Merging Sessions: Optimal Unicast Multi-Sources Routing in Communication Networks

摘要

Unicast routing algorithms transmit packets from one source node to only one destination node. In fact, corresponding to each destination node, the routing algorithm is separately performed once. If the number of destination node increases to n ≥ 2, the unicast routing algorithm must be performed n times, Repeating unicast routing algorithms causes conflict among separated unicast sessions and wastes time, bandwidth and network efficiency. Reducing the number of execution of routing algorithm can decrease time and bandwidth consumption and can increase network efficiency. By merging some sessions together to make a single session, we can reduce the number of algorithm execution. In this paper we present a new approach which combines n ≥ 2 sessions together to make a single session. It means that new approach for fulfilling n unicast sessions performs routing algorithm only one time and there is no need to n performances. In fact, our new approach solves the merged unicast routing problem instead of solving n separated unicast routing problems. The simulation results and numerical computations will show that our new approach as compared to available unicast routing algorithms has more data rate and efficiency and reduces computation time and bandwidth consumption. Our approach which is hased on linear programming, supports quality of services and multichannel paths, and we call it as: Optimal Unicast Multi-Sources Routing (OUMSR).