A self-sizing framework for adaptive resource allocation in label-switched networks

Abstract

For adaptive resource allocation to maintain its crucial timing requirements in large-scale networks, fast analysis and synthesis algorithms are required in order to process freshly collected traffic data. In this paper we build on previous work to introduce a general measurement-based self-sizing framework for resource allocation in label-switched networks. Furthermore, we study the scaling behavior of the simulated annealing variant used in the analysis and synthesis algorithm. We propose a hierarchical approach to improve adaptation performance for large networks and include extensive simulation results indicating significant improvement in performance as networks get larger.