IT managers at a conference last week said they face daunting tasks in trying to implement information life-cycle management approaches, including classifying data based on its business value so it can be stored on different types of devices.
Another big challenge is changing the attitudes of business units that are used to getting premium storage for every byte of data, according to about 10 attendees at the inaugural ILM Solutions Conference, which was sponsored by the Storage Networking Industry Association.
Users and storage vendors alike agreed that most ILM strategies are still works in progress. In addition to the internal challenges that IT managers have to contend with, ILM standards are lacking, attendees said. They also cited a need for more automated data migration capabilities and better integration between applications and storage devices.
But the sheer volume of information makes it imperative that companies begin classifying data and managing it via automated policies, said users such as Scott McIntyre, CIO at Quantum.
The need to comply with various regulations, such as Sarbanes-Oxley Act "provides the opportunity for us to bludgeon our business units to where we need to go" with ILM, McIntyre said. Quantum, which makes disk and tape drives, is moving from a storage approach built around data warehousing to what McIntyre described as an information retention model. But, he added, the company is "in the early stages" of the project.
The potential business ramifications of not adopting ILM "are absolutely huge", said Joel White, lead IT architect at Allstate Insurance. White added that Allstate's ILM strategy is twofold: the insurer is trying to achieve an enterprise-wide view of all its data while consolidating and centralizing its storage infrastructure.
Mike Peterson, an analyst at Strategic Research, and program director for the SNIA's Data Management Forum, said the key issues driving ILM adoption include an inability to effectively manage data and a desire to get information out of the silos where it has traditionally been stored. But dealing with the sheer mass of data "is really an ominous challenge for the IT group", Peterson said.
Conference participants also identified the need to keep old applications running in order to access and read archived data as another significant concern.
Peter Delle Donne, CEO of Iron Mountain, said the archiving services vendor recertifies older business applications on its systems after upgrades to ensure that archived data can still be accessed. Delle Donne added that he has sometimes been able to avoid the costly practice by saving data in Portable Document Format files. With PDFs, "we can present the information for a long period of time", he said.
But others disagreed that PDF files are a viable means of staving off business application upgrades for archived data.
"PDF is not the answer," said Tom McGuire, director of IT at a large pharmaceutical company. "In fact, it's a big problem." McGuire, who asked that his company not be identified, noted that an upgraded version of PDF software he purchased last year wouldn't access data archived in an earlier format.
Jim Zhou, a senior programmer/analyst in the IT architecture and engineering group at Genentech, said storage managers need to convince their company's business units that not all data has to be stored on high-performance disk arrays.
Storage vendors should develop standardized procedures for IT departments to follow in classifying data and setting service-level agreements with business units, Zhou said. He added that business managers could then choose from a menu of ILM options.
SNIA plans standards work to help boost ILM usage
Less than 10 percent of all corporate data is transactional, according to consultancy McKinsey & Co. The rest is fixed data that potentially could be stored on low-end disk arrays. Yet prior to being archived, most data is still stored on primary disk systems with annual operating costs of between 45 and 55 cents (US) per megabyte.
"If you're not accessing it, don't leave it there," said Lou Harvey, a member of the Storage Networking Industry Association's Data Management Forum. "That's where your cost savings will be realized."
How to lower storage costs through the use of standardized ILM practices was the focus of several discussions led by members of the SNIA's ILM Technical Working Group at the ILM Solutions Conference. Working group members also detailed plans to develop best-practice guidance for users and vendors to follow in automating the migration of data across tiers of storage devices.
Chief among the efforts that will take place over the next two years are initiatives to standardize the way different industries classify their data for storage and create interoperability standards between business applications and the middleware used to manage data migration procedures.
"Classify the data, and we can build whatever storage facilities we need to store it," said Jim Carlson, a senior technical staffer at IBM's storage systems division and a member of the ILM working group within the SNIA. Carlson said it should take the working group about six months to come up with a methodology for classifying data.
Jim Doustou, technical director at storage software vendor Permabit, is involved in the SNIA's efforts. He said that once standards are established, data classification shouldn't be an arduous task for users. Generally, the process shouldn't take more than six to eight weeks per line of business, he said.
"Do not do this as a big bang," Doustou advised. But, he added, "you can't deploy (ILM) technology until this is in place".