of-breed solutions involve complexity. Because different vendors supply different parts of the answer, those pieces must be integrated, and projects that require a lot of integration often have many points of failure. But implementing solutions that introduce more points of failure runs counter to the instincts of most professionals responsible for safeguarding company data. Although senior executives give lip service to Return on Investment (ROI) being the most important criterion in deciding which technology to use, for people on the front line, minimizing risk is the first priority. If a system fails and crucial data is lost, the cost-effectiveness of the implementation no longer matters.
For many vendors, ILM starts as a service offering. Consultants will help companies assess where their data resides and how to reengineer the storage structure. In other words, vendors will help companies think about how they can look at their current setups through the lens of ILM. But the thought process is just beginning. The outcome remains out of sight.
