As we all know, all organizations are in fact becoming a software business collecting and analyzing data with the vision to build better products faster. However, the challenge with new DevOps processes driving agile development is that multiple copies of data are requested by the DevOps team at any point in time. There are four problems associated with these so-called ‘down level environments’ (non-production):
- Speed – How fast can you clone the data? On average it takes 2 weeks to clone a database right from the initial request to the final provisioning.
- Costs – Who is going to pay for it? Storage, infrastructure, power, and cooling costs all add up.
- Quality – Most often, customers end up provisioning a subset of data. As a result, you cannot do regression testing, performance and scalability testing early on. Bugs are found very late in the test cycle at which point you have to go back to the drawing board or release buggy software.
- Security & governance – With so many clones floating around, it’s a nightmare scenario from a security and privacy perspective.
Denny Cherry in his blog in Gestalt IT articulates many of the challenges in keeping down level environments populated. In particular, he describes many of the challenges associated with security and privacy and human effort required to keep them populated and the need for a software solution that integrates with DevOps processes.