Agile for infra delivery
Hi
interesting question:
we need to deliver a piece of software by fourth quarter this year. The software is a commercial off the shelf product from a major vendor, we need to design the infrastructure, integration into our environment and configuration.
typically, we would use waterfall. A series of phases leading up to a launch. Often, we would pick a date for launch, eg 7 November, that would be based on those phases. Of course, this rarely works out since we can predict exactly how long those phasss will last.
but I can’t see how agile would help here since we don’t have sprints as such either. We need to build servers, firewalls and such in datacenters for example
both approaches seem to have challenges. Any advise?
From your description it seems that there is virtually no volatility in the requirements, you know every requirement upfront, and just have problem with exact time estimation (totally justifiable). In these circumstances I do not really see a convincing reason to be agile, therefore it can be difficult to use it as an argument of why so late :)
On the other hand, you could meet together, decide on what to concentrate on during next 2 weeks, then do it, and then meet again, see if you have delivered "the sprint goal", maybe even show it to someone (stakeholders) interested in project progress,check if you could have done something more efficiently, and then repeat...
Where does the risk or complexity lie with this product, and would inspection and adaptation through frequent release help to control it?
Thanks.
Filip - I like the idea of 2 week sprints to accomplish something, even it's not necessarily not a software release.
Ian - the risk lies in the vendor's inability to meet our requirements and also the MASSIVE amount of process and tollgates we have here. Example, in order to start evaluating the product in our test environment, we need to have a test plan already created. This seems odd.