Friday, February 12, 2010

Is "Big Bang" a good way to implement Infrastructure?

I was recently involved with a UK client who have just cancelled their Infrastructure deployment project, which was originally going to be implemented as a "Big Bang" deployment.
Part of the reason for this was the underlying risk of such a disruptive deployment.
To put this in context, when I was first brought in to review the project, I realised that the new technology was so disruptive that it would actually be far easier and quicker to have a clean switch-over, rather than trying to incrementally upgrade the infrastructure. My view was that the risk could also be managed.
Since then, the Client's Business has moved on, the risk analysis was reviewed, and they decided to move forward incrementally. This will mean only getting (say) 40% of the improvements in the same timescale. Nevertheless, because of the changed business circumstances this makes sense, and so I supported the change of Strategy.
As an IT Production Consultant, I am generally unhappy with big-bang deployments. I prefer the gradual incremental approach which is more risk-averse, and more in tune with the culture of IT Production generally.
It would be interesting to have comments from other consultants in this area.