Seeing Through the CloudBy Tony Kontzer | Posted 02-23-2009
When Howard Rubin speaks, people listen. And when the subject is cloud computing, today's IT leaders--many of whom are attempting to chart nascent cloud strategies--are bound to lean forward a little farther in their chairs.
CIO Insight contributor Tony Kontzer recently spoke with Rubin, an IT benchmarking guru, professor emeritus at Hunter College and a senior adviser to analyst company Gartner, about his developing theory on applying the concept of the commons--a piece of land for which certain rights, like grazing, are granted to people other than the owner--to solve some of today's computing challenges. This is an edited version of that conversation.
CIO Insight: You point out that 85 percent of the world's computing capacity sits idle at any given moment, consuming as much as $250 billion worth of power per year. How did IT waste reach such staggering levels?
Howard Rubin: One way to look at it is that servers are the SUVs of computing. Servers came along when companies were flush with money, and until people became environmentally aware and virtualization became an alternative to up your resources, they were an easy platform to grow. It wasn't really allowed; it was just sort of natural. When you get new economics and an understanding of what's going on with computing and the environment, you learn and you start taking action.
What caused you to draw a parallel between the concept of the commons and the evolving IT landscape?
Rubin: There was an article in The Economist a while ago about the concept of the commons. The interesting thing that's happening now is that the economy is driving economies of scale. Companies need scale beyond what they can reach to be competitive.
If you look at what's evolving in the market, you have a kind of e-mail commons. Gmail is an example of a service that can be supplied as a cloud by a vendor. Those large-scale services can be used as commons by large companies. Why can't we build megascale platforms for common on-demand computing?
What needs to occur for an IT commons to become reality?
Rubin: Pooling and sharing of resources always are accompanied by barriers. People have to drop the barriers to playing in the sandbox with partners that are in or out of their industry. Everyone needs e-mail. Everyone needs processing. It's breaking down the barriers of sharing, it's breaking down the barriers of having service-level requirements that aren't needed, and it's building service commons that don't exist today.
How would a CIO fearful of anything new without a service-level agreement be persuaded to join in a technology "greenfield"?
Rubin: You can establish different levels of commons. It's really just looking at things in new ways, and always looking at optimizing based on the size of the market and the unit price. CIOs have to get over the belief that the market gives crappy service levels, which it doesn't. The crappy service levels are inside companies themselves.
Would a technology commons need to take into account the difference between internal and external clouds?
Rubin: Companies are always going to have to look at the marketplace commons versus the commons they can create. They'll have to determine if there are service levels they need that can't be had on the open market. If you go far enough out with this concept, you'll see absolute blur. The companies will be relying on a commons, and there will be very little internal [IT]. The economics won't make sense.
With the clouds out there and standardization, a company that's doing it themselves won't be able to compete with companies that are doing it out there. It's like having your own generator for your company versus buying into a grid.
What efforts are currently under way to help push the development of a true technology commons?
Rubin: The pioneers are out there, and they are giving us the models--the Googles, the Amazons and the IBMs, if you go back to older models. The psychology of sharing is starting to take hold. It's sort of techno-Darwinism.
Things like the cloud stuff, even the Internet itself--which is absolutely a commons in its own right--take that into the world of processing, and companies learn to share and leverage external resources. I don't know what it's ultimately going to look like, but the market forces are going to make it happen.