Abstract
Due to the rapid growth of Software-Defined Networks, we can manipulate network traffic easier than before. One of the major issues on SDN is the multipath routing protocols. In this paper, we focus multipath routing algorithm on Multipath TCP (MPTCP). MPTCP is an extension of TCP that increases the throughput of TCP communication significantly by utilizing multiple paths transmission rather than single path. In other words, an MPTCP flow generates multiple sub-flows for end-to-end communications. We consider the multipath routing problem as a k Max-Min bandwidth disjoint paths one. The problem is to find k disjoint paths with relative higher throughput and the smallest bottleneck bandwidth of the k paths is the maximum. Since this problem is NP-complete, we propose a heuristic algorithm to solve this problem in polynomial time. The simulation results show that our proposed algorithm perform better than previous works in terms of average throughput and average hop count.