Cloud - What else is needed?

To say there has been (and continues to be) a buzz around Could computing is probably the understatement of the century! Almost everyone I talk to has something to say about it and usually an opinion on what it all means for the future of computing and IT.

A colleague recently said that that he could take a £10 note and write 'Cloud' on it and then sell it for £20 to an array of waiting venture capitalists. Certainly the array of Cloud events, press-releases and industry conferences testifies to the fact that there must surely be something behind the buzz.

Another colleague recently made the transition from Grid to Cloud and even the BCS's old Parallel Computing group has been resurrected and renamed under the catchy title of 'Distributed and Scalable Computing Applications and Services Specialist Group'. So all eyes are focused skywards into the Clouds, but what's really going to happen next?

My opinion is one shared by many others out there who are busy gathering up the component parts to make a public Cloud really work. What is a public Cloud?

Well to me it is a real marketplace where I have a choice as to where, how, who, when and for how much my computing needs are satisfied and preferably in a dynamic enough way that I can switch as my needs change without a heavy penalty.

This really means that I'm going to set a policy based on my selection criteria and a resource broker is going to do the rest for me. Okay, so this only satisfies those things that are considered asynchronous today as synchronism implies a more rigid, more real-time approach.

The primary point though for anything to take-off rather than be confined to the history books as a nice or a failed experiment is that a market is created where people have a choice, where suppliers can complete and consumers can select based on their needs, which more often than not now, can vary wildly with time.

So, a resource broker is one component, those exist already today but what else?

Well, the ability to understand your costs if your are a supplier of resource into a public cloud marketplace is pretty important. Not just direct fiscal cost but also environmental cost, carbon. What? You don't think cloud customers of the future won't ask you how many megawatts, how efficiently and how many kilos of carbon it's going to 'cost' to service their requirement?

Of course they will! It will be one of the elements the policy based broker will have to cope with. 'I want to run service this compute need for the minimum possible carbon footprint because the carbon tax in my country will offset any marginal increase in service cost by a factor of two'.

Maybe this seems silly to you now but in the UK we already have carbon budgets set by the government and they are coming to a company near you shortly! The direct 'tax' may be punitive to begin with but there are other things to consider such as impact on brand value that could be much worse.

Could computing as some of my colleagues at the DCSG say is the future, it makes sense from an efficiency perspective, it makes sense from a cost perspective but there is still some way to go before it makes sense from a 'business perspective' for all but a lucky few.

Image from: http://www.flickr.com/photos/donabelandewen/470780785/