A cloud-end data multicast method includes: obtaining a multicast packet, the multicast packet carrying a tenant identifier, a destination address, and a source address; and searching for a multicast group according to the tenant identifier and the destination address. The multicast group includes multicast members. The method also includes obtaining routes corresponding to the multicast members; generating a routing tree according to the routes corresponding to multicast members; and obtaining member addresses corresponding to the multicast members. The method also includes obtaining an address list containing addresses of target members that are identified to receive the multicast packet by performing address filtering according to the source address and the member addresses; encapsulating the multicast packet; and delivering the encapsulated multicast packet to the target members according to the address list and the routing tree.

A provider edge (PE) device and a method implemented thereon are disclosed for Ethernet virtual private network (EVPN). According to an embodiment, a first PE device performs label assignment procedure with a second PE device such that the first and second PE devices share an Ethernet segment identifier (ESI)-excluded label and know a correspondence between the ESI-excluded label and a label combination of an ESI label and a VPN label. The first PE device encapsulates a packet of broadcast, unknown unicast or multicast (BUM) traffic, with the ESI-excluded label instead of the label combination. The first PE device sends the encapsulated packet to the second PE device.

Techniques for limiting forwarding of multicast communications are described herein. For example, the techniques intelligently forward data along paths of a network where members of a multicast group are located. As such, a node that does not lead to members of the multicast group may be configured to selectively and intelligently forward multicast messages it receives. This can reduce network communications, ultimately conserving processing, communication, and/or battery resources of the nodes and improving performance of the network.

Various aspects of the present disclosure generally relate to wireless communication. In some aspects, a user equipment (UE) may receive an indication of a mode for mapping multicast broadcast quality of service (MB-QoS) flows to logical channel identifiers and group radio network temporary identifiers (G-RNTIs); identify an MB-QoS flow from a medium access control (MAC) transport block (TB) based at least in part on the indicated mode; and decode data included in the MB-QoS flow. Numerous other aspects are provided.

Methods, systems, and devices for wireless communication are described. Generally, the described techniques provide a base station receiving an indication from a core network to serve multicast/broadcast traffic to one or more UEs. The base station may select a radio bearer mode from a plurality of radio bearer modes for delivery of the multicast/broadcast traffic to at least one UE of the one or more UEs. For example, the base station may select a multicast/broadcast only mode, a mixed multicast/broadcast mode and unicast mode, or a unicast mode for service of the traffic. Dependent on the selected mode and/or quality of service, the base station may select a dedicated radio bearer (DRB) for communication of the traffic in a unicast mode (e.g., to a particular UE) or a multicast radio bearer (MRB) for communication of the traffic in a broadcast/multicast mode.

An integrated circuit chip has a set of communication units, each unit being configured to operate according to a protocol in which a data packet sent by one unit is receivable by one unit only, each unit being configured to send at least one packet having one of a plurality of tiers to at least one other unit and being configured to specify, for each tier, a subset of destination units to which packets of that tier are to be sent, wherein each unit is configured to: receive a packet having one of the plurality of tiers; determine the tier of the received packet; and sequentially send packets having a different tier to the tier of the received packet to each of the respective subset of destination units for the different tier.

A first network device is configured with a rule preventing network traffic from travelling from the first network device to one or more other network devices. The first network device is configured to receive and distribute network traffic to the one or more other network devices. A second network device receives and distributes network traffic to the one or more other network devices. The first network device determines that the second network device has failed. In response to determining that the second network device has failed, the first network device removes the rule so that the first network device receives and distributes network traffic to the one or more other network devices.

A network device configured to perform scalable, in-network computations is described. The network device is configured to process pull requests and/or push requests from a plurality of endpoints connected to the network. A collective communication primitive from a particular endpoint can be received at a network device. The collective communication primitive is associated with a multicast region of a shared global address space and is mapped to a plurality of participating endpoints. The network device is configured to perform an in-network computation based on information received from the participating endpoints before forwarding a response to the collective communication primitive back to one or more of the participating endpoints. The endpoints can inject pull requests (e.g., load commands) and/or push requests (e.g., store commands) into the network. A multicast capability enables tasks, such as a reduction operation, to be offloaded to hardware in the network device.

A network device configured to perform scalable, in-network computations is described. The network device is configured to process pull requests and/or push requests from a plurality of endpoints connected to the network. A collective communication primitive from a particular endpoint can be received at a network device. The collective communication primitive is associated with a multicast region of a shared global address space and is mapped to a plurality of participating endpoints. The network device is configured to perform an in-network computation based on information received from the participating endpoints before forwarding a response to the collective communication primitive back to one or more of the participating endpoints. An injection policy comprising the issuing of credits enables each endpoint to limit the amount of collective communication primitives injected into the network simultaneously to reduce network congestion caused by increased network traffic due to the multicast capability of the network devices.

A network device configured to perform scalable, in-network computations is described. The network device is configured to process pull requests and/or push requests from a plurality of endpoints connected to the network. A collective communication primitive from a particular endpoint can be received at a network device. The collective communication primitive is associated with a multicast region of a shared global address space and is mapped to a plurality of participating endpoints. The network device is configured to perform an in-network computation based on information received from the participating endpoints before forwarding a response to the collective communication primitive back to one or more of the participating endpoints. The endpoints can inject pull requests (e.g., load commands) and/or push requests (e.g., store commands) into the network. A multicast capability enables tasks, such as a reduction operation, to be offloaded to hardware in the network device.