A lukewarm response

It's rare that a storage product announcement these days isn't somehow connected to information life-cycle management, or ILM.

Vendors, whether they are selling disk, tape or storage software, are promoting ILM. Product offerings vary, but the sales pitch is fairly consistent: Customers need ILM to manage staggering amounts of data. Vendors say ILM will provide a comprehensive framework for managing data from creation to deletion.

How ILM proposes to do this is tough to describe. The idea is to automatically migrate data to the most cost-effective storage platform based on the data's value to an organization. The goal is to optimize the use of expensive storage resources.

Vendor enthusiasm for ILM is unmistakable. But for government customers, it isn't necessarily a top priority. "It's kind of a new concept to us," said Tom Urbanik, systems software manager at the New Mexico Department of Transportation, which has been more concerned with shrinking its backup

window.

Indeed, ILM may be nowhere near the top of the to-do list for organizations absorbed with data-protection strategies, e-mail archiving and other storage priorities. And even if customers were ready for ILM, many observers believe a complete ILM solution has not yet arrived.

"We're two to three years away from vendors providing full ILM suites that resemble the visions we're seeing from them today," said Peter Gerr, a research analyst at Enterprise Storage Group.

Those issues aside, customers may begin laying the foundation for ILM even as they grapple with more immediate problems, industry executives say. Indeed, ILM is so broad in scope that many customers may already possess elements of the approach, even if they don't know it.

ILM illuminated

ILM has emerged as industry's antidote for customers' greatest storage challenges. Vendors see it as the vehicle through which they can address the big data problem, said Bill Yaman, vice president of the software division at Advanced Digital Information Corp., a tape library vendor. Through ILM, vendors expect to address a range of data types — structured data such as database records, semistructured data such as e-mail messages and unstructured data such as word-processing documents.

This ambitious vision encompasses a variety of storage products. At the hardware level, ILM presupposes the existence of a multitiered storage environment. Such an environment might include primary disk storage, a middle tier of near-line, less expensive disk storage and then archival storage, usually in the form of tape. Various software products shuttle data to the appropriate storage layer.

But ILM is as much about process as it is about products. "We attempt to get an understanding of the business process first," said Steve Alfieris, vice president at EMC Corp.'s federal systems division. Customers, he said, need to answer several questions: How is data created? How is it used through its life? How is it accessed? How is it valued?

In such an assessment, the objective is to appraise data at the application level so it can be allocated to the appropriate storage tier, Alfieris explained.

Mission-critical data, for instance, may be assigned to high-performance, high-end disk arrays, while data of lesser value may be stored on less-expensive Serial Advanced Technology Attachment disk devices.

The assessment/classification process is manual to some degree. But storage resource management (SRM) solutions can lend support. Softek Storage Solutions Inc.'s Storage Manager SRM product, for example, contains data-classification and data-profiling capabilities. The tool can extract information on file age, size, owner, type and naming conventions, said Gerard Svartz, product marketing manager at Softek.

Enterprise content management (ECM) is another software class that can aid in the classification process. EMC's Documentum division focuses on unstructured data, offering metadata tools for categorizing information. "We provide the intelligence to help organizations understand the value of their information," said John Magee, vice president of product marketing at Documentum.

Intelligence, regardless of its source, drives the valuation task. The next step is policy creation.

"ILM is the process of first making those value determinations, and then setting up protection, movement and retention policies based on those relative valuations," according to a March 2004 Enterprise Storage Group report on ILM. A policy might call for e-mail to be stored for two weeks in primary storage, then transferred to near-line storage for 90 days and then archived on tape for seven years.

Such policies, however, must be automated for ILM to deliver all of its benefits. "It's all got to be automated," said Charles King, a research director with the Sageza Group Inc., a market research firm. "Otherwise, management costs just go through the roof."

Vendors offer data life-cycle management (DLM) products that put policies into motion and automate the progression of data from one tier of storage to another based on an organization's policies. Yaman calls such policy-based data management the foundation of ILM.

Other executives also view DLM as a critical ILM component. "Obviously, there needs to be a policy engine that can drive the movement of data through each of these different layers, or storage tiers," said Chris VanWagoner, director of product marketing at CommVault Systems Inc.

The need to better manage e-mail inspires current customer interest in DLM, said Jeff Lundberg, senior product marketing manager at Veritas Software Corp.

Benefits and limitations

As for specific benefits, ILM promises to save money through the more cost-effective use of storage resources.

ILM aims to avoid situations in which seldom-accessed data is stored on an expensive high-performance platform. The combination of tiered storage and policy-driven data movement helps customers achieve the best cost/performance ratio, VanWagoner said.

ILM methods can also help organizations improve their data-protection programs. The task is to keep noncritical or aging data from accumulating on top-of-the-line, high-availability storage devices. "Data-protection strategies become faster and more efficient because [customers] are not processing old, stale data over and over again," VanWagoner said.

Finally, ILM may help customers comply with legislative mandates regarding data retention. Organizations can create and automate storage policies that are consistent with those directives.

Still, ILM is not necessarily a priority for government customers. Other storage concerns are more pressing.

The main storage issue for New Mexico's transportation department has been improving backup time. The department deployed backup appliance technology from StorServer Inc. to shrink file restore time from an average of about two hours down to a few minutes. "It's made a huge difference for us," Urbanik said.

In general, customers tend to focus on problems such as backup and data protection, rather than grand strategies such as ILM, industry observers report.

For the Corrections Department in Washington, D.C., the availability of mission-critical databases is a primary consideration. The department practices data replication and local tape backup. Backup sessions via the Internet to an AmeriVault Corp. data center provide another layer of protection, said Keith Godwin, chief of network operations in the Correction Department's Office of Management Information Technology Services. Godwin said ILM isn't practical for the department at this time.

Other government entities are taking on portions of ILM such as DLM. "Data life-cycle management is probably the end for us in the near term," said Terry Duncan, chief of the engineering branch of the Beam Control Division at the Air Force Research Laboratory. "Taking the next step up to [ILM] would be costly and offer limited benefits. Our problem is that the time and cost of configuration, design and implementation is too high for the marginal benefit."

Vendors are aware of the concern that ILM will be expensive to implement.

Customers fear that ILM "is going to add complexity and cost and management overhead," VanWagoner said. He said organizations that aren't careful in pursuing ILM could potentially end up with more copies of data, rather than less, as they move data across different storage layers.

Another issue arises with major applications such as customer relationship management that combine structured and unstructured data formats. "The long term challenge is, How do you provide some ability to link structured and unstructured data together?" VanWagoner said.

Yaman added that data classification is another area that needs refinement. The trick is being able to distinguish between an MP3 file that contains an executive's speech and one that contains pop music.

But vendor executives say government customers should start building the basis for ILM now, rather than waiting for all of the components to fall into place. Tiered storage, data-classification tools and DLM are all available today. Customers can begin to reap the benefits of ILM with those precursor solutions, vendor officials say.

"There are things they can do today that can provide the foundation that will drive efficiency in the organization