A centrality-based selection method for SDN measurement nodes in large-scale networks

Abstract

In Software Defined Network (SDN), selecting a subset of switches for traffic monitoring is crucial for ensuring network security at minimum cost. However, identifying the smallest subset of switches with maximum coverage is NP-Hard. To address this challenge, this research proposes a fast measurement node selection algorithm based on the centrality metric in graph theory, which considers cases where flow path information is known and unknown. The importance of the nodes is ranked by considering the difference in the amount of information of the nodes, followed by the measurement node reduction method to select the measurement nodes with the coverage ratio in mind. Results show that our algorithm reduces the average unit time overhead by 74% while ensuring coverage in both scenarios compared to existing methods.