Int J Performability Eng ›› 2018, Vol. 14 ›› Issue (8): 1848-1856.doi: 10.23940/ijpe.18.08.p22.18481856

• Original articles • Previous Articles     Next Articles

Sequenced Switch Migration for Balancing Controller Loads in Large-Scale Software-Defined Networking

Lan Yao, Tao Hu, and Julong Lan   

  1. National Digital Switching System Engineering R&D Center, Zhengzhou, 450002, China

Abstract:

The proposal of the multi-controller has improved the scalability and reliability of software-defined networking (SDN), and the entire network is divided into several subdomains with the self-governed controller. Due to the dynamic changes of network traffic in different subdomains, a new challenge has emerged for balancing loads on the distributed controllers. Switch migration, a promising method, is designed to solve the unbalanced distribution of controller loads. However, existing schemes implement switch migration without considering migration sequences and are characterized by long migration time and poor controller processing efficiency. To solve the above problems, this paper proposes the Sequenced Switch Migration (SSM) approach, which considers migration parameters (e.g., delay, traffic, residual capacity, and failure probability) for multi-objective optimization and achieves efficient switch migration using a sequenced manner. By executing the parallel and heuristic algorithms, SSM can reduce the execution time of migration while ensuring the performance requirements of controllers. The simulation results show that compared with the existing schemes, SSM reduces the total migration time by about 32.7% and decreases controller load variance by 28.5% when controllers are under high load condition.


Submitted on May 16, 2018; Revised on June 13, 2018; Accepted on July 24, 2018
References: 17