A Performance Analysis of Distributed Algorithms in JavaSpaces, CORBA Services and Web Services

Implementation of distributed parallel algorithms on networked computers has always been very difficult until the introduction of service-oriented architectures (SOA) like JavaSpaces service, CORBA services and Web Services. Algorithms of the type Master/Worker pattern are implemented with relative ease using the SOAs. This project analyzes the performance of such algorithms on three contemporary SOAs namely JavaSpaces service, CORBA services and Web Services. These architectures make the implementations of distributed algorithms reasonably fault tolerant and highly and dynamically scalable. Also, the systems built on these architectures are generally loosely coupled and operate asynchronously. In this project we measure and analyze the latency, speed-up and efficiency metrics of an insertion sort of 0 (n^2) complexity on all the three SOAs. We then draw conclusions of overall performance and scalability on all the three architectures