Abstract: The present disclosure generally discloses a multicast capability configured to support multicast that is based on use of Bit Indexed Explicit Replication (BIER). The multicast capability may be configured to support deployment of BIER into a network in a manner that continues to support use of Protocol Independent Multicast (PIM) for multicast control within the network. The multicast capability may be configured to support deployment of BIER, into a network including an Interior Gateway Protocol (IGP) area, using a BIER domain that is commensurate with the IGP area or that is only a portion of the IGP area. The multicast capability may be configured to support deployment of BIER into a network in a manner supporting stitching of traditional PIM access networks via a BIER core network.

Abstract: A method of cyber-attack protection is provided. The method includes receiving one or more internet group management protocol (IGMP) membership report messages on one or more ports of a network switch. A table is maintained associating each of the one or more ports to the one or more group addresses of which any hosts coupled to that port are members. An IGMP group leave message is received at a first port of the network switch. A group address in the group leave message is compared to all group addresses associated in the table with the first port. If the group address in the group leave message does not match any group address associated with the first port, the group leave message is dropped without being acted on in accordance with IGMP.

Abstract: A network device may receive, from a first neighbor network device, information indicating that the network device is to remove a bit indexed explicit replication (BIER) header from a multicast packet prior to transmitting the multicast packet to the first neighbor network device. The network device may receive the multicast packet from a second neighbor network device, where the multicast packet includes the BIER header and is to be transmitted to the first neighbor network device, and where the network device, the first neighbor network device, and the second neighbor network device are included in a same BIER domain. The network device may remove the BIER header from the multicast packet, and may transmit the multicast packet to the first neighbor network device after removing the BIER header from the multicast packet.

Abstract: An apparatus for routing multicast data packets, the apparatus includes an ingress port to receive data streams of multicast data packets and status data about egress ports available to transmit the multicast traffic data streams. A processor coupled to the ingress port, to identify source data of the multicast data packets of the data streams to match the multicast data packets with available egress ports. The processor to determine, using the identified source and status data which of the multicast data packets matches the available egress ports. The processor to select a first data path coupled to the egress port to transmit the matched multicast data packets to available egress ports where the selected first data path is configured to enable the direct transmission of the matched multicast data packets to available egress ports.

Abstract: A method for implementing a software-defined network (SDN) includes receiving, by an SDN controller, virtual network information sent by a cloud computing platform, where the virtual network information includes a correspondence between information about a virtual machine (VM) and information about a server running the VM. The method also includes acquiring, by the SDN controller, physical network information, where the physical network information includes a correspondence between information about a top-of-rack (TOR) switch and the information about the server. Additionally, the method includes acquiring, by the SDN controller, a correspondence between the information about the VM and the information about the TOR switch according to the virtual network information and the physical network information.

Abstract: Systems and methods for providing multicast group multicast local identifier (MLID) dynamic discovery on received multicast messages for a relevant multicast global identifier (MGID) in a high performance computing environment. By allowing InfiniBand (IB) clients to associate local queue pairs (QPs) with the MGID(s) of relevant multicast group(s) without requiring any join request to the subnet manager (SM)/subnet administration (SA), it is possible to receive relevant multicast (MC) messages without imposing the SM/SA overhead of a conventional multicast group join request. After receiving, at an end-node of the subnet, a multicast packet including a multicast global identifier and a multicast local identifier, the end-node can inspect the multicast packet to learn the multicast local identifier and include the learned multicast local identifier in the multicast group record at the end-node for the received multicast global identifier.

Abstract: Aspects of the present disclosure involve methods and systems for mapping network service paths between endpoint devices of a network. An aggregate network traffic database is used to store network traffic data retrieved from network devices of a network. A network service mapper then queries the database using endpoint device identifiers corresponding to endpoint devices to determine which network devices interacted with the endpoint devices during provisioning of previous network service instances. The set of network devices is then compared and analyzed to determine one or more network service paths between a pair of endpoint devices. In certain implementations, network architecture rules and/or network device configuration data is also analyzed to determine an order of the network devices along the one or more network service paths.

