On the Stability of Autonomous Decentralized Flow Control in High-Speed Networks with Asymmetric Configurations

In high-speed network, the principle adopted for a timesensitive flow control mechanism should be autonomous decentralized control. In this mechanism, each node in a network handles its local traffic flow only on the basis of the information it is aware of, although it is desirable that the decision-making at each node leads to high performance of the network as a whole. To implement this mechanism, we need first to investigate the relationship between the flow control mechanism of each node and network performance. In particular, the stability of the network performance is important. Our previous studies have proposed a simple mechanism for autonomous decentralized flow control and have shown that it has desirable stability. However, since we used a simple, symmetric, and homogeneous network model having a single bottleneck in the evaluation, the results cannot be generalized. In this paper, by using more complex network configurations in evaluations, we show that the principle of autonomous decentralized flow control can be applied to high-speed network with asymmetric configurations.