Recently, I presented to the UK CMG’s London conference on the subject of outage prevention - you can take a look at the presentation speech here. CMG is an independent industry forum focussed on performance, capacity and service management. As such, its conferences attract a range of IT professionals, from capacity planners through to problem management teams, all of them essentially focussed on keeping services running, and running well. It’s a great audience from which to gauge the state of the nation on how service management is really done in large organisations, and I distilled some interesting observations from talking with conference delegates and presenters
Firstly, an interesting statistic quoted on ITIL was that only 30% of people who attend training courses on the subject, subsequently see it implemented in their companies. If accurate, what does that say about ITIL’s practical effectiveness in real life and what does it say about the maturity of IT in our organisation? A subject worth returning to in a later blog, I think.
Secondly, listening to discussions in the capacity planning forum reaffirmed that most real-life capacity forecasting is still done in Excel and involves drawing lines through points on CPU charts etc. The great divide between IT and the business is alive and well, so that few organisations use business demand forecasting to drive infrastructure planning, and the whole thing generally still relies on pretty crude techniques.
There was also a great presentation from Sysload on the many and subtle difficulties of predicting and measuring performance and capacity on virtualised systems, (we also covered virtualisation issues from a different perspective, in this blog last year). Sysload seems to have some great answers to some of the challenges, but it’s clear in other discussions I’ve had with clients that there’s a lot of hesitation around virtualisation generally, because of the unknowns it introduces into critical systems.
Combined with the above two points, this had me wondering just how many other IT cost-saving opportunities can’t get executed in practice because of poor control and visibility leading to unacceptable risk. Given the period of the economic cycle that we’re now entering, that’s a big opportunity lost. It would be much improved by a little more insight, of the type we regularly discuss in this blog.
