We're all getting familiar with the basic graph justifying information lifecycle management. The business value of data changes over time so that its initial residence on fast expensive disk gives way to storage on nearline or secondary disk and then on to archive media. But what is the 'business' value of data? How is it calculated?
Let's look at some business data - a supplier's identity for an office desk. What is its value? We could say it is the value of the desk. As the desk gets depreciated over three or five years its value decreases to zero. But the value of the data does not. We need a new desk. So we contact the supplier and order a new ZF-53C or whatever.
We can't really say that the desk identity data has undergone a lifecycle management process. It needs to sit on a facilities management computer system so that when it is needed it can be recovered.
Let's take a customer ID number. It's not a lot of use on its own. A customer record is different. It's value is the value of the customer's purchases over the duration of the customer's interactions with the company. If that customer's records are stored in a call centre, an accounts payable system and a dispatch system then who is to say which tier of storage those records should be on? When you want the data you want the data.
I don't think the business value of data changes over time. I think its access frequency does. I also think that sales, manufacturing and support data records are more important to a business than a junior marketing person's PowerPoint slide decks. But how you compute the relative value of these data items requires a clever mathamatician.
Where this leading is to ILM. How on earth do you set up different silos of information for different importance levels of information? Unless you have a disk array set with the same operating software and good, meaning transparent data migration tools you simply can't move data easily between tiers.
Data has to be classified and it has to be classified into as many storage tiers as you have. But not all data is the same. Is Centera data of the same type as manufacturing records.? I think not. I think ILM can only be applied piecemeal and is a real fertile ground for services organisations looking to save you money -maybe - and make themselves money - certainly.
They'll bring in compliance and start talking about data management strategies. Look at this pitch from Glasshouse:-
- With ever increasing regulatory requirements in the US and UK inaction is not an option, but simply buying more products is not the solution, unless they meet all the storage needs of the enterprise as a whole. If this seems too much to ask of a project, product or vendor then perhaps it is. Projects, products and procurement are all tactical considerations. As Sun Tzu wrote several millennia ago "Tactics without strategy is the noise before defeat".
- The strategic business importance of ILM is now widely recognised when properly planned and implemented. It develops service management best practice to align storage services with business requirements. ILM, within a service provider model, creates an effective business strategy.
- There are new dynamics within the forces at play that cannot be managed in isolation. Risk management depends on records management. Storage management requires data management. Business governance depends on IT governance. None of these tasks can be performed in isolation. Only a business that can manage its data can manage its costs, risks and customer service.
- Every business has its own style of cost, risk and service management, but the winning formula is a matter of business strategy. This cannot be dictated by accounting practices, organisational constraints or technology prejudices. However, tactical product selection is often the least critical part of the process. To complete Sun Tzu, "Strategy without tactics is the slowest route to victory."
This is fine marketing stuff but consider this sentence from it: "Only a business that can manage its data can manage its costs, risks and customer service." Ask yourself what business has been doing for the past 300 years? Has it not been managing its costs, risks and customer service?
Phil Tee,CTO of nJini, said, "Personally I think ILM is bunkum."
In these days of lightning fast storage developments with SAN, ISCSI, IP SAN, NAS, clustered NAS, FC disk, FATA disk, SATA disk, SAS disk, VTLs, hybrid disks, DPMs, continuous data protection, RAID 1,2,3,4,5, and 6, WAFS and grid storage services like Cleverspace plus Zetera's oh so cheap storage ideas the whole storage scene is just not stable enough to have an ILM infrastructure layered on top of it.
ILM is a concept whose time has just not come. It's like storage resource management a few years ago. Where are the SRM companies now? Scattered to the four winds. Park ILM at the edge of your storage radar screen and watch and wait.