Model for Controller Assignment and Placement to Minimize Migration Blackout Time with Load-Balancing Platform in Software-Defined Network

A software-defined network (SDN) is a network that the centralized SDN controller controls multiple SDN switches. Load-balancing platforms can realize the distribution of the load of the switches between multiple controllers. The platforms allow controller processing capacity to be used efficiently. When the assignment between switches and controllers and the controller placement are changed, migration blackout time that the controller temporarily stops processing messages can occur. The migration blackout time can result in failure to meet delay requirements between switches and controllers. This paper proposes a model that determines the controller assignment and placement while minimizing the migration blackout time with the load-balancing platform. The proposed model can be used when the controllers in the network are overloaded and the controller assignment and placement need to be changed. We formulate the proposed model as a mixed-integer second-order cone programming problem. We develop a migration procedure used in the proposed model. In the procedure, each switch can be controlled by multiple controllers with a load-balancing platform. The load-balancing platform allows status messages sent from a switch to be sent to multiple controllers. This allows status messages sent from the switches to be processed in order and the migration blackout can be avoided. The proposed model is compared with a baseline model based on the previous works. In the baseline model, the migration blackout time always occurs when the controller assignment and placement are changed. Numerical results show that the migration blackout time in the proposed model becomes smaller than that in the baseline model. The results also show that the number of controllers placed in the proposed model is smaller than that in the baseline model.