A Game-Theoretic Approach for Dynamic and Adaptive Managers Selection in Service Specific Overlay Networks

Abstract

The introduction of Service Overlay Networks (SON) enabled service providers to offer new services with end-to-end quality of service guarantees. As a consequence, modern networks have become more service centric, which require the application of service specific management approaches. However, increased heterogeneity of network components and access technologies made management tasks more tedious and complex. In this article, we propose the use of game theory in the management tasks of overlay networks to reduce cost and complexity. We propose a fully distributed, self-organizing, and self-adapting algorithm for overlay nodes to efficiently select their managers. The proposed algorithm copes with the highly decentralized and dynamic nature of SON and adapts to changing network conditions. Extensive simulation results validate the effectiveness of the approach compared to existing solutions and the optimal solution.