by April 20, 2011 0 comments

Andy Mulholland, CTO, Capgemini

Disruption is beginning to beat innovation as the most used word to describe what is happening. My belief is that the scale of the shift in the way we do, or will, or are, doing things is starting to become all too apparent. The term ‘innovation’ sounded much cosier, just something really interesting we found in a corner of our market or business, something that felt much more like an evolution in truth and didn’t threaten business as usual.

The latest potential for mass disruption appeared in a quiet way in The Economist under the title ‘A Market for Computing Power’ to launch SpotCloud , the first spot market for buying and selling computing power. It’s the normal concept of a spot market in matching surplus from a seller to shortage from a buyer, in this case just simple computational power. At this point you may recall that Amazon Web Services moved to create ‘spot instances’ some months ago, but this is different and the implications are profound.

Amazon Web Services are in the business of providing computational services, they are all set up and organised for it, just as you might expect for an enterprise that makes its money out of providing ‘services’ of this type. But that’s not what SpotCloud is all about, it’s using technology to allow any IT shop in theory to sell off any unused capacity, or any IT shop to buy and use that capacity. Neither buyer nor seller will be visible to each other, it’s all anonymous, with SpotCloud managing the transactions via Google App Engine.

The buyer of the computing power places their work as files on to the CloudSpot App Engine space and SpotCloud sends them to the seller’s servers to run. And that’s where the disruption comes in. The simple mechanism by which this has been created and works illustrates several things about where we are moving to;

  • First — just how simple, universal and common a ‘lego’ style approach has developed already, and frankly is going to go on becoming.

  • Two — how much the nature of the work is changing away from monolithic predictable and sizable stable applications towards ‘task based’ variable activities.

  • Three — that it really is time to recognise that the fundamentals in the supply and use of technology enabling elements are just on the tipping point of change.

It’s that last point that is the real disruption, the question to be considered, and shows SpotCloud as being part of the answer. After all it’s not likely that monolithic enterprise applications are going to be supported via SpotCloud!

If all those who have announced their intentions to build and offer cloud centres do so, and I am thinking particularly of new entrants such as telcos, there will be an over supply of capacity. More importantly it will be ‘bare metal’ capacity suited to this kind of use; think of the same situation around the installation of fibre optic bandwidth at the same stage of the market with PCs and networks as the disruptors leading to an over supply of ‘dark fibre’. On the other side think of all the users and devices, tablets, smartphones, sensors, etc, growing at a tremendous rate with the consequential rise in all the ‘apps’ they will script and consume. Together they constitute a huge new market place of buyers and sellers.

Now here is the tricky bit; with these points in mind take a careful look through all the predictions for 2011 to 2013. Just to make it simple here is Charlene Li’s library of predictions collected from across the industry, and most important here is Charlene’s very thought provoking blog ‘hot or not; Disruptive Technologies to watch’ . Charlene’s point is the lack of reasoning provided by various pundits as to why they think particular technologies are ‘hot’ and disruptive.

My build on this is to point out that individually they may not prove that disruptive, and probably can be handled within the current concept of ‘some innovation’. Collectively though, they combine to really make a total disruption in the use of computational power in terms of access, use, management, and therefore provisioning and payment. With this in mind go through the trends 2011 list and ask how many could you enable and support from your current ‘customised’ computational capabilities.

No Comments so far

Jump into a conversation

No Comments Yet!

You can be the one to start a conversation.