Corporate management is just now beginning to really buy into the notion of enterprise-wide information lifecycle management (ILM). The data center will increasingly become the main nerve center for ILM, as the concept becomes reality over the next few years and as new and more creative, policy-based tools emerge.
Some observers believe there may even be some new types of ILM-related jobs in the data center, but that forecast is far from certain given the bumpy economic climate and the need for most organizations to cut costs and improve efficiency.
Of course, data centers have long engaged in a rudimentary form of ILM via nightly backups of production data that are put onto tape. At most shops, those tapes are then shipped to a secure location. But this just scratches the ILM surface, according to observers.
Data centers have traditionally been focused on backup and other related functions, but ILM is actually much broader and deeper than that. It involves the totality of business continuity, where backup and recovery are significant pieces of a much larger puzzle.
At its core, ILM is about putting information on the correct storage medium for the type and value of the information at hand. What you wind up with are different classes or tiers of storage based on the business needs and requirements of the people using the information. The tiers and definitions that guide them will be slightly or vastly different among companies.
Arun Taneja, founder of the Taneja Group, in Hopkinton, Mass., describes the problem that ILM seeks to solve. "In my current environment, I'm still keeping all the ERP data on my most expensive EMC box, even though 60% of the data hasn't been accessed in two months. And I'm still backing it up and taking snapshots," of all the old data, he explains. Compounding the problem is that, a lot of this backup and data movement are handled manually at most companies.
With ILM, "I'm now going to say that ERP and CRM are the most important applications, and they need to stay on my DMX box and be backed up and maintained," Taneja continued. "But then I'm going to move all my stale data -- and staleness is defined differently by different companies -- onto less expensive storage." The idea is that there's still access to this "stale" data, but now that information lives on a storage platform that matches its value -- whatever that value is perceived to be.
Despite the differences in how information is valued, the process and thought needed to put ILM into practice will be similar across most firms. ILM requires figuring out parameters to measure information's value -- by frequency of use, the age of the information, who owns it. Once a team of IT and business professionals decides on those parameters, policies can be developed to put those parameters into use. Policy-based software tools then automate much of this movement among the different
tiers of storage devices and help manage it.
For example, a company may decide to keep all of its current e-mail on its most accessible, fastest storage tier. After 30 days, anything that hasn't been accessed then goes on the second-tier storage system. And then after 60 days of inactivity, it goes on a third tier, until finally it goes into the permanent archives.
However ILM is accomplished, most agree that the data center will be at the heart of the operation. "For this strategically to make sense, companies will need to have centralized control over data storage based on business processes," said Clay Ryder, executive vice president at The Sageza Group, a consultancy in Union City, Calif. "You can't have one department" approach this differently from another, or it's not really ILM, he points out.
As ILM hits the data center, jobs will undoubtedly morph and responsibilities will change. The policy-based tools at the heart of ILM will free up operations and maintenance personnel to do other things. Ed Broderick, principal analyst at the Robert Frances Group, a consultancy in Westport, Conn., sees ILM as a potential opportunity for new classes of jobs, such as an archiving officer, data protection officer or disaster recovery officer.
"IT is being deluged" from the CEO and other higher-ups, Broderick said. "IT is looked upon as the one function that spans the entire enterprise and therefore becomes the owner of all this -- not to do it alone, in a vacuum, but to be the controlling agency that gets it done."
The first data centers to feel the ILM squeeze will be those in the most heavily regulated industries -- health care, financial services and government contractors. Not producing the required information to prove a legal claim false "isn't an option," Broderick said.
For their part, vendors are scrambling to help make ILM more of a reality. Mark Lewis, a senior product marketing manager in StorageTek's ILM group, admits that "software tools haven't really been mature enough." The idea of bringing together disparate tape, disk, SAN and NAS devices, and then automating data flow based on user-specified policies, has been difficult to make come to fruition.
But now, Lewis said, "our efforts are focused on delivering more unified capabilities" -- to bring all the various piece-parts together, and "we're starting to deliver on the realities of ILM." Ideally, these standards-based tools will both operate in a multivendor world and provide the intelligence necessary to move data among different storage tiers.
As all this unfolds, data center personnel will need to learn to think about their jobs differently, said Ken Steinhardt, director of technology analysis at EMC. "The IT administrator becomes more of a catalyst to help the organization apply technology more effectively, as opposed to being experts on the technology itself."
And one of the best ways to learn about the business is for IT to partner with a services organization that will work with business users to determine the value of information.
"The starting point is different with ILM than it is with traditional storage buys," he said. "Instead of starting by figuring out what hardware you need, you start with the intellectual capital of what information you have and its relative value to the organization. Then you figure out which policies to put in place, and then you decide which software tools you want to use to enable all that."
In this scenario, hardware is the last step -- not the first. With ILM, data center employees will have to think differently about projects instead of making decisions based on whatever IT infrastructure is already in place.
