An apparatus and method for operating a ring interconnect are disclosed. The ring interconnect has a plurality of nodes that are used to connect to associated components, and is arranged to transport a plurality of slots around the ring interconnect between the nodes in order to transfer items of traffic allocated into those slots between components connected to the nodes. For each item of traffic, one of the components acts as a source to allocate that item of traffic into a slot, and another components acts as destination to seek to remove that item of traffic from the slot. In a default mode of operation, the ring interconnect is arranged to allow all of the slots to be available for transfer of any items of traffic. Special slot management circuitry is provided that is responsive to a throughput alert trigger indicating a potential for occurrence of a throughput inhibiting condition, to cause a slot amongst the plurality of slots to be reserved as a special slot that is constrained for use only when one or more determined conditions are met. Further, the one or more determined conditions are arranged to cause the special slot to be used in a manner that seeks to avoid the throughput inhibiting condition arising.

Multiple nodes (2) of a ring network communication system (1) include a heteronomous node (22) and an autonomous node (23). The heteronomous node (22) operates in a heteronomous mode in which data generated by a data generation part (8) of an own node is transmitted as transmitted data Dt only when received data Dir received by a data reception part (13) includes a data transmission request. The autonomous node (23) operates in an autonomous mode in which the data generated by the data generation part (8) of the own node is transmitted as the transmitted data Dt regardless of reception of the received data Dir including the data transmission request. Data including the data transmission request is not required for the autonomous node, and the control period is shortened.

Embodiments for providing enhanced multicast data transfer for ring topology based artificial intelligence systems are disclosed. Multicast data is sent to a plurality of disjointed cores in a multicast group according to a first multicast mode, a second multicast mode, or a third multicast mode, where the first multicast mode sends a first half the multicast data on first multicast ring and a second half on a second multicast ring, the second multicast mode sends the multicast data on either the first multicast ring or the second multicast ring, and the third multicast mode replicates the multicast data and sends the multicast data to both the first multicast ring and the second multicast ring.

Embodiments for providing enhanced multicast data transfer for ring topology based artificial intelligence systems are disclosed. Multicast data is sent to a plurality of disjointed cores in a multicast group according to a first multicast mode, a second multicast mode, or a third multicast mode, where the first multicast mode sends a first half the multicast data on first multicast ring and a second half on a second multicast ring, the second multicast mode sends the multicast data on either the first multicast ring or the second multicast ring, and the third multicast mode replicates the multicast data and sends the multicast data to both the first multicast ring and the second multicast ring.

An on-board network system includes an arithmetic unit and a plurality of control devices connected to the arithmetic unit via a communication network. The communication network is configured such that each of the control devices are communicative with the arithmetic unit and/or other one(s) of the control devices via at least two communication paths. In each of the communication paths, a priority order of communications is determined according to a type of communication data. If a communication disturbance is detected in one of the communication paths, the arithmetic unit or one(s) of the control devices adjusts the traffic of the communication data in accordance with the priority of the other one of the communication paths.

An object of the present invention is to provide a communication system and a communication method that can suppress an increase in the total fiber distance even if the number of secondary devices increases and that can make complex processing for band allocation for uplink optical signals unnecessary. The configuration of the communication system according to the present invention is a mixture of a ring type and a branch type, in which an optical fiber cable (primary path) from a communication station to each user's home is arranged in a ring (loop) shape, and the primary path is branched and laid to the user's home. Furthermore, unidirectional communication is enabled by a ring-type primary path, and the transmittable amount can be made uniform by further combining time division multiplexing.

According to one or more embodiments of the disclosure, a particular networking device located in a ring of networking devices of a network receives an indication from a supervisory service that the particular networking device has been designated a ring manager for the ring of networking devices. The particular networking device determines that the supervisory service is unreachable by the ring of networking devices. The particular networking device obtains telemetry data regarding a new device connected to the ring of networking devices. The particular networking device onboards, based on the telemetry data, the new device to the network, when the supervisory service is unreachable by the ring of networking devices.

According to one or more embodiments of the disclosure, a particular networking device joins a ring network of networking devices that has a ring topology. The particular networking device monitors the ring network for a multicast frame used within the ring network to detect link failures. The particular networking device determines that a link in the ring network has failed, based on the particular networking device not receiving the multicast frame within a threshold amount of time. The particular networking device initiates repair of the ring network, when the particular networking device determines that the link in the ring network has failed.

According to one or more embodiments of the disclosure, a particular networking device located in a ring of networking devices of a network receives an indication from a supervisory service that the particular networking device has been designated a ring manager for the ring of networking devices. The particular networking device determines that the supervisory service is unreachable by the ring of networking devices. The particular networking device obtains telemetry data regarding a new device connected to the ring of networking devices. The particular networking device onboards, based on the telemetry data, the new device to the network, when the supervisory service is unreachable by the ring of networking devices.

Reception information of frames received at a redundant port may be stored to a data structure in group of data structures such that reception information of all frames in a sequence are not stored to the same data structure and the reception information can be quickly accessed for duplicate detection. The data structure includes reception information for a portion of consecutive frames in a sequence of frames transmitted by a specific transmitter node. A communications layer address of the transmitter node and information capable of identifying a portion of consecutive frames in a sequence of frames are derived from a received frame. The reception information for the received frame in the data structure is read on the basis of derived information capable of identifying the received frame within the portion of consecutive frames.