Note: You are experiencing only the raw content of this site, without the intended layout and design. Either your browser has ignored the Cascading Style Sheet (CSS) files for this site, or you are using an outdated browser which does not support Web Standards. Learn more.

EDS' Next Big Thing Blog: Read and Respond to What the EDS Fellows Say About Technology

Read and respond to what the EDS Fellows have to say about the future of technology on EDS' Next Big Thing Blog on eds.com.

What should the price of utility computing be based upon?

by Charlie Bess

I mentioned a while back that I am going to be at Purdue this week and I spent some time talking with a number of folks about a variety of topics.

One of the people I talked with works in the power management space and we got into a discussion about how organizations charge for CPU time and utility computing. He suggested that compute cycles might not be the best way to charge for usage in the future. The computer processor is only a small portion of the operating expenses anymore and yet its usage is the current focus of measurement. Maybe we should charge for the power consumed in the process. Why should an I/O bound job get charged less when it actually may consume many more watts of power, spinning all those disk drives?

This would be a big shift in the view of charging for value delivered and it may not be accepted by the market, but it does make sense. It would also take a different approach to measurement in the data centre.

Published Friday, February 10, 2006 1:50 PM

Subscribe to this post's comments using RSS

Comments

Posted by vinnie mirchandani Monday, February 13, 2006 1:47 PM

The biggest cost in an "IT utility" is people cost. That needs to get automated more, lowered through shared services, remote services, scale etc.

EDS has over 100,000 employees. The average Fortune 500 CIO has 500 IT employees. If EDS has that much more scale and still is 2-3X cost of an internal employee, something is off in the industry.
as I wrote in this post

http://dealarchitect.typepad.com/deal_architect/2005/10/utility_computi.html

Posted by Charles Bess Tuesday, February 14, 2006 1:39 PM

I agree with the thought that the biggest cost is the people, unfortunately with the energy concerns one of the costs that large IT organizations have the least amount of control over is their power consumption (as long as they take a passive approach).

I am not sure why you would say EDS or other organizations in this space are 2x-3x the cost of an internal employee - you must be making some kind of apples to oranges comparison without rolling up all the costs (e.g., benefits, corporate risk), so I will not be able to talk about that effectively here.

I thought you were going to go down the path -- I can understand why power companies charge by the watt. They are in the power business. Utility computing organizations should be providing business value, not power. Shouldn't they be charging based upon the amount of business value delivered. Oh well, I guess I'll wait for someone else to bring up that point.

Posted by vinnie mirchandani Thursday, February 16, 2006 3:44 AM

Charlie, $ 100 an hour is a very conservative rate for most US or European firms times 1800 hours a year - that double of what the average IT staff makes, even after burden and taxes etc. Many firms charge much more and also for overtime ...

Posted by Bert Armijo Friday, March 10, 2006 7:19 AM

While interesting at an intellectual level, such dscussions are really premature. Only recently have we begun to see hardware and software designed with the intent of providing utility computing so our vision of such services are likely to change completely over the next few years.

For instance, the suggestion here assumes that my application and yours are segregated in a way that makes power measurement meaningful. However, services in which that isn't true are already being tested.