Abstract: A method for controlling wireless communication between a mobile device and an electronic device includes enabling wireless connection between the mobile device and the electronic device via a first wireless communication link; maintaining the wireless connection for a period of time so as to enable communication of information between the mobile device and the electronic device using the first wireless communication link; and terminating the wireless connection between the mobile device and the electronic device in response to a timeout event.

Abstract: Methods and apparatus for enabling a peripheral processor to retrieve and load firmware for execution within the constraints of its memory. The peripheral processor is allocated a portion of the host processor's memory, to function as a logical secondary and tertiary memory for memory cache operation. The described embodiments enable the peripheral processor to support much larger and more complex firmware. Additionally, a multi-facetted locking mechanism is described which enables the peripheral processor and the host processor to access the secondary memory, while minimally impacting the other processor.

Abstract: A method which is implemented by an access point, and in which requests are received from a number of user devices, in which each of the requests specifies one or multiple requested file blocks of at least one file. A set of multicast groups is determined based on the received requests, and each multicast group contains a set of indices of one or multiple of the user devices and a set of indices of the file blocks which are to be transmitted to the user devices indicated in the multicast group. For each multicast group of the set the file blocks indicated in the multicast group are retrieved and encoded, and a forward error correction process is used. For each of the multicast groups of the set the encoded file blocks indicated in the multicast group are transmitted to the user devices indicated in the multicast group.

Abstract: In some examples, a controller for a network includes a path computation module that determines, for a plurality of LSPs or other flows having a common source, shortest paths of the network from the common source to respective destinations of the plurality of LSPs based at least on a minimum bandwidth. The path computation module further determines, after determining the shortest paths, a shortest path for the LSP of the plurality of LSPs as the shortest path of the shortest paths of the network from the common source to a destination for the LSP. A path provisioning module of the controller, after the path computation module determines the shortest path for the LSP and in response to the path computation modules routing the LSP to the shortest path for the LSP on a network model of the network, installs the LSP to the network as routed to the shortest path.

Abstract: Disclosed is a packet forwarding method, comprising: receiving a packet to be sent to a bit-forwarding egress router (BFER) based on traffic engineering for bit index explicit replication (BIER-TE), and a path identifier, wherein the path identifier is used for indicating that the packet is sent from a first bit-forwarding router (BFR) to a primary path and a backup path of a next-hop BFR of the first BFR; determining, according to the path identifier, a bit string used for forwarding the packet to the BFER; and filling the bit string in the packet and forwarding the packet according to the bit string.

Abstract: Computerized aspects monitor listening virtual local area network ports for multicast packets from hypervisor virtual machines that meet a multicast group definition, and retrieve source addresses for original hypervisor sources of detected ones of the multicast packets, bind the retrieved addresses to the multicast packets to generate encapsulated packets, and forward the encapsulated packets as user datagram protocol unicast packets to a destination address of a virtual extensible helper daemon defined for a destination hypervisor of a second virtual local area network. The destination address meets the multicast group definition for the first virtual local area network, and the second virtual local area network is geographically remote from and different from the first virtual local area network, and thus the aspects send the first user datagram protocol unicast packet through an internet protocol router to the designated receiving port.

Abstract: A lighting system comprises at least one light source (1,3), a mobile system, e.g. a mobile device (5), and at least one further device (6-8). The mobile system comprises a light sensor (11), a communication interface (12) and a processor (13). The mobile system is configured to receive from a light source (1) a light emission (2) in which an identifier has been encoded, e.g. using Visible Light Communication (VLC) techniques, and to determine the identifier from the light emission (2). The identifier is associated with a group of devices that does not comprise mobile devices that are not receiving a light emission in which this identifier has been encoded. The mobile system is configured to join this group and to communicate with at least one further device (6) from this group of devices (5,6).

Abstract: A first map request message is sent from a source network device to a mapping network device to determine a destination network device associated with a destination endpoint device and a security association between the source network device and the destination network device. A first response message is received at the source network device that includes data indicating a mapping between the destination network device and the destination endpoint device and data indicating a security association between the source network device and the destination network device. The data is stored at the source network device. A second map request message is sent from the source network device to the mapping network device to update the data indicative of the mapping or the security association. A second response message is received at the source network device from the mapping network device.

