Accounting for link load correlation in the estimation of blocking probabilities in arbitrary network topologies

In this paper, we discuss an efficient analytical framework that can accurately account for the link correlation in generic-topology circuit-switched networks. The model takes all active paths, as well as all free links, as independent objects on the network topology. Therefore, the actual sets of sequentially connected busy links (paths) and free links are seen as they appear in the network and the link correlation can be straightforwardly considered. Comparisons with results obtained from simulations show the excellent performance of the proposed analytical modeling for a variety of network topologies.