Iterative planning
We're currently working on a significant upgrade to a key piece of infrastructure. The implementation plan will run over several weeks as we replace older components with new ones and run them in parallel with old infrastructure before making the new components primary. Each step gives us an opportunity to validate the software configuration, automation flow and firewall rules before the new infrastructure goes fully live. We'll even be able to run live data through the system in real time to validate the upgraded software. It's not often you have an opportunity to do that, so when it becomes available you make the most of it. QA is great, but nothing beats real-world data!
We've had some very productive planning sessions to date. Earlier in the week I sat down with the Release Manager and we built a plan based on our prior knowledge and assumptions. Once we had a high level plan built we released v0.1 to the Subject Matter Experts (SMEs) for their review.
A further meeting with the SMEs fixed a few minor problems with the plan and an opportunity to validate the automation flow earlier was identified. Version 0.2 is now out for further assessment.
Having a couple of people put together a baseline plan to build upon is a lot easier than having all the SMEs together in a room create something from scratch. It also gives them the head space to identify improvements on your baseline. Very rarely will a baseline plan have to be entirely re-written, but it's usually in response to something that wasn't apparent until the baseline plan had been studied.
Eventually we'll have a v1.0 of the plan, and once the change request has been sent out for approval maybe even a v2.0, but I'll be confident that it will be a plan that meets our objectives.
- Previous: What do you do when you approve a change?
- Next: Spectrum of personality traits and the personality star
