Optimal Multicast Control for Simple Network Coding

We propose a network coding strategy for multicast applications called the simple network coding scheme, which takes network coding chances raised among adjacent nodes. The proposed scheme incorporates both intrasession and intersession network coding strategies and effectively improves multicast throughput. We characterize the capacity region of the proposed scheme and derive an optimal control algorithm for the proposed scheme. We perform a complexity analysis for the proposed control algorithm and provide some insights into its practical implementation. For a comparison, we also formulate the capacity region of conventional multicast schemes and provide performance evaluation using a linear programming solver. In empirical analyses, we investigate how the proposed scheme improves multicast throughput gains from various perspectives (i.e., the number of flows, the number of network coded packets, and split–multicast) and find out that most of the gains can be achieved by simple pairwise network coding with non-split–multicast. We observe dramatic throughput gains up to 75 $\%$ beyond the conventional schemes on random topologies with ten nodes.