Topology-Based Discrete Water Wave Optimization for Software Defined Network Controller Placement

Software-Defined Networking (SDN) separates data and control planes for flexible management, but large-scale networks need multi-controller collaboration to avoid singlepoint failures, leading to the NP-hard Controller Placement Problem (CPP). This paper proposes a Topology-based Discrete Water Wave optimization algorithm (TDWWO). It discretizes traditional WWO for 2D geographic coordinate topology graphs, integrating node location and connection info. By using WWO’s propagation, breaking, and refraction operators, it balances exploration and exploitation for optimal controller deployment. Notably, when discretizing WWO, node coordinate info is stored via TDtree and applied in propagation-stage searches. A hash table and controller number array accelerate calculations by eliminating redundant traversal. To verify the effectiveness of the TDWWO algorithm, experiments are conducted on five real topologies, and comparisons are made with multiple standard and state-of-the-art algorithms as well as the global optimal solution. The results show that TDWWO outperforms four heuristic algorithms, reducing the SC-avg and SC-worst latencies by up to 65.14% and 31.99%. It ensures calculation stability within 7.28% margin for large-scale topologies and maintains an error rate < 1.18% versus the global optimal solution.