SDN-based Optimized Controller Deployment Strategy for Satellite Network
Haiyun Liu, Tao Zhang, Dingyuan Shi · 2024
In Software-Defined Networking (SDN), the positioning of controllers and the division of switches determine the upper limit of the network’s responsiveness to network events. However, highly dynamic satellite networks pose a significant challenge to the controller placement problem. The difficulty is to achieve efficient control over a long period of time with low overhead. Existing methods mostly perform computations independently and greedily in each time slot, leading to frequent changes in the controller placement and the control relationships between controllers and switches. In this paper, we present a comprehensive model to measure the control plane overhead. By defining migration overhead and anticipated overhead to ensure the continuity of control strategies, we formulate the controller placement problem in Software-Defined Low Earth Orbit (LEO) satellite networks as an optimization problem. To tackle this issue, we draw inspiration from hierarchical clustering techniques and introduce a heuristic algorithm named Hierarchical Clustering-based Minimizing Overhead Algorithm (HCMOA). Evaluations of our simulation system show that HCMOA significantly outperforms relevant approaches in terms of control overhead and average response latency.