Utility computing – websites like power outlets 8. July 2008 – 11:26 by Bengt Feil (TuTech Innovation GmbH)

The vast amount of different websites and services on the web seems to be impossible to categorize. Can we even call something like Google docs a website or is a more a less a piece of software running in a browser (which is an even broader term)? But there is a growing group of website that have distinct properties distinguishing them form the mass. These sites behave more like utilities (water or electricity) than a service – hence the term describing them: utility computing. These perform a simple task very well and provide the tools to others to build on top of it.

A popular and still recent example would be Twitter. This microblogging site gets much more powerful when services build on top of it are used than if one just uses Twitter itself. Examples of these added services are Tweetscan (a search engine tailored to Twitter), Twitterfeed (a tool that allow you to integrate your own RSS-Feeds to Twitter) or Twellow (Yellow pages for Twitter). Another example would be Google Maps - the service in of it self in basic but delivered in a very sophisticated and high performance way. But the additional services and the uses of Google maps by others really make it what it is.

The way this interaction between websites is made possible through so called APIs, short for Application Programming Interface. These are a set of rules which describe how the interactions have to be performed, which kinds of interactions are possible, etc. APIs power many different services on the web and also on local computers. The browser you are using most certainly interacts with your computers operating system through its APIs right now.

As with public utilities this kind of websites develop cottage industries improving and building on what is provided by them. But has these added services highly depend on the utility site there are some consequences. Utility sites have to be highly reliable and cannot change their service to much along the way. On a more technical note the API has to be consistent and well worked out. If a website chooses to go along those lines and the cottage industries takes of there may even be a new kind of business model attached to that (take a look at that idea at Mashable).

There seems to be a general trend to more focussed services on the web which interact among each other to perform more complex tasks. The rise of utility websites is also an indicator for this trends besides for example the discussion about data portability, initiatives like Open Social or apps build on social networking platforms like Facebook.

PS: For the more technical minded readers: I suggest to listen to the FLOSS weekly episode on the ROCKS where later on in the program the developers discuss using utility computing (in this case Amazon S3) to build large cluster computers (e.g. super computers).

Tags: APIs, Google, inenglish, Twitter, utility computing, web services