It's interesting how this title gets a direct to comment effect. If I may summarize the post, I am not asking what ITIL is, nor what it has. What I am asking is, if you are running an autonomic million core data center that is based on Google's three rules, then does it matter as much (i.e., really matter). Like I said in the original question, I don't know. Should there be a hardware change process in a 200k square foot, dynamically orchestrated server environment, that has less than 20 people in the building where computing power runs like an electricity grid? In fact, does a model like Six Sigma become more of a play in an environment that?

I'm just play devils advocate here. I read your blog and know you are very knowledgeable..

If a server fails and the system autonomously self-heals, do humans need to do anything? Do we need to know it happened? i think so. So there was an incident. The fact it self resolved does not mean we don't still want the incident recorded, and sommeone may need to do something if only after the fact.

These are clusters of anywhere from 100 to a 1000 commodity systems where the work is processed in parallel and failure is expected (google's three rules).

What if they keep failing?

They don't. They throw the cpu board away (google's three rules).

+At what point does the resulting degradation impact service delivered? Who do we tell and when?

Will we ever make a change to the environment? Will we ever replace a main power component? Upgracde the operating systems? Do humans need to consider and approve this change? I think so.+

I am not sure if at this point we are moving into the engineering domain and again maybe be something like Six Sigima is more appropriate. If you look at a GE Aircraft, Nuclear, or a Motorola plant you find a lot less ITIL and a lot more Sigma. The real question is will IT make the "Big Switch" as Carr says, and if so does IT become more like a utility (electricity, water, gas, ...). I know ITIL wants to grow beyond just IT and if it does it might still make sense even then.

With all that said, none of this is going to happen to the mainstream anytime soon. However, hears the rub, IMO the true ITIL adoption rate has been very slow among E5k customers and if it takes to long with ITIL miss it's window as a nother failed methodogy?. It's like what Mark Twain says about the weather... Everyone's talking about it but nobody does anything about it.

How do we predict growth of the farm? On what basis do we order new hardware? At which point will the farm grow beyond the absolute capacity of the technology? of the budget?

I see less than 20% of my E5k customers even using ITIL in a serious capacity and I have never seen any of that 20% anywhere near doing capacity planning as part of ITIL.

Just my 2 cents

You got me...

If it's ITSM we are both out of a job. Like I said I posted the question to try and get some real debate on the subject. I work with a lot of the Tivoli products and I posted this question on the Tivoli mainling list and the debate went on for almost a month. The problem was most of it was complaints on how ITIL is being pushed down their throats. My main area of expertise is not really ITSM however it is more focused on a subset - ESM. However, I have a fond interest in how utility computing might effect both.

johnmwillis.com