I seem to have been doing more conferences and meetings than anything else this year, and it's always fascinating to keep an ear out for what the latest buzz word.
Right now, it is cloud computing, but it is remarkable just how many people are using for a huge range of topics that may, or may not, be what others consider to be cloud computing.
Certainly its true that cloud computing means lots of things to lots of people. To me its the natural progression from what I was working on a few years ago, grids. Grids and cloud computing share many of the same features - large sets of computers that can be used to handle 'work'. The distinction between grids and clouds is that in a grid you tend to have a huge array of computers working on a single tasks or purpose, whereas in a cloud the computing power is a general resource that we can use for pretty much anything.
Others are applying the cloud computing to any large data center of computers. And here cloud computing can be applied, but again a distinct exists in that within the typical data center you would expect computers to be dedicated to individual tasks and to have a certain amount of administration/management overhead. Within cloud computing, the computers are a resource and you can change and redistribute work between machines in the cloud. There also tends to be more autonomy within a cloud than within a typical data center.
So are we all wrong about clouds? Not at all. I think it's clear that cloud computing, like clouds, is a wide ranging description that covers a range of potential uses. What one sees as a cirrus, to another is a cumulonimbus. Both are types of cloud, but having a different structure.
What remains to be seen is whether the current bandwagon of cloud computing will solve some of the wider issues that sit behind these types of installation. Issues such as how cloud computing can help with power consumption, cooling and general computing power issues are in many cases yet to be addressed, but they still apply.