A Cluster Approach to Scaling OpenFlow Control Plane
Intervenant(s) : Jin Zhao (Fudan University, China)
The centralized control plane in OpenFlow network has raised scalability concern in controllers, which may suffer from performance bottleneck for managing large-scale networks. Most of the existing controller designs are either a single node structure (with limited scalability), or in a distributed manner (with increased complexity). In this work, we seek to exploit the massively-parallel processing power of a cluster to address the bottleneck in current control plane of OpenFlow based network. We propose cCluster, a scalable OpenFlow controller architecture which can balance controller load across multiple servers in a cluster. cCluster is totally transparent to the existing OpenFlow switches and the upper layer applications. Our simulation results show that the proposed architecture is more efficient and scalable in terms of response time and controller scale, with only slight latency penalty when scaling up. We also implemented a cCluster prototype using commodity off-the-shelf PC hardware. The real-world evaluation results show that the load from OpenFlow switches can be well balanced, especially in the case of load bursts.
Short bio: Jin Zhao received the B.Eng. degree in computer communications from Nanjing University of Posts and Telecommunications, China, in 2001, and the Ph.D. degree in computer science from Nanjing University, China, in 2006. He joined Fudan University as an assistant professor in 2006. He stayed at University of Goettingen, Germany for 3 months as a visiting scholar in 2010. His research interests include P2P networks, media streaming and network coding theory. He is a member of IEEE and ACM.
stefano.secci (at) nulllip6.fr