ARTICLE POSTED February 13, 2006 ILM: If it's not a product, what is it? By Mary Brandel The concept of information lifecycle management has been around for a while, but the storage industry, it seems, is still trying to figure out exactly what ILM is. Consider, for instance, that in a recent survey of 252 users conducted by the SNIA End User Council, 42% said they "neither agreed nor disagreed" with the definition of ILM developed by the SNIA in November 2004. Certainly the driving forces for ILM—namely, cost, complexity and compliance—continue to grow, and user interest in adopting ILM practices is high. That's no surprise, with data growing at a rate of 30% to 50% per year for many organizations, as well as increased government regulations that call for companies to retain certain data for set periods of time. "Many organizations have ILM initiatives or are interested in starting one because of all the publicity," says Jim Damoulakis, CTO at Glasshouse Technologies. "But the biggest driver is that storage costs are growing at an astronomical rate, and they're being mandated to cut the storage budget." The vision of ILM is to classify data in terms of its business value and then align that value with how the data is stored, managed and protected. The result: reduced costs, maximized storage utilization, minimized redundancy and protection in regulatory compliance situations. But that can only happen when users fully grasp what ILM means to them and their particular situation. According to the EUC survey, that isn't happening on a wide scale. Respondents said they had yet to develop fully defined ILM requirements or a business case for ILM. "It's worse than virtualization as far as, 'What does it mean?'" says Cliff Dutton, executive vice president and chief technology officer at Ibis Consulting, Inc., a provider of electronic discovery and compliance solutions in Providence, RI. Golden rule of ILM: Don't think 'product' Part of the problem lies in the fact that ILM, above all else, is not a technology or product that can be purchased. It is a process that requires defining the value of data in the organization and then storing and protecting that data accordingly, even as its value changes over time. Value is defined by factors such as how critical the data is, how frequently it is accessed and how long it needs to be retained. There's a range of technologies that can help companies discover, classify, store, archive and automatically move data, according to Jon William Toigo, CEO of storage consultancy Toigo Partners International. These technologies include HSM software, content management systems, virtualization tools, e-mail/database archiving products, backup software, SATA arrays and content-addressable storage arrays. However, no set of tools will move you an inch closer to ILM until you've done the laborious front-end work. "One of the biggest mistakes an organization can make is to center the ILM discussion on infrastructure, not business or policy," Damoulakis says. In fact, a lot of people make the mistake of equating ILM with a "tiered storage infrastructure," where you create layers of storage that move from low-cost/low-performance to higher cost/higher performance. "Tiered storage by itself without accompanying good management practices really could have negative consequences," says Damoulakis. "Rather than tiers of storage, you may be better off with several classes of storage that are implemented on a single storage platform." For instance, you might store your data in a top-end system, with some data requiring remote replication and the rest not, depending on what the business or application requires in terms of availability and recoverability. "If you don't know what you're trying to do from a business standpoint, you're going to make the wrong decision," says Tom Ko, director of data center operations at HIP USA, a $4 billion HMO in New York. "Anyone who says it's a tool is going down the wrong path." Second golden rule: Keep it simple Another problem is that with all the hype propagated by ILM, people can easily become overwhelmed by grand visions of ILM nirvana. For instance, according to Glasshouse Technologies, the idea that vendors are touting today about moving things around automatically based on value assigned to the data may never be fully implemented. "Most applications aren't designed in such a way that data can be moved around underneath them and still function," Damoulakis says. In addition, the idea of an organization cataloguing and creating meta data for all the data in their entire enterprise can also stop users dead in their tracks. "You could chuckle when you look at the layout of tasks in some of the vendor literature on implementing ILM," says Damoulakis, "like, 'The first step is to classify all your data.'" According to Bernard Shen, an independent contractor specializing in storage architectures and server consolidation, "Maybe there's a subset of data that is business-critical and has certain meaning, but cataloguing everything on a grand scale is just not practical." While you need to develop a framework that can scale and keep your focus enterprise-wide, it's equally important to not get mired in impossible tasks. "ILM is the kind of thing you can only get better at; it will never be done," Dutton says. Labor-intensive Dutton warns that the most time-consuming part of implementing an ILM strategy is defining the business requirements for data storage and protection. "The first pass is extraordinarily difficult," he says. "That's certainly where the most energy is required." But this stage is also where you determine how you're going to capture return on investment. "You need to clarify the goal—is it disaster recovery, business continuity, regulatory compliance, or meeting service level agreements with external or internal clients?" he says. "The business benefits will help define what amount of money is worth spending." After defining data policies, Ibis implemented a storage architecture that uses an Acopia Networks NAS gateway to migrate data across storage tiers based on rules set up in Acopia's policy engine. Data is stored on a 200-terabyte, ATA- and Fibre Channel-based storage array from BlueArc Corp. and is archived on a tape library from ADIC. Currently, migration to tape is not managed by Acopia. "We don't have an integrated framework for ILM," Dutton says. "We have an integrated set of policies that embody ILM." Invisible tasks Another often-overlooked job is establishing a business case for ILM that resonates with business users. Many business departments don't bear any responsibility for paying for storage and so are unaware that poor storage management policies result in escalating costs. "They're happy using expensive disk if they don't have to pay for it," Damoulakis says. "The benefits of ILM reside in the IT organization, and those benefits don't necessarily translate to business users." It's a matter of making those costs apparent to the business, Dutton says. "When they realize it's an enterprise cost to buy capital equipment called 'disk' to store junk, they'll realize the benefit of doing this," he says. One way to drive home the point is to implement chargeback policies, which ILM can enable. Such is the case at HIP USA. According to Ko, the biggest reason HIP USA began its 18-to-24-month ILM project was spiraling storage costs for its 60 to 70 terabytes of data. "We started the ILM process for the same reason as everyone—hardware and maintenance dollars were going through the roof," he says. "But as the project evolved, we started asking, 'What does it mean for us to provide services that the business needs to do business?' That's when we started to engage more and more with the business because when we begin charging everyone for storage, we want people to feel comfortable with what they're paying." Thus far, Ko's group has been working with the various business departments to establish data criticality and regulatory retention needs for their various applications and pieces of data, which get translated into SLAs and retention lifecycle policies. The group recently completed an eight-month RFP in which it selected EMC and Hitachi to provide the platforms on which the data will reside. And to establish migration criteria, it contracted Glasshouse Technologies to classify the data already residing on its storage systems. HIP USA chose an EMC Symmetrix and a Hitachi 7700 for Tier 1 data, with data replicated to a third-party vendor. Tier 2 data will reside on an EMC Clariion CX system, and a StorageTek 8055 tape library will serve as Tier 3 for backup. The standard migration policy allows for Tier 1 data to be stored for 14 days, with Tier 2 data stored for 30 days and Tier 3 for a year. Separate policies will exist for other applications that have specific regulatory needs. An evolving definition At this point, Damoulakis suggests the term "ILM" could be morphing into a metaphor for storage managers simply trying to implement good management practices to reduce costs. "Conceptually, ILM is like motherhood and apple pie," he says. "It's a worthy goal, but it falls very short from a practicality standpoint at this stage in the life of data management." About the author Mary Brandel is a freelance writer based in Newton, MA. Feedback: feedback@snwonline.com Questions: questions@snwonline.com Technical difficulties: webmaster@snwonline.com