A fully-pipelined fixed-latency communications system with a real-time dynamic bandwidth allocation

摘要: The present invention provides for an on-chip communications method with fully distributed control combining a fully-pipelined, fixed latency, synchronous bus (35) two-level arbitration scheme where the first level of is framed, time division multiplexing and second fairly-allocated round-robin implemented using token passing mechanism. Both latency bandwidth allocation are software programmable in real-time operation system. also communication system access to shared resource (10, 15, 20) controlled by protocol. Access from subsystem through interface module (40, 45, 50, 55, 60). modules 60) provide indirection between (25, 30) be connected 20). This allows decoupling performance requirements. Communication over memory mapped.