Cost-Efficient and Scalable Multicast Tree in Software Defined Networking

摘要

Multicast can effectively reduce the cost of network resources, and Software Defined Networking (SDN) makes Steiner tree a feasible and promising way for multicast. However, multicast still suffers from a scalability problem when the number of groups is large since the flow table size is limited. In this paper, therefore, we propose the Degree-dependent Branch-node Weighted Steiner Tree (DBWST) problem, which is NP-hard. This problem aims to minimize the total cost of edges and branch nodes. The cost of a branch node is degree-dependent. We design an approximation algorithm, named Path-Vector based Harmony Search Algorithm (PVHS), to solve this problem. The path vector means a solution vector in a harmony and denotes the ordered set of nodes from source to a destination in the multicast tree. Globle and local optimization are combined appropriately. Simulation results on randomly generated topologies indicate that the trees obtained by PVHS are more cost-efficient and scalable over the existing ways.