A BPSO-based Controller Placement Algorithm for Hierarchical Service Function Chaining

With the emergence of hierarchical service function chaining (hSFC), Internet providers are able to offer more flexible and scalable network functions for various user requirements via collaborations of datacenters situated at different geographical locations. However, considering scalability, how to cope with the controller placement in multiple domains of hSFC is still challenging, as corresponding signaling interactions are complicated and heavy. Thus, in this paper, we propose a binary practical swarm optimization (BPSO)-based algorithm to minimize signaling costs among hierarchical controllers through a reasonable placement. Particularly, we first formulate the hSFC controller placement problem as an integer programming model. Then, we propose a heuristic placement algorithm (BPSO-HPA) to approximate the optimal solution, along with a greedy-based placement algorithm (G-HPA) as a benchmark. Finally, we evaluate the performance of the proposed algorithms with extensive simulations, and the results prove that the proposed BPSO-based algorithm is able to approach the results derived from the optimization model with less running time, and outperform the G-HPA algorithm.