Traffic Allocation for Heterogeneous Links in Satellite Data Relay Networks

Abstract

Efficient traffic allocation for heterogeneous links in satellite data relay networks (SDRNs) has been an attractive topic nowadays. However, different traffic types and link heterogeneity bring plenty of difficulties to the link delay evaluation which directly impacts the traffic allocation efficiency. Owing to the unique determination of the probability distribution of arrival processes, we succeed in characterizing any type of traffic with the help of the moment generating function (MGF). According to the alternation and discontinuity of satellite ground links (SGLs), we propose a Markov chain Monte Carlo embedded service matrix construction method to model the SGL service. Being apt at combining service processes, the stochastic network calculus based min-plus convolution is employed to depict the relay service. By deriving the tight upper bounds of the delay in heterogeneous links based on the MGF, min-plus convolution, and Markov chain, respectively, taking advantage of the Chernoff bound and Union bound, we calculate the optimal traffic allocation ratio in terms of minimum system delay. Finally, numerous simulations are conducted to show the relationships among the optimal traffic allocation ratio, traffic type, and link heterogeneity.