A mixed integer programming model for multicast routing with multipath delay jitter reduction

Multipath routing is the technique of assigning multiple paths to destinations which is beneficial to satisfy QoS requirements, especially when there are strict delay requirements and violation from Service Level Agreement (SLA) leads to significant revenue loss. However, it has not yet analytically been analyzed that how multipath routing can improve different QoS parameters. In this paper, first we propose a mixed integer programming model to minimize reservation costs while satisfying end-to-end delay constraint by letting to reserve at most two redundant paths for each multicast destination. Since the mathematical model is complicated and highly nonlinear, we propose a heuristic algorithm to find redundant paths to satisfy delay requirements. To guarantee delay jitter reduction, redundant paths must be maximal disjoint-minimal delay different. We use simulation experiments to survey the effect of multipath routing on QoS for different network topologies and parameter settings and compare results to show performance of our proposed model.