Emulation Performance Study of Traffic-Aware Policy Enforcement in Software Defined Networks

Modern networks require robust traffic handling policies in order to minimize unwanted traffic and maximize performance. These policies are enforced by the placement of rules on network devices such as routers and switches. For high-speed processing, the rules are stored in Ternary Content Addressable Memory (TCAM) [3]. Because it is expensive, there is only a limited amount of TCAM on each network device. Software Defined Networking (SDN) [2] and the OpenFlow [1] protocol provide the capability to control the way rules are generated and placed within a network. This allows researchers to design algorithms that control TCAM usage while optimizing performance. The goal of this project is to develop a test bed for researchers to easily implement and evaluate their performance-optimizing algorithms. In this paper, we describe a test bed that emulates SDN activity and captures network performance data for rule placement algorithm evaluation.