Partitioning Application Using Graph Theory for Mobile Devices in Pervasive Computing Environments

Abstract The very first phase in software partitioning is to choose an appropriate clustering methodology before moving the clusters to multiple computational nodes since this stage can impact on the distribution of clusters and eventually on the performance of the overall application. In pervasive computing environments, certain memory and cpu intensive applications would prefer to use the computational nodes (anything with at least networking, storage and cpu capabilities) available in the environment. To this end, this paper investigates the very specific area of clustering or partitioning an object-oriented application running on mobile devices. Graph theory has been used to model an application and a cluster analysis algorithm has been proposed. Details about the implementation are covered, followed by a laboratory application partitioning using the proposed approach. Researchers and software designers willing to investigate software partitioning can consider this practical and easy implementable approach.