Abstract: In some examples, a method includes receiving, by a first provider edge (PE) device of a layer 3 network, configuration data configuring the PE device to provide, via an Ethernet segment with an edge device, active-active multi-homing layer 2 (L2) virtual bridge connectivity to the edge device using an Ethernet Virtual Private Network (EVPN) instance; receiving, by the first PE device, a multicast Join message for multicast traffic, the multicast Join message identifying a second PE device that also participates in the EVPN instance; processing, by the first PE device, the multicast Join message to generate multicast forwarding state; and forwarding, by the first PE device based at least on the multicast forwarding state, the multicast traffic.

Abstract: What is disclosed is a system and method to automatically configure the fanout modes of ports of a switch in a network system. The network node coupled to the port of the network switch is discovered. A plurality of fanout modes for the port is determined. The port is configured with one of the plurality of fanout modes. The system may also use neighboring ports to determine the maximum speed of a network interface card of the network node.

Abstract: A method and apparatus is executed by a network device in a source packet in routing (SPRING) network. The method is to establish a multipoint to multipoint multicast distribution tree where the network device is selected as a rendezvous point for the multipoint to multipoint multicast distribution tree. The method includes advertising the rendezvous point with a multicast group identifier and multicast segment identifier via a control plane protocol, receiving at least one interest registration from receivers of the multicast group via the control plane protocol, and constructing the multipoint to multipoint multicast distribution tree with the rendezvous point as a root and each of the receivers as leaves using unicast tunnels and replication points.

Abstract: The present invention provides a transmitter including a first source network interface connected to a first transmission path of a plurality of transmission paths, an nth source network interface connected to an nth transmission path of the plurality of transmission paths, a video signal segmenting unit configured to segment a video signal into a plurality of pieces of data, a sequence number allocating unit configured to allocate a sequence number to each of the plurality of pieces of data, an IP packet generating unit configured to generate a plurality of IP packets by adding IP headers to the plurality of pieces of data, and a transmitting unit configured to allocate each of the plurality of IP packets to one of the first source network interface to the nth source network interface, respectively, and to transmit the IP packets independently.

Abstract: Example methods are provided for first host to perform multicast packet handling in a software-defined networking (SDN) environment. The method may comprise: in response to the first host detecting, from a first virtualized computing instance, a request to join a multicast group address, obtaining control information from a network management entity. The control information may include one or more destination addresses associated with one or more second hosts that have joined the multicast group address on behalf of multiple second virtualized computing instances. The method may also comprise: in response to the first host detecting an egress multicast packet that includes an inner header addressed to the multicast group address, generating one or more encapsulated multicast packets based on the control information and sending the one or more encapsulated multicast packets in a unicast manner or multicast manner, or a combination of both.

Abstract: Methods and systems for executing a penetration test of a networked system by a penetration testing system so as to determine a method by which an attacker could compromise the networked system, and/or for distributing common sets of data to nodes of a networked system. The methods and systems include identifying network nodes which have shared broadcast domains.

Abstract: A participant in a network is configured to query a data storage system to determine whether there have been any changes to a network. The participant receives a response to the query and determines, based at least in part on the response, whether reconfiguration is necessary. If the network has changed (e.g., if the network topology has changed), as indicated in the response, the system performs one or more actions in accordance with the response. Multiple participants in the network may query the data storage system so that, collectively, network updates are initiated by changes to data in the data storage system. The network may be an overlay network that allows communication according to a communication protocol, such as multicast, that may not be completely supported by a physical network substrate.

Abstract: A method for group communication of a broadcasting server in a wireless communication system is provided. The method for group communication of a broadcasting server includes receiving a data packet from a content server, generating a transmission packet by copying the data packet, and transmitting the generated transmission packet to a base station.

