Service Decomposition and Task Allocation in Distributed Computing Environments

Abstract

Abstract.Highly competitive and open environments should encompass mechanisms that will assist service providers in accounting for their interests, i.e., offering at a given period of time adequate quality services in a cost efficient manner. Assuming that a user wishes to access a specific service composed of a distinct set of service tasks, which can be served by various candidate service nodes, a problem that should be addressed is the allocation of service tasks to the most appropriate service nodes. This scenario accounts for both the user and the service provider. Specifically, service providers succeed in efficiently managing their resources, while users implicitly exploit in a seamless way the otherwise unutilized power and capabilities of the provider's network. In general, service task allocation is founded on general and service specific user preferences, service provider's specific service logic deployment and current system & network load conditions. The pertinent problem is concisely defined, mathematically formulated, optimally solved and evaluated through simulation experiments.