RPA-RA: A Resource Preference Aware Routing Algorithm in Software Defined Network

Abstract

In Software Defined Network(SDN), every switch contains multiple flow tables, each flow table containing multiple flow entries. The number of flow entries determines the number of flows that can be forwarded by the switch. Link resources(like bandwidth) is no longer the only one to be considered when routing because node resources(like flow table) may also cause network bottlenecks. Moreover, from a view of network traffic, different flows have different preference to these network resources. In this paper, we use Analytic Hierarchy Process to analysis the characteristics of network traffic, then based on K-shortest path algorithm, a novel Resource Preference Aware Routing Algorithm, simplified as RPA-RA, is proposed. RPA-RA selects the most proper path for a flow from k shortest paths by considering the resource preference of the flows. The aim is to balance the usage of link bandwidth and the flow table simultaneously when routing, so that the network can accept and process more flows. At last, our algorithm is validated with POX and Mininet. The simulation results show that RPA- RA can balance the usage of the flow table and bandwidth, result in fewer resource bottlenecks, and make the network accept more flows relatively.