Abstract: A common access ring (CAR) network includes a main ring and one or more sub-rings. The main ring includes one or more masters, one or more slaves, and one or more bridges. Each sub-ring is coupled to the main ring through a corresponding bridge. Each node of the CAR network is assigned a unique identifier, thereby implementing a global flat address space. One or more masters may issue requests on the CAR network, such that multiple transactions are simultaneously pending. Multiple masters may simultaneously issue requests to the same slave. However, each master cannot make more than one request at a time, and must wait until a current request is completed before making another request. The ring architecture ensures that no more than one request arrives at a slave at any given time. Requests received by busy slaves are returned to the originating masters, and may be subsequently re-sent.

Abstract: A common access ring (CAR) architecture that supports multiple masters and slaves is provided. One or more masters may make a request on the ring at the same time, such that multiple transactions are simultaneously pending. Moreover, multiple masters may simultaneously make a request to the same slave. However, each master cannot make more than one request at a time, and must wait until a current request is completed before making another request. The ring architecture ensures that no more than one request arrives at a slave at any given time. If a request arrives while a slave is processing a previous request, the arriving request is not serviced, and the master that originated the arriving request is asked to retry the request at a later time. Atomic shadow-write operations are supported by including all shadow registers in a dedicated sub-ring of the CAR architecture.

Abstract: A common access ring (CAR) architecture that supports multiple masters and slaves is provided. One or more masters may make a request on the ring at the same time, such that multiple transactions are simultaneously pending. Moreover, multiple masters may simultaneously make a request to the same slave. However, each master cannot make more than one request at a time, and must wait until a current request is completed before making another request. The ring architecture ensures that no more than one request arrives at a slave at any given time. If a request arrives while a slave is processing a previous request, the arriving request is not serviced, and the master that originated the arriving request is asked to retry the request at a later time. Atomic shadow-write operations are supported by including all shadow registers in a dedicated sub-ring of the CAR architecture.

Abstract: A network switch port receives, stores and then forwards incoming cells. The network switch assigns each incoming cell to one of a set of flow queues, each of which is allotted a portion of space in a cell memory. The switch port periodically computes a average of the number of cells assigned to each flow queue stored in the cell memory during a preceding period, and assigns a discard weight to each incoming cell that is a function of the amount by which the average for the cell's assigned flow queue exceeds a threshold level. The switch port randomly discards incoming cells assigned to the flow queue with a probability that increases with the incoming cells' assigned weights. The switch port stores incoming cells that are not randomly discarded in the cell memory and later forewords them from the cell memory.

Abstract: An input or output switch port for a network switch converts each incoming packet into a cell sequence stores each cell in a cell memory. The switch port includes a traffic manager for queuing cells for departure from the cell memory and then signaling the cell memory to read out and forward cells in the order they are queued. The traffic manager selectively queues cells for departure on either a cell-by-cell or sequence-by-sequence basis. When cells are queued for departure on a cell-by-cell basis, cells of two or more sequences may be alternately read out and forwarded from the cell memory. Thus cells of different sequences may be interleaved with one another as they depart the cell memory. When a cell sequence is queued on a sequence-by-sequence basis all of its cells are read out of the cell memory and forwarded as a contiguous sequence and are not interleaved with cells of other sequences of the same departure queue.

Abstract: A traffic manager for a network switch input or output port stores incoming cells in a cell memory and later sends each cell out of its cell memory toward one of a set of forwarding resources such as, for example, another switch port or an output bus. Data in each cell references the particular forwarding resource to receive the cell. Each cell is assigned to one of several flow queues such that all cells assigned to the same flow queue are to be sent to the same forwarding resource. The traffic manager maintains a separate virtual output queue (VOQ) associated with each forwarding resource and periodically loads a flow queue (FQ) number identifying each flow queue into the VOQ associated with the forwarding resource that is to receive the cells assigned to that FQ. The traffic manager also periodically shifts an FQ ID out of each non-empty VOQ and forwards the longest-stored cell assigned to that FQ from the cell memory toward its intended forwarding resource.

Abstract: A traffic manager for a network switch input or output port stores incoming cells in a cell memory and later sends each cell out of its cell memory toward one of a set of forwarding resources such as, for example, another switch port or an output bus. Data in each cell references the particular forwarding resource to receive the cell. Each cell is assigned to one of several flow queues such that all cells assigned to the same flow queue are to be sent to the same forwarding resource. The traffic manager maintains a separate virtual output queue (VOQ) associated with each forwarding resource and periodically loads a flow queue (FQ) number identifying each flow queue into the VOQ associated with the forwarding resource that is to receive the cells assigned to that FQ. The traffic manager also periodically shifts an FQ ID out of each non-empty VOQ and forwards the longest-stored cell assigned to that FQ from the cell memory toward its intended forwarding resource.

Abstract: A network switch port receives, stores and then forwards incoming cells. The network switch assigns each incoming cell to one of a set of flow queues, each of which is allotted a portion of space in a cell memory. The switch port periodically computes a average of the number of cells assigned to each flow queue stored in the cell memory during a preceding period, and assigns a discard weight to each incoming cell that is a function of the amount by which the average for the cell's assigned flow queue exceeds a threshold level. The switch port randomly discards incoming cells assigned to the flow queue with a probability that increases with the incoming cells' assigned weights. The switch port stores incoming cells that are not randomly discarded in the cell memory and later forewords them from the cell memory.

Abstract: An input or output switch port for a network switch converts each incoming packet into a cell sequence stores each cell in a cell memory. The switch port includes a traffic manager for queuing cells for departure from the cell memory and then signaling the cell memory to read out and forward cells in the order they are queued. The traffic manager selectively queues cells for departure on either a cell-by-cell or sequence-by-sequence basis. When cells are queued for departure on a cell-by-cell basis, cells of two or more sequences may be alternately read out and forwarded from the cell memory. Thus cells of different sequences may be interleaved with one another as they depart the cell memory. When a cell sequence is queued on a sequence-by-sequence basis all of its cells are read out of the cell memory and forwarded as a contiguous sequence and are not interleaved with cells of other sequences of the same departure queue.