Agent-based Architecture for Autonomic Web Service Selection

In principle, with Web services, application creation is simply a matter of finding and selecting the right services and composing them into a solution. However, current techniques do not address the problem of selecting the best or even a desirable service from among those matching a given interface. We propose an approach wherein middle agents serve as proxies for Web services to assist an application in selecting implementations that best match the quality criteria of the application. This approach supports the dynamic and potentially optimal selection of services and supports the natural formation of communities of agents helping each other evaluate the same services. This approach gets us close to vision of autonomic computing, wherein computational resources are selfmanaging and self-configuring. A prototype of this system is under development and evaluation.