Dynamic controller assignment problem in software‐defined networks. Issue 8 (1st July 2018)

• Record Type: Journal Article
• Authors: Zhang, Bang; Wang, Xingwei; Huang, Min
• Is Part Of: Transactions on emerging telecommunications technologies. Volume 29:Issue 8(2018)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2018-07-01
• Subjects: Telecommunication -- Periodicals; 384.05
• DOI: 10.1002/ett.3460
• Languages: English
• ISSNs: 2161-5748

Abstract

Abstract: In large‐scale software‐defined networks, multiple controllers need to be deployed in order to guarantee network reliability and scalability in a cooperative way, where the assignment of multiple controllers and switches is very significant and usually consists of static controller assignment and dynamic controller assignment. However, static mapping among controllers and switches is not adequately efficient and realistic due to the dynamic network traffic load. To this end, this paper pays attention to and solves the dynamic controller assignment problem. We propose an adaptive controller assignment scheme for multiple controllers, and it can dynamically adjust the number of active controllers and the mapping relationship among controllers and switches by three algorithms. (1) When some controllers are overloaded, that is, the load ratio of controllers reaches a certain level, other controllers are activated to share the load of these controllers. (2) When some controllers are lightly loaded, these controllers are deactivated to reduce the control traffic overhead and improve controller utilization. (3) To further improve network performance, the load among controllers is balanced after reassignment. Simulation results demonstrate that the proposed adaptive controller assignment scheme has better performance than others in terms of flow setup time, control traffic overhead, and controller load balance, comprehensively. Abstract : We propose an adaptive controller assignment (ACA) scheme for multiple controllers. (1) When some controllers are overloaded, that is, the load ratio of controllers reaches a certain level, other controllers are activated to share the load of these controllers. (2) When some controllers are lightly loaded, these controllers are deactivated to reduce the control traffic overhead and improve controller utilization. (3) To further improve network performance, the load among controllers is balanced after reassignment.