A switching architecture for asynchronous time sharing

The concept of asynchronous time sharing is based on a multiclass network model and a distributed scheduling algorithm that provides guaranteed quality of service. There are three traffic classes for transporting user information and a fourth class for network management and control. Access to resources is resolved through a scheduling mechanism based on time sharing and space partitioning. A generic switching architecture that supports asynchronous time sharing and guarantees quality of service as negotiated at call setup is presented. The architecture exhibits switch fabric independence and only requires that the switch fabric be nonblocking. An operating implementation is described in detail