Abstract: Embodiments are directed to scaling services, transitioning from a first service version to a second version and to implementing an external system integration service. In one scenario, a computer system establishes a message broker service that maintains message queues that allow communication between services. The message queues receive messages from publishers and transfer messages to subscribers. The computer system indicates a specified message queue for each service, where the specified message queue is configured to maintain messages for that service. The computer system also moves at least one of the services to a second, different computer system, while the specified message queue maintains messages for the moved service. The computer system further allows an external system integration service to be implemented which subscribes to specified, registered messages and forwards the registered messages to selected external entities.

Abstract: In one embodiment, a method includes receiving from a host node comprising a tenant multicast application with a single tenant deployment, an encapsulated multicast packet over unicast at a virtual replicator, the encapsulated multicast packet comprising receiver information for the virtual replicator, replicating the encapsulated multicast packet at the virtual replicator, and transmitting the encapsulated multicast packet over unicast to a plurality of receiving nodes based on the receiver information. The receiving nodes comprise the tenant multicast application and the virtual replicator operates in a cloud without multicast capability. An apparatus is also disclosed herein.

Abstract: Method and apparatus are provided for synchronizing execution of a plurality of threads on a multi-threaded processor. A program executed by a thread can have a number of synchronization points corresponding to points where execution is to be synchronized with another thread. Execution of a thread is paused when it reaches a synchronization point until at least one other thread with which it is intended to be synchronized reaches a corresponding synchronization point. Execution is subsequently resumed. A control core maintains status data for threads and can cause a thread that is ready to run to use execution resources that were occupied by a thread that is waiting for a synchronization event.

Abstract: A multicast forwarding method and an apparatus, where a bit-forwarding egress router (BFER) in a bit index explicit replication (BIER) network may determine, according to information about a multicast group in a multicast receiver list obtained in advance, an Internet Protocol (IP) address of a registration router that supports the multicast group, the BFER generates a BFER registration message according to the information about the multicast group and the IP address of the registration router that supports the multicast group, where the BFER registration message includes the information about the multicast group, a destination IP address of the BFER registration message is the IP address of the registration router, and a source IP address of the BFER registration message is an IP address of the BFER, and the BFER sends the BFER registration message to the registration router.

Abstract: An example method includes exchanging targeted hello messages to establish a targeted neighbor connection between a first routing device and a second routing device, wherein one of the routing devices comprises a central routing device, and wherein another one of the routing devices comprises an ingress routing device. The example method further includes processing a source-active register message that specifies a source address and an identifier that are collectively associated with a multicast stream, and wherein the source-active register message further indicates whether the multicast stream is active or withdrawn.

Abstract: Various example embodiments are disclosed relating to the establishment of reliable multicast/broadcast sessions in a wireless network. According to an example embodiment, an apparatus may be configured to receive, from a wireless recipient station, a request to establish a reliable multicast/broadcast session with the recipient station. The apparatus may be further configured to transmit, to the recipient station, a response to the request to establish the reliable multicast/broadcast session. The response may include one or more retransmission fields describing a retransmission of data for the requested reliable multicast/broadcast session. For example, the request may include a retransmission multicast address to be used for retransmission of data for the multicast/broadcast session.

Abstract: In accordance with various implementations, a method is performed at a multicast gateway node within an operator network, where the multicast gateway node includes one or more processors, non-transitory memory, an ingress interface, and one or more egress interfaces. The method includes determining a multicast identifier for a user device in response to obtaining a registration request associated with the user device. The method also includes generating a header for a multicast data stream based at least in part on the multicast identifier in response to obtaining a multicast flow join request. The method further includes populating a packet associated with the multicast data stream with the header. The method further includes forwarding the packet to the user device via a portion of the one or more egress interfaces, where the portion of the one or more egress interfaces is associated with the user device.

Abstract: A multicast group establishment method in a fat-tree network is disclosed, the method includes: sending, by a network node according to an address of a target multicast group, a join request to a management switching node of the target multicast group, where the join request is used to request to make the network node join the target multicast group, and the management switching node of the target multicast group is a core switching node to which the address of the target multicast group belongs. The fat-tree network includes multiple switching nodes, a top layer is a core switching node, and each core switching node manages multiple multicast group addresses.

