Dynamic SDN Controller Association Mechanism Based on Flow Characteristics

In software-defined networking with the distributed control plane, the switches are associated with controllers using only the number of the flow requests as the basis of resource allocation, ignoring the characteristics of flow. To address this issue, we first analyze the control resource consumption of flow requests with different characteristics, and point out that the control resource consumption can be reduced through changing the controller association relationship. Then, we propose the dynamic controller association mechanism based on flow characteristics and design the greedy set coverage algorithm with the coalitional game to reduce both control resource consumption and control traffic overhead. The mechanism is divided into two phases: first, the greedy set coverage algorithm provides a fast algorithm to yield a controller association scheme but leads to an unbalanced association. Based on this algorithm, a coalitional game strategy is introduced to achieve the near-optimal association scheme between controllers and switches through switch migration. The extensive simulations demonstrate that while keeping control traffic overhead low, our mechanism can reduce control resource consumption by about 36% in comparison with the controller load balancing mechanism.