VOI and ROI of Release and Deployment Management
I was recently asked about the ROI and VOI of Release and Deployment Management.
Let me start by acknowledging that there is a lot of confusion about the difference between Release and Deployment Management and Change Management. Change Management is a risk management, governance process. Release Management is more actionable – it is about bringing one or more changes to life through defined pre and post production activities.
You could almost call it the DevOps process.
The growing requirement for rapid (some would say continuous) deployment does not undermine the need for quality releases. In fact, a structured approach to rapid deployment is more critical than ever since there is less time to flush out errors. You can make the process more agile by building release models for different types of releases. The models can match the rigor associated with building, testing, implementation testing and deployment with the complexity, risk, business need and impact of what is being released. The model will also address the need (or lack thereof) for documentation, early life support, user preparedness and communication.
Essentially, an individual release model will pre-define what needs to be done by both Dev and Ops for this specific type of release. As time goes by, the models may be a critical success factor in instilling an agile DevOps culture into your environment. Release and Deployment management is not bureaucratic by design – so go ahead and build some different models for common types of releases and test them out.
Here’s the VOI and ROI - failed changes and releases are very expensive and in some environments the damage could impact human life, market perception and/or loss of customers. With just about everything moving online – whether via the cloud or through traditional web services – the need to be able to “drop code” multiple times a day will continue to increase. Dev and Ops will need to cross the cultural divide, perhaps using release models as the bridge. Someday, this process may be as significant to DevOps as Incident Management is to the Service Desk.