Aries: Adaptive Cluster Routing With Flexible Link Allocation for Optical MEO/LEO Satellite Networks

In recent years, optical MEO/LEO satellite networks (OMLSNs) have gained significant attention for their complementary advantages, high flexibility, and high-speed data transmission capabilities. However, the uneven spatial-temporal distribution of satellite network traffic and highly dynamic network topology pose significant challenges for traffic routing. Existing routing schemes struggle to effectively address the highly dynamic and constantly changing traffic patterns in OMLSNs, resulting in poor link utilization and network congestion. To address these challenges, this paper proposes Aries, an adaptive cluster routing scheme for OMLSNs. Aries employs a game theory-based dynamic clustering method that divides LEO satellites into multiple regional group networks based on network conditions and traffic patterns, which effectively alleviates the impact of the dynamic change of network topology. Additionally, Aries dynamically determines optimal transmission thresholds for inter-cluster traffic based on actual demand and available bandwidth. It optimizes for long-distance and hotspot traffic by flexibly allocating additional direct links. Such a scheme enables flexible bandwidth resource scheduling and effectively mitigates congestion caused by traffic bursts while enhancing available bandwidth utilization. Comprehensive simulation results demonstrate that Aries outperforms baseline schemes across various satellite network scales and traffic patterns, significantly reducing average end-to-end delay while substantially increasing throughput.