Hop-by-Hop Dynamic Congestion Control with Contact Interruption Probability for Intermittently Connected Deep Space Information Networks

摘要

We investigate congestion control in intermittently connected deep space information networks (DSINs) in which the impact of link outage and fairness guarantee over traffic streams is considered. First, we model deep space backbone layer of DSINs as a set of hypergraph, and we also quantify a finite discrete incomplete contact interruption probability distribution as a σ-fairness Renyi entropy with the fairness factor along the corresponding link. In view of different scenarios of hop-by-hop fluid flows based on the hypergraph model, we then propose two hop-by-hop dynamic congestion control schemes, namely, noncooperative dynamic congestion control model under multi-session uni-path and autonomic congestion control strategy under uni-session multi-path. Specifically, we formulate noncooperative dynamic congestion control model as a differential game mathematically, and obtain the feedback Nash equilibrium solution by solving the partial differential equations. To guarantee the buffer constraint of joint node, we also devise an optimal rate update policy. In addition, a monotone transition function is defined in order to effectively regulate the instant rate in autonomic congestion control strategy, and an autonomic rate control algorithm via τ_1/τ_2 search is further presented. Finally, our results are performed by numerical simulations to demonstrate the effectiveness of the proposed schemes.