“Data expands to fill the space available for storage” is a popular (in the Systems Management community at least) corollary of Parkinson’s Law [1].

But data-sprawl is no joking matter to the frustrated IT executive, handling the latest demand for more storage and backup capacity to hold the ever increasing tomes of company data (most of which will never be used again – except to justify demands for more data storage).

Data-sprawl increases operation cost, degrades system performance, and inhibits system update and migration initiatives. Therefore, it is important to integrate data storage costs into your Application Value Management. Unfortunately, storage costs are often lumped together in infrastructure budgets making it difficult to understand true application costs and make decisions based on value.

DE:CRAP is a humorous but practical suggestion for approaching your data storage decisions (and supporting Application Value Management). The rules are simple: The less valuable the data, the less you should be spending on it.

Document, Evaluate: Compress, Replatform, Archive or Purge.

The first stage is to document and evaluate your data: repositories; categories; retention policies; what business processes does the data enable or support (very important); what kind of access is required; how much of the infrastructure cost is related to storing it. You can now better understand the value of each dataset to the business and how much should be spending on storing it, protecting it and making it accessible to users (if anything at all).

Armed with your evaluation, informed decisions about what to do with it become easier.

If the data is of high business value and needs real-time access, then this is where you should be placing your investment in maintenance and improvement of the supporting infrastructure.

However, if the data is of lower value (but you need to keep it online in some form) then this is where you look to cut costs. You can consider tools to Compress the data so that it consumes less storage space on your existing, or Replatform it to lower cost infrastructure (such as cloud storage).

If the data has little or no current value to the business, then you want it off your production infrastructure ASAP. Purge it if possible, but if you are bound by regulatory requirements or other retention schedules then your only choice may be Archival. Nevertheless, you need to make value-based decisions about the type or archival, and keep track of the ongoing costs and schedules: eventually you will be able to eliminate the archive, and lower cost alternatives may appear for your existing archives

Happy De-Crapping!
Andy

References:
[1] https://en.wikipedia.org/wiki/Parkinson%27s_law