Quocirca's Straight Talking: Storage meets business

Like IT itself, a recent trend in storage is to make it strongly influenced by business priorities. Quocirca's Jon Collins looks at what it will take to make this a reality.

The origins of the term Information Lifecycle Management (ILM), coined in late 2002, are obscure, but most likely it came from the thinking of certain strategists at StorageTek.

It's a great theory. Spawned from the idea that there is more to storing data than, well, storage, the thinkers set about making it work better.

The central concept is that storage can be categorised according to how fast and expensive it is. This is a sliding scale - fast (think high-end disk) tends to be expensive, and slow (think tape) tends to be cheap, with several categories in-between. Now, thought the thinkers, what if data could be similarly categorised? If so, high-value data could be stored on top-end disks, low-value data on low-end tapes, and so on.

Where the idea gets really clever is the inclusion of software to shift data from one storage platform to another, depending on how it is being used.

Hang on, you storage diehards might be thinking, isn't this just a re-hash of Hierarchical Storage Management (HSM), a dynamic archiving technique? Well, yes, and no. First, HSM was one-dimensional, covering only the time value of data and not taking into account its value to the business. Second, and because of this, ILM requires classification of data by real human beings. Finally, HSM was largely about archiving old data. If ILM is done right, the concept of archiving goes away and is replaced by a continuity of dynamic data movement.

As a practical reality ILM has existed for about a year now. For a while, there was a danger that (like many good ideas) it would be turned into a sales vehicle by the storage vendors' marketing departments. EMC was quick to jump on the bandwagon, and a number of other companies including Veritas have adopted the idea of 'lifecycle management' - in Veritas's case, it is called Data Lifecycle Management. Hewlett-Packard and IBM use the ILM phrase, but it is buried somewhat in their messaging.

How has it evolved? As marketing took hold, it became clear there were a number of issues with how ILM was being presented. First, while ILM might be a good idea in principle, it was not articulated in a straightforward manner and nobody on the customer side understood it. Second, in a number of cases (including StorageTek, which thought of it), it really was little more than marketing - both because companies saw it as so and because they didn't put their whole product lines behind delivering on it. Third and finally, to work, it required a much more comprehensive approach to storage solutions than had existed in the past, requiring a holistic architectural view and services to support its delivery.

Despite these issues, the vendors have stuck with ILM; indeed, they have emphasised its importance. The reason being that it really is the right answer in the same way that service-oriented IT is the right answer: of course storage should be treated with attention to its business implications. There are other reasons, not least the emergence of compliance as a key driver for improved storage solutions. A compliant solution is an integrated solution, and ILM offers an approach to support such integration.

As a result, vendors have recognised they need to make a bigger effort than previously if they want to make things work. Not least, they are - on the whole - understanding that ILM needs to be services-driven. For example, EMC has announced an assessment and classification service that is an entry point to ILM, and StorageTek has augmented its services organisation with ILM in mind. Channel companies like Dimension Data and Glasshouse Technologies are also reacting to the trend, augmenting their storage offerings accordingly.

Perhaps even more important are the growing partnerships with content management (CM) companies. EMC has acquired Documentum and StorageTek has partnered with Ixos, which is merging with Opentext. The reasons why ILM can benefit from content management are many, but it suffices to say CM has long had both feet firmly planted in the business, even while storage management is trying to open the door to the data centre.

Progress is being made with ILM, but there is still a way to go. Computer Associates have just announced the addition of security features to its BrightStor range. While this is perhaps laudable, it begs the question of why such features weren't considered as standard for the past three decades. There are also ongoing integration requirements to improve management capabilities: for example, incorporating policy-based ILM with other aspects of enterprise management. And it is still early days for such ideas as adaptive storage and automated management, both requirements for ILM.

Dynamic provisioning and storage accountability are even newer, as is building in the financial aspects of storage service delivery. For ILM to work, a data store should be able to report exactly how much it is costing to store the data, down to the byte if necessary. Only in this way can the vision of ILM - moving data according to a dynamic assessment of cost and benefit - be addressed. Once this exists it can be linked to application costs and ultimately the business can have a clear picture of how much its storage is costing. At that point, there is a question of whether it should still be called ILM, but it's as good a name as any.

Ultimately, however, even if ILM is the right answer, it can only work if businesses want it to happen. This is a bigger issue than some might imagine. Businesses don't necessarily feel they need this level of control over their stored data, and thus aren't prepared to explain to vendors what they need in order to adopt it.

We can come up with all the acronyms we like, but they are nothing more than ongoing efforts to make IT work for the business. For ILM to succeed, it needs to be considered in the round, alongside other business-oriented aspects of IT.