The ABCs of ILM

By Heidi Biggar

Information life-cycle management (ILM), or data life-cycle management, is quickly becoming one of the storage industry's most overused and abused terms. ILM has come to mean anything from disk-based backup to hierarchical storage management (HSM) to data provisioning (or capacity management) to data migration. And more than a few vendors claim that the term is synonymous with elements of their storage management software—or even hardware—portfolios.

"ILM is becoming an in-vogue term used by vendors to describe what is often simply a suite of storage management products," says Phil Goodwin, senior program director, infrastructure strategies, at the META Group consulting firm.

"It's becoming one of those meaningless 'marketecture' words," says John Toigo, CEO and managing partner of Toigo Partners International. "ILM needs to be seen for what it is to avoid ill-advised buying decisions."

"As it's being positioned today, ILM is a lot of hot air," says Richard Scannell, vice president of corporate development and strategy at GlassHouse Technologies, a provider of storage services.

However, despite the negativity there may be something to the hubbub. ILM will eventually address a very real data management problem—one that has already been addressed in the mainframe arena but that has made little progress in the open-systems market for a variety of reasons, including ongoing issues with hardware/software interoperability, data management, and vendor posturing.

"Life-cycle management is a good thing," says Randy Kerns, a partner at The Evaluator Group consulting firm. "It's all about getting a better handle on your data and having your data in the right spot at the right time and at the right cost."

"ILM is about time management of information and knowing the value of information throughout its life cycle," says GlassHouse's Scannell. For some industries (e.g., engineering) the data life cycle is cyclical; for others, it is flat (e.g., medical) or short-lived (e.g., publishing), he says.

Whatever the industry, ILM is about being able to characterize the value of information to an organization over time and then building the appropriate infrastructure to manage the data cost-effectively. And analysts agree that ILM is not a hardware issue, nor can it be realized with software alone.

"ILM is really about process, not product, and IT organizations should not be tempted to substitute the latter for the former," says META's Goodwin in a recent report, No Silver Bullet for Data Lifecycle Management. Choosing and/or deploying products should be the last step, not the first step, in the process, he advises.

According to Goodwin, the general rule of thumb about ILM is to categorize business data types first and then relate business rules (e.g., access rules, retention periods, protection/compliance practices) to established data groupings. Once these groupings have been set, service levels can be matched to the business rules. These service levels should also be matched to appropriate storage architectures, and then products can be selected that meet the criteria.

However, analysts caution organizations about purchasing any storage technologies because of vendors' ILM claims, primarily because none of the products fully enable ILM yet.

"None of the technologies that are in the market today—or that are on the drawing boards—offer the level of granularity required to make ILM a reality," says Toigo. "ILM is about matching data to the right platform based on the performance requirements and cost characteristics of the platforms and on access frequency.

Some of the basic technology exists (e.g., data movers, storage resource management software, policy engines, etc.). However, a lot of work still needs to be done to make applications more storage-aware, or vice versa, so that applications are more aware of the types of data they are producing, associated business rules and service levels, and application/data access patterns.

"ILM, when it becomes real, will determine inside the application which data is more important and then set policies accordingly," explains Scannell. "This will allow organizations to focus on the type of services they need, not on the transaction."

The bottom line, analysts say, is to buy a product because it addresses a current pain point, not because it's labeled "ILM."

This article was originally published on January 01, 2004