A Density-aware SDN Controller Placement in LEO Satellite Networks

The rapid deployment of Low Earth Orbit (LEO) satellite constellations necessitates efficient network management solutions due to their dynamic topologies and high node mobility. Software-Defined Networking (SDN) offers a promising architecture for centralized control and programmability in such environments. However, optimal controller placement remains a key challenge due to the time-varying nature of satellite positions and traffic demands. In this paper, we propose a density-aware SDN controller placement strategy that considers local satellite density, isolation distance, and neighborhood compactness to select controllers that are both traffic-aware and spatially welldistributed. The proposed approach employs a greedy selection mechanism followed by refinement to minimize worst-case switch to controller delays. We evaluate the proposed method using the Iridium-NEXT constellation and compare its performance with baselines, including Standard K-Means, K-Means++, and Spectral Clustering. Experimental results show that the proposed method significantly outperforms the baselines in terms of the worst-case switch to controller delay, achieving up to a $93 \%$ improvement.