Abstract: Systems and methods for providing default multicast local identifier values per subnet partition as additional subnet management agent attributes in a high performance computing environment. In accordance with an embodiment, a single multicast local identifier can be associated with multiple multicast group identifiers when the multiple multicast group identifiers are associated with a same subnet partition. That is, a default MLID for each subnet partition (i.e. either all subnet partitions or those subnet partitions that are defined to have multicast communication) can be provided. A subnet node can have a default multicast local identifier table for storing default multicast local identifiers. A relationship between the partition key table of the node and the default multicast local identifier table of the nod can map default assigned multicast local identifiers to relevant partitions.

Abstract: Embodiments of the present application disclose a method for advertising information about a PE device and a related apparatus, which are applied to a BGP-based MVPN. The method includes: a PE device in the MVPN determines a role of the PE device in the MVPN and a working status corresponding to the role, where the role is a root node and/or a leaf node, and the working status is active or standby; the PE device generates a BGP update message, carrying an IP address of the PE device, an identifier of the MVPN, the role of the PE device, and the working status of the PE device; and the PE device sends the BGP update message to a BGP neighbor of the PE device in the MVPN. The method and apparatus may alleviate a configuration workload and reduce an error occurrence during an application of a BGP-based MVPN.

Abstract: According to one aspect, a method includes an Intercloud Fabric Switch (ICS) included in a public cloud and an ICS cluster obtaining a packet, and determining if the packet is obtained from a site-to-site link that links the ICS to an enterprise datacenter. If the packet is obtained from the site-to-site link, it is determined whether the packet is an unknown unicast packet. If the packet is an unknown unicast packet, the packet is dropped, and if not, the packet is provided to an access link that links the ICS to a virtual machine. If the packet is not obtained from the site-to-site link, it is determined whether the packet is obtained from an inter-ICS link that allows the ICS to communicate with the ICS cluster. If the packet is obtained from the inter-ICS link, the packet is dropped if it is an unknown unicast packet.

Abstract: Systems and methods configuring a process that uses IPv4 communication without associating the IPv4 process with an IP loopback address are disclosed. Embodiments may include receiving a message to configure an IPv4 process. Embodiments may also include determining if a configuration parameter associated with the received message is set to indicate that one or more default IP loopback addresses are to be used as IP loopback addresses to associate with the IPv4 process. Embodiments may further include completing the configuration of the IPv4 process without associating an IP loopback address with the IPv4 process when the configuration parameter is determined to not be set to indicate that one or more default IP loopback addresses are to be used as IP loopback addresses to associate with the IPv4 process.

Abstract: A method for method for streamlining communications between groups of users uses multiple user accounts and a remote server. Each user account is associated to a user PC device. Additionally, the method associates each user with a representative group. The user account is identified as a primary account or a secondary account. As a result, primary accounts are granted greater messaging permissions than secondary accounts. To that end, the secondary accounts are able to broadcast messages to all of the primary accounts that are associated to a representative group. The primary accounts on the other hand are able to selectively send messages to one or more individual secondary accounts. Moreover, the primary accounts are able to broadcast messages to the primary accounts that are associated to any of the representative groups.

Abstract: The present invention relates to a device comprising at least one computer machine and a software for executing a correlation analysis mechanism during performance degradation of an application chain comprising a hardware and software arrangement for storing a measuring repository, said measuring repository comprising a hardware and software arrangement for measuring, by consumption probes, the level of use of each resource on the set of servers constituting the application chain during periods of performance degradation, then storage in a memory of these levels in the measuring repository, in association with the period, said device being characterized in that it further comprises an editing hardware and software arrangement of a configuration repository, a categorization module of the performance problems as a function of the measuring and configuration repositories.

Abstract: A network interface of a network user having at least one physical interface for connecting the network interface to a network interface of a different network user, at least one data selector, which is connected to the physical interface and which is suitable for receiving data from the physical interface and sending data to the physical interface, and at least one data switch, which is connected to the data selector and which is suitable for receiving data from the data selector and sending data to the data selector.

