Cooperative Distributed Scheduling for ATS-Based Broadband Networks

A proactive, distributed, cooperative control algorithm for resource scheduling of broadband networks based on asynchronous time sharing, is presented. Underlying its feasibility and eeectiveness are traac predictions of correlated input traac streams into network nodes. The processes modeling the traac load are stationary with nite mean and autocorrelation function. The traac predictions are used to obviate the problem of propagation delays, a major roadblock in inter-nodal cooperation in high speed networks. Detailed implementations of the adaptive, feedback control mechanisms of distributed asynchronous algorithms that project future overload conditions are presented. Nodes projecting to be under stress signal their neighbouring upstream nodes for the latter to readjust their ow rates, while guaranteeing quality of service as negotiated at call set up. The design allows the underlying single node link scheduling algorithm to be run in either a cooperative or non-cooperative mode of operation.