Towards a biologically-inspired framework for multimedia service management

Abstract

The advent of service-oriented architecture (SOA), internet and ubiquitous delivery technology bas resulted in multimedia services (e.g. repurposing, streaming and conferencing services) being accessible at any time, from any devise, through any network. However, there are still some problems related to heterogeneity, scalability and QoS demand of the management of such multimedia services. Some of the existing solutions are centralized, which evolve scalability problems in terms of the number of concurrent requests for the target service composition. Other solutions are distributed, which depend on the use of traditional algorithms (e.g. Dijkstra, Bellman Ford). Such distributed solutions also use replicated services, which can also result in scalability problems for large networks. In order to mitigate the above problems, this thesis proposes a framework for multimedia service management that is based on a biologically-inspired approach. It utilizes an ant-colony-based selection algorithm for collecting the QoS requirements from the individual repurposing service in order to select the most suitable one for the desired composition process, which ensures higher scalability and efficient load balancing. It also develops a QoS-aware service selection algorithm for a multimedia repurposing service. The proposed framework's performance is validated through both simulation and prototype implementation.