Bidirectional Matching Strategy for Multi-Controller Deployment in Distributed Software Defined Networking

IEEE Access

卷:6

Peng Yi¹  Julong Lan¹  Zehua Guo²  Tao Hu²  Jianhui Zhang³ 

[1] Technology Research Center, Zhengzhou, China;

[2] National Digital Switching System Engineering &x0026;

[3] University of Minnesota, Minneapolis, MN, USA;

关键词: Software defined networking;    OpenFlow;    multi-controller;    load balancing;    matching;   

DOI  :  10.1109/ACCESS.2018.2798665

来源: DOAJ

【 摘 要 】

Software defined networking (SDN) realizes the full decoupling between control plane and data plane, and it could effectively simplify network management and promote network innovation. In the large SDN network, there are the single point of failure and the lack of control resources for one centralized SDN controller, so several researchers have presented deploying multi-controller to improve the reliability and scalability of the network. However, unreasonable deployment of multiple controllers may cause the unbalanced distribution of controller loads, and it is harmful to the network stability. In order to address the above problems, this paper proposes bidirectional matching strategy (BMS) and designs its corresponding algorithm to implement multi-controller deployment. First, through collecting the information of hops, delay, and traffic in the network, we build the matching lists of switches and controllers, respectively. Then, according to the specific sequencing of elements in the matching lists, we design a mutual selection strategy that the switch and controller will select the optimal elements from its list to implement matching in turns until realizing the construction of the distributed network. Meanwhile, we implement the strict mathematical derivation to demonstrate the feasibility of BMS. Finally, simulation shows that BMS can achieve the better multi-controller deployment when compared with the existing algorithms.

【 授权许可】

Unknown