Who really needs ILM anyway?

In much of the literature around information lifecycle management (ILM), vendors have unfortunately seized upon regulatory compliance as the underlying business case for their "solution." You need to bolt on some ILM technology to your storage infrastructure, they claim, in order to bring your stored data in line with SEC regulations and other governmental requirements.

This is far from the truth. Regulatory compliance requirements are a people and process issue, not a storage technology issue. Don't take my word for it: go look at the FEI study posted on Computer Science Corp.'s Web site to get your facts directly from more than 500 corporate financial officers. Contextualizing the requirement for ILM technology within corporate governance is the worst kind of marketecture, and the vendors should be ashamed of themselves.

And bogus ILM at that

Moreover, most of the half-baked ILM strategies on the market today are not truly ILM at all. While some may offer an effective "data mover" technology (that is, a way to move data to specific devices), the burden remains with the end user (that is, you) to undertake the laborious and time-consuming process of deciding what data needs to be moved where.

With virtually every ILM solution offered today, defining what data needs to be moved requires somebody to conduct a meticulous review of the data, as well as the applications and business processes that generate it. Similarly, defining where the data needs to be moved requires a meticulous analysis of the performance characteristics and cost characteristics of each of the heterogeneous storage platforms in your infrastructure. No ILM product performs these tasks for you, and performing them yourself constitutes a massive hidden cost of ILM.

Think about the effort you expend going through your shoebox and getting your receipts in order prior to filling out your income tax. Now, consider the effort required to sort through the huge data pile that your company has amassed over the years (and the equally onerous burden of characterizing the heterogeneous equipment that you use to host stored data on a cost/capability basis) to get ready for ILM. Time for Excedrin? You betcha, as my friends in Fargo might say.

Without these careful data definitions and platform characterizations, ILM buys you NOTHING in terms of cost savings, risk reduction or business enablement that you wouldn't get from a solid disk mirroring or effective tape-archiving system. Whatever TCO advantages that vendors claim will accrue from deploying their technology are specious at best. Don't waste your money.

ILM isn't really necessary, anyway

Truth be told, complex ILM systems would be completely unnecessary if companies were doing their IT right in the first place. For example, companies wouldn't need ILM for their file systems at all if they instituted policies that forced end users to assign names to their files that included well-defined "use codes," identifying whether the files they were about to save were subject to regulatory requirements, or were critical to the organization in a disaster, or needed to be made accessible to the broadest number of people, or needed to be handled with close attention to privacy. Any data mover worth its salt could reference the codes in a file name and use this information to put the file where it needed to go.

The same logic holds true for databases. All of the database archiving/neo-ILM products offered in the market today are fundamentally unnecessary. Using any of the current crop of RDBMS (Relational Database Management System) products, a good database administrator has the ability to design ILM features into his or her applications. You can readily include functions that sort down the database to identify stale data and write it to a separate table or dataset, while simultaneously purging it from the active dataset and preparing it for offload to an archival or reference data platform. If we designed such functionality into our databases at the outset, we wouldn't need a third-party product to do this kind of ILM.

The problem is that neither file-coding schemes nor database-archive functionality has been designed into the applications fielded by most companies. Why not? In most cases, the explanation is as simple as a slap in the face: Nobody ever made it a requirement for the designer.

Half-baked ILM order up

Like disaster recovery planning, a data management strategy is often an afterthought. In its absence, we look to vendors to give us a "bolt-on" solution to cover our collective tuccus with the auditors and regulators. And in the manner of good marketeers, the vendors are delighted to oblige us with half-baked technology (in most cases, just a rehashing of their archiving, storage resource management or hierarchical storage management software) that does little to actually resolve the problem and leaves the bulk of the work on the consumer's shoulders.

Their real agenda is to get your data into their proprietary system so that they own your storage budget going forward. In the final analysis, ILM is the biggest lock-in scam since the last one: Fibre Channel SANs.

So, why aren't we hearing this from Gartner, IDC and the rest? Ask yourself that question the next time they contact you to re-up on your services contract.

Going forward, you need to start doing something about your data management requirements - not just for regulatory compliance but so that your broken storage model doesn't drive you into bankruptcy.

Jon William Toigo is an author, journalist and speaker who has written more than 1,000 articles and 12 books, including two on storage. Current titles include Disaster Recovery Planning and The Holy Grail of Storage Management.


Copyright © 2004 IDG Communications, Inc.

7 inconvenient truths about the hybrid work trend
Shop Tech Products at Amazon