- There is a total lack of use cases, yep SaaS, IaaS and such terms have been bonded around but where is the detail
- Management is being spoken about as if you can just wave your hands and suddenly you get a compute service delivered - but really, where are the overlaying / orchestration management tools or in fact a list of those which could be used
- Technical ratification of all of the major players? oops- that’s not there either (believe it or not – oracle does account for something like 60% of my estate – so I would like to be able to use whatever I deploy in conjunction with this tool)
- If I deploy this nice new shiny cloud along with tools – i now sit in the middle of even more tool sets rather than consolidation - I can’t get rid of my existing tool set as i still have to manage Sparc / AIX / HPUX resources and now i am getting additional toolsets thrown at me with no roadmap or view into how I am going to solve this.
- Open API's / ways of integrating.. you know what XML is a great way of letting me "do stuff and manage" to a compute environment but i still need something to orchestrate this... i am starting to feel that open api's is a new way of the vendors saying that the orchestration bit is just just a little to hard and leave it to the end-user... This is a fine approach – but also suggest a way forward / management framework along with tools.
- Overstated "easy to do" configs... A number of these have been presented at very large forums, an example being Virtual Hosted Desktops bounded around as a no-brainer... I personally have just been involved with / at the back end of a significant deployment - and it has been by far, one of the most difficult things i have done (if you want some details and guidance - please shout) – don’t buy this as being easy...
- Alliance's, so i am not sure what is going on here... I have alliances, new partner agreements and all sorts of stuff spinning up – its coming out of my ears to be honest... If you don’t have an alliance on the way, apparently you are “behind the times....” – but what do they really mean??? Even basic interoperability sheets are yet to be fully populated, along with application & ISV support.
- Go to market model... Is there really a go to market model, other than loads of salesmen trying to sell me bits of stuff that they don’t understand? In the worse case I have actually had someone come up to me, and say - do you want to buy a cloud... I mean... come on vendors - please sort this out...
- And by the way - when you come and sell this stuff to me, I really don’t expect to tell you how your solution is being sold, how an alliance is being formed and what you can and cannot do with the infrastructure - surely that’s your job – (and before anyone asks - yes this did really happen!)
Thursday, 28 January 2010
So... this term cloud computing is being bounced around, alliances springing up everywhere and apparently there is a paradigm shift to this new cloud services model... but you know what, i think it should be called fog (or foggy) computing
Why - well reasoning goes something like this:
Also – whilst it is sold as simplifying environments – this may be true for those doing the provision after infrastructure has been built, in truth – the underlying complexity and layers of abstraction is really quite scary... Performance fault finding as an example – has just become orders of magnitude harder...
So back to the title - why foggy computing - well, it’s sort of unclear, un-instrumented, undefined, unworkable and quite frankly- unreal at this time..