Reputation Based Selection for Services in Cloud Environments

Distributed environments hosting specific and dedicated computing resources are serving nowadays more and more scientific and commercial applications. Finding the right resource in distributed environments in term of performance is not an obvious task. This paper is addressing the issue of selecting the right service deployed in a distributed environment for any applications or workflow engines that need access to best service´s endpoint in term of performance. The service selection is made according with a reputation function, defined in this paper. The reputation function considers the execution cost as a performance criterion. The experimental results highlight the behavior of proposed reputation function and the comparison with other reputation functions. The papers present the possibility of enhancing the proposed solution in a real environment represented by a workflow engine based on Business Process Execution Language (BPEL).