Cloud Computing - What Does it Take?

I've recently read a number of articles and blogs on Cloud Computing, as well as marketing materials, and I find that there is ever more hype, and ideas, but not a lot of substantial "this is the way we do it" stuff. One idea presented by several is that Cloud Computing requires some form of brokerage functrion in order to be successful. See, for example, http://www.datacenterknowledge.com/archives/2010/01/22/cloud-computing-brokers-a-resource-guide/ or this press release http://www.gartner.com/it/page.jsp?id=1064712. The latter is also referenced in the first article.

For any type of online, automatic brokerage to actually work in a non-proprietary manner, many prerequisites must be fullfilled. This includes fully "machine readable" and standardised service definitions and service levels, automated and standarised security and deployment systems, as well as payment systems that can be trusted for transactions at an enterprise level. And with the promise (or hype) of Cloud Computing, all of this needs to function at a global level, where customers can take advantage of resources and spare capacity across the entire internet.

As I see it, there is a long way to go before this can get off the ground. On the service definition side, the SPACL consortium (see http://manag-e.blogspot.com/2010/01/can-service-catalogs-be-standardised.html for a few words on this) is heading in the right direction, but they have a long way to go. But there are no non-proprietary solutions that have come any further...

In the mean time, Cloud Computing will rely on more manual forms of contracting for services than the automated, brokered dream that a number of us share.