|
At a high level, a rule of thumb would be to keep high value data on your premium disk storage technology, then you could store less frequently accessed, perhaps aging data on lower-cost disk and finally, all remaining data that is kept for archive purposes on nearline, or even offline media such as tape, WORM, optical disk, etc. High value data in this context is, revenue generating or business-critical data.
However, this involves a manual data migration process that many software vendors have been trying to automate for quite some time. This is what has been driving the ILM efforts for some time but unfortunately, the options are still somewhat limited in terms of automation.
One of the obstacles in determining the value of data is that value is not easy to determine for certain types of data. For example, you can't always easily put a dollar value to medical records or government records. Sometimes, the value of data is not calculated in terms of lost revenue but rather, in terms of liability, loss of public confidence, or political embarrassment, which is much harder to quantify.
A simpler way to approach this may be to base your decision on access frequency and time. It begins to make sense to keep frequently accessed data on disk storage that can handle the performance and availability requirements associated with this type of data. Data that becomes less frequently accessed can then be migrated to lower-cost disk (which could, but not necessarily come with fewer costly performance and availability features). When storage decisions are made based on data access frequency, there are a few more options in terms of software products that can help identify the data to be migrated or even help automate the migration (i.e.: HSM solutions).
|
