ADAPTIVE BACK-PRESSURE CONGESTION CONTROL-BASED ON LOCAL INFORMATION

The problem of distributed congestion control as it arises in communic ation networks as well as in manufacturing systems is studied in this paper, In particular, a multistage queueing system that models virtual circuit and datagram communication networks and a class of manufactur ing systems are considered, The topology may be arbitrary, there are m ultiple traffic classes, and the routing can be class dependent, with routes that may form direct or indirect loops. The model incorporates the functions of transmission scheduling, how control, and routing, th rough which congestion control is performed in the network, A policy i s given that performs these functions jointly, According to the policy , heavily loaded queues are given higher priority in service, A conges ted node may reduce the flow from upstream nodes through a flow contro l mechanism, Whenever routing is required, it is performed in such a m anner that the lightly loaded queues receive most of the traffic. For arrival processes with bounded burstiness, the policy guarantees bound ed backlogs in the network, as long as the load of each server is less than one. The actions of each server are based on the state of its ow n queues and of the queues one hop away, Therefore, they are implement able in a distributed fashion, An adaptive version of the policy is al so provided which makes it independent of the arrival rates.