Information Lifecycle Management -- the idea of managing the entirety of an enterprise's data throughout its useful life -- is rapidly taking hold in American business. Information Lifecycle Management, or ILM, deals with a lot more than just storage. However it has an immediate impact on storage, especially in the storage architecture, which must support the enterprise's vision of what ILM is.
The notion of vision is important because ILM is still a work in progress, somewhere between buzzword and codified best practices. Everyone tends to interpret the phrase differently in the light of each organization's needs. How it is interpreted in your organization is likely to have a major effect on how difficult and costly ILM will be to implement there. That makes it doubly important that storage managers participate early and effectively in shaping the corporate ILM policy.
The basic idea behind ILM is an observation that is intuitively obvious to every storage manager: Data has declining value in time, but the value of different kinds of data declines at different rates. All a storage manager has to do is look at number of accesses to data to see the decline. As data ages, it is accessed fewer and fewer times until eventually most of it trails off into a netherworld where it may sit for months or years without anyone accessing it at all.
The technical expression of this truth is found in Hierarchical Storage Management -- an architecture that moves data among various classes of storage as it ages, starting with fast, expensive storage today and eventually ending up on relatively slow but cheap storage such as offline tape. Information Lifecycle Management does the same thing, but it is focused on the entire enterprise and not just the way storage is organized. ILM starts by looking at the enterprise's information and how it needs to be managed rather than with the technical roadmap that characterizes HSM. Both HSM and ILM aim at more or less the same goal, but ILM takes a much broader view.
That broader view means that unlike HSM, which was almost entirely an IT initiative, ILM starts with the entire enterprise before it moves to the technology. Storage managers are certain to be involved in ILM projects, but they're not likely to be the leaders or the instigators.
ILM starts with the people who use the data -- or the people who have the legal responsibility for maintaining it. A lot of the interest in ILM is being driven by new requirements for reporting (in the case of financial records of public companies) or privacy (in the case of patient records in health care). In these cases the enterprise has to adapt its information management practices to the new legal environment, whether the aim is better transparency of information or better protection from unauthorized access.
Whether by force or law or the needs of the business, the users have to decide how long to keep information and how quickly they will need to access it. Both of these requirements are tricky for storage managers because of the different rates at which information loses value. And to further complicate matters, diminished value doesn't necessarily imply the need for rapid access.
Some information, such as legally required records, maintains high value for its entire life, but is seldom going to be accessed. Other information, such as a lot of accounting information, declines in value slowly, although the uses to which is it put change as it ages. Often it stops being current working information and becomes the basis for reports and analyses, the kind of thing better kept in a data warehouse. Most data shows a steep decline in value and accesses over the first 60 days after it is recorded, settling into a seldom-used limbo after about 90 days. A successful ILM storage strategy has to reflect these mixes of value and access as the data ages, and this usually means more than spooling the stuff off to tape at the end of 30 days.
Obviously this can lead to a pretty complicated data retrieval situation, because data that ages at different rates is ideally kept on different storage cycles. One concomitant of ILM is the notion of 'path management', as EMC calls it. This is the ability to find a particular piece of data no matter what file, volume or disk or tape it happens to be on. Path-management software, which is rapidly developing, virtualizes the process of data retrieval by keeping track of where everything is in the storage cycle.
There is a considerable body of information available on Information Lifecycle Management available from various perspectives. For an overview of ILM, StorageTek and Sun have a Powerpoint presentation.
One central source for information on ILM in both the government and the private sector is at the DOD's Defense Information Technology Testbed, which offers links to a variety of sites.
ILM isn't on next week's agenda for most storage managers, but in one form or another managing the lifecycle of information is going to become part of the storage manager's world over the next couple of years. Considering the potential impact on an organization's storage, it makes sense to get a jump on the subject.
Rick Cook has been writing about mass storage since the days when the term meant an 80K floppy disk. The computers he learned on used ferrite cores and magnetic drums. For the last twenty years he has been a freelance writer specializing in storage and other computer issues.
