Throughput Capacity of LEO Satellite Networks Under Differentiated Inter-Satellite Link Rates

Abstract

Throughput capacity is an important metric for evaluating the traffic carrying capability of low Earth orbit (LEO) satellite networks, which is constrained by inter-satellite link (ISL) rates. During the operation of LEO satellite networks, the ISL rates are inevitably differentiated due to the factors including relative satellite motion and space environmental effects, etc. Such differentiation may cause inadequate matching between traffic allocation and ISL rates, which would degrade throughput capacity. Existing studies commonly adopt the ideal equal ISL rate to analyze the throughput capacity for tractability, which ignores rate differentiation and overestimates the performance of the LEO satellite network. To bridge the gap, we investigate throughput capacity under the practical ISL rates. Specifically, to formalize throughput capacity bottlenecks, we propose rectangular block cut sets through graph partitioning, which divide the network into two complementary parts. The proposed cut sets recast the formalization from detailed traffic allocation over the entire network to a comparison of the traffic flow rates supported by different rectangular block cuts. The cut set attaining the lowest supported traffic flow rate, i.e., the worst rectangular block cut set, is shown to yield a throughput capacity upper bound. Numerical results indicate that the throughput capacity decrease shows a sublinear increasing trend as ISL rate variance rises. Moreover, increasing ISLs associated with the worst rectangular block cut set almost achieves the same throughput capacity gain as increasing all ISLs, with throughput capacity linearly increasing with ISL rate increase.