A search architecture for grid software components

Today, the development of Grid applications is a very difficult task, due to the lack of Grid programming environments, standards, off-the-shelf software components, and so on. Nonetheless, we can observe an emerging trend: more and more services are available asWeb Services, and can be linked to form an application. This is why we envision a market where developers can pick up the software components they need for their application. A natural process of evolution in this market will reward components that are faster, cheaper, more reliable or simply more popular. In this work, we present our vision of GRIDLE, a search engine for software components. It will rank components on the basis of their popularity, their cost and performance, and other users' preferences. We built a prototype of GRIDLE, which works on Java classes. It is able to give them a rank based on the social structure of Java classes.