Summary Until now Ajax has only been used to implement Rich Internet Applications, however Ajax can and will be used for more types of applications. I will demonstrate in this article the use of Ajax in a Distributed Computing environment by a simplified example, distributed link validation.

Ajax does broaden the application field of distributed computing. You certainly will have been thinking that Ajax can’t be used for critical information, such as business forecasting, or applications where assemblers are needed. And you’re right. You should not use Ajax here but another distributed computing method.

In the other cases however you should always consider Ajax since it is extremely user friendly. Although the technology to develop a giant supercomputer by using the idle cycles from normal internet computers is already available for a couple of years now, the biggest limitation has always been the human nature. The modal user doesn’t care about scientific investigation nor does he care about business forecasting. And he certainly will not download special software to contribute to these fields.

But neither is he against the use of the spare computation time. As long as it will pass by unnoticed, but of course with his permission and without extra work for the end user, he will not care if there is a guest application contributing to research. And this is the strength of Ajax. Typically, there will be minimal interaction between your site and the user, leaving the processor unexploited. But with Ajax you can run a process on the background while the user is reading on article on your site. If you play it ethically right, this has very minimal implications to the end user, but can deliver you a lot of processing time.

And there is another big advantage of Ajax, broadening the field of distributed computing. It is based on JavaScript and XML and as you know JavaScript is optimized for a network environment. To fully understand the implications, I shall look deeper in to the second condition of a good exploit the coarse-grained parallelism, namely a “high computation-to-communication ratio”.

read more...