Many IT infrastructure decisions are made for reasons which have little to do with the capability of the technologies and very few are even made with due consideration of investment returns, long term costs and even fewer are revisited with the light of truth shone upon them.
So it is a wonder that any IT infrastructure works at all?
Well not really, as we have moved into a pretty much homogenised environment where all is interchangeable and pretty much all is good enough; the decisions are going to be made for reasons other than technology.
Many decisions are made simply are the grounds that more of the same is the path of least resistance. You have already learnt to love what you hated about a product and you are comfortable with it. You might have grown close to the account team, they know all your favourite restaurants and sporting events; why change? And change is costly.
Of course, then you get the obverse; you have learnt to hate what you loved and the account team has grown far too comfortable. Perhaps there’s been a change in account manager or simply you decide that you’ve spent too much money with a company. Of course at this point, you suddenly find that what you have been paying is far too much and the incumbent slashes their costs to keep the account. But you’ve had enough and you decide to change.
Then you get the principled decision; the decision which could be based on the belief that open-source is the right thing to do or perhaps you believe the security through obscurity myth. Sometimes these look like technological decisions but they are really nothing to do with technology in general.
So have we moved to a market where the technology is pretty much irrelevant and why?
I think that we have and for a pretty good reason; you can’t manage what you can’t measure and quite simply, we are still lousy in measuring what we do and what it means. It means that all decisions have to made based on reasons which often have dubious links with reality.
For all discussions about metering and service-based IT; I don’t believe that we are anywhere near it. Internal metering tools are often so expensive and invasive to implement that we don’t bother.
And what is worse, we are often working in environments which do not care really care; who really cares if solution ‘X’ is cheaper over five years than solution ‘Y’ as long as solution ‘Y’ is cheaper today. Tomorrow can look after itself, tomorrow is another budget year.
So not only is measurement not easy; perhaps we simply don’t care?
Perhaps the only option is just carry on doing what we think is as right as possible in the context that we work in?