I have never been a big fan of ILM (information lifecycle management)—the term, not the concept. This is mostly because the concept is so broad but the term itself seems so limiting.
When people hear ILM they think information, which reminds them of data. They think of data and it reminds them of storage.
Next thing you know, ILM is a storage-centric solution. My problem with the term then is that information management is about much more than storage.
Every single ongoing IT-related project is ultimately has information management issues. Be it data warehousing, server consolidation, application integration, supply chain management, its all fundamentally about information management.
The problem with big concepts is we tend to approach them one piece at a time.
So if you are a former database guy like me you tend to only see the benefit of database archiving or subsetting. If you administer the corporate e-mail server and are being hounded about compliance or legal discovery issues, you tend to focus on the e-mail archive. If you are a storage administrator, you are focused on offloading data to cheaper drives.
The list goes on and on. Each little interest group solves its niche issue in a vacuum until one day the many little problems we solved become the cause of some even larger problems.
Years later some bright person gets the idea that all of these small niche problems were actually the same problem.
This is an age-old issue, really. Some of us experienced this as distributed systems became more and more prevalent. We realized that the problem of having to maintain a number of point products to monitor and manage layers of infrastructure (such as the server/OS, database, network and applications) became harder to overcome than the original problem we were trying to solve.
The solution then was a unified management framework. The frameworks applied standards to handle messaging from the various monitoring agents and a common infrastructure to handle event management and more.
Software vendors like Computer Associates and BMC Software came out with these unifying frameworks, including UniCenter and Patrol, to address the larger overall problem of systems management.
Of course the problem was that most of the point product software that these software frameworks were suppose to integrate were not built from the ground up for that purpose.
Needless to say there was a great deal of pain that some organizations are still suffering when it comes to a seeing true ROI on a unified systems management approach.
The ILM issue, like systems management, is such a broad concept that it touches virtually every aspect of the IT infrastructure (not just storage) and impacts the success of virtually every IT initiative.
Right now, in your organization, special interest groups are out there evaluating niche ILM solutions and believe me there are plenty of ILM vendors that only do one or two niche things and are happy to engage with those groups.
Taking a Unified Approach
If you are a high level IT manager I urge you to stop this activity now. Why repeat the mistakes of the past if you dont have to? With all due respect, the distributed systems management market had to grow up and mature. They had no real example to learn from.
The ILM space has no such excuse. ILM is an enterprise strategy not a series of niche approaches. Information is information, whether it is structured data in a database table, semi-structured data in an e-mail or unstructured data in a document. At the end of the day you should desire a single point of management.
There are many reasons—technical and business—for a unified approach. But Ill just mention one. Some day, someone is going to need to search for some information. They may not know if that information is in a database, contained in an e-mail, or in a Word document. They just need to find it.
It is likely they will be under some time constraint. Wouldnt a unified ILM software architecture enable that search better than a hodgepodge of point products, each with its own metadata store and user interface?
Lets not fool ourselves either; enterprise search is the ultimate killer app for ILM. In speaking with Sai Gundavelli, CEO of Solix, an ILM vendor, he articulated this vision best when he told me that Solixs mission is to organize all enterprise information.
Now, organizing information is like organizing your tax records or your garage. The goal is ultimately to make not only your day-to-day work more efficient but to make things easier to find later on when you need it. That is a simple yet very profound vision when applied to enterprise data.
We are now seeing the pace of consolidation heating up in the ILM space with EMC and Hewlett-Packard leading the way. Both have the vision, I believe. But like the systems management vendors of the past, they are trying to accomplish it through a series of acquired point products.
One can only assume that they, like CA and BMC before them, are busy figuring out how to cobble together a unified ILM software architecture from a bunch of piece parts. I call this approach a market vision which typically precedes any actual technical vision, unfortunately.
At least for me, small companies like Solix offer some hope of relief, as they not only have a unified ILM vision for the market but have designed their software from day one to eventually handle such a task.
They are the first, but hopefully not the last vendor in the ILM space to deliver software built with the realization that ILM is really one broad concept, not a niche strategy.
Its not important that any vendor can actually do ILM for any source of data on any platform today. But we should at least ask the question: Have they been designing their software for this eventuality?
After all, no one wants to repeat the mistakes of the past because thats just silly.