Abstract: A method of batch processing implemented using a pre-classification mechanism of a network interface controller is provided. The method includes receiving, by the pre-classification mechanism of the network interface controller, rules for populating a first receive queue and a second receive queue, receiving, by the pre-classification mechanism, packets corresponding to a first protocol intermingled with packets corresponding to a second protocol, and populating, by the pre-classification mechanism of the network interface controller, the first receive queue exclusively with the packets corresponding to the first protocol and the second receive queue exclusively with the packets corresponding to the second protocol based on the rules such that a central processing unit is able to obtain a first batch of the packets from the first receive queue and process each of the packets from the first batch using a same node graph path.

Abstract: Exemplary methods include a first network device participating in an election process to determine a designated bit forwarding router (D-BFR). The methods include in response to determining the first network device is elected to be the D-BFR, performing D-BFR operations comprising assigning one or more bitmask positions (BMPs) to one or more bit forwarding egress routers (BFERs) and advertising the assigned one or more BMPs. The method may further include one or more of determining an elected bitmask (BM) length based on maximum local BM lengths advertised by other BFRs and determining an elected tree type based on supported tree types advertised by other BFRs.

Abstract: Message summarization and flood suppression may be provided. A proxy (e.g., an IGMP Proxy) process may be used to reduce the flooding of messages (e.g., IGMP messages) over a network (e.g., an EVPN network). A triggering process may also be provided for provider edge (PE) devices to setup their underlay multicast tunnels. The proxy may comprise two components: i) a proxy for reports (e.g., IGMP reports); and ii) a proxy for queries (e.g., IGMP Queries).

Abstract: The present disclosure relates to a method for message flooding in a wireless communication system, and is characterized in that it includes the steps of: broadcasting, by a transmission node, a flooding message to a surrounding node through a first common channel; checking, by a reception node, whether the first common channel has received the flooding message; transmitting, by the reception node, an ACK signal through a second common channel or a NACK signal through a third common channel, based on a result of the check; and determining, by the transmission node, whether the second common channel or the third common channel has detected a signal, to determine whether to re-transmit the flooding message.

Abstract: A client and content provider are connected by a plurality of simultaneous transport connections. The number of the transport connections that are used to transfer data is selected based on the size of the data to be transferred and may change after transfer of data has commenced based on the amount of data left and the attributes of the transport connections. In another aspect, data to be transmitted over the transport connections is organized into frames such that each frame includes data from only one data stream. The frames are sized to be less than or equal to a control window of the transport connection over which they are transmitted. Each frame may be assigned to a transport connection in a round robin fashion or based on the size of the frame and the sizes of the control windows of the transport connections.

Abstract: A system includes a plurality of servers that are configured to execute a distributed matching system for the trading of one or more items.

Abstract: A method for processing an intrusion in a communication network including a plurality of node equipment, including a current node, which: discovers of a neighborhood of the current node, including assigning a resilience group to the neighboring node, according to at least one piece of information representative of a resilience level of the neighboring node to at least one type of attack; detecting an intrusion affecting at least one suspect node of the neighborhood of the current node; establishing a consensus concerning the at least one suspect node in a neighborhood by counting a number of resilience groups having detected the intrusion in the neighborhood of the suspect node and a total number of resilience groups represented in the neighborhood of the suspect node; and deciding to change a status of the suspect node based on a result of the consensus by comparison of both numbers.

Abstract: The present disclosure describes traffic forwarding in a network where a virtual local area network (VLAN) exists in multiple network sites that are geographically dispersed and connected via virtual links. A first edge device (ED) at a first site receives, from a second ED at a second site, information identifying a VLAN of the second ED through a virtual link. The VLAN identified in the information is an active VLAN of the second ED. An association between (i) the VLAN of the second ED and (ii) the virtual link through which the information is received from the second ED is recorded. When forwarding traffic to the VLAN of the second ED, it is determined the virtual link associated with the VLAN of the second ED is determined based on the recorded association and the traffic is forwarded via the determined virtual link.

Abstract: A CSI measurement method, a CSI acquisition method, a terminal and a network device are provided, to measure CSI. The terminal performs interference measurement to acquire initial interference measurement result, determine the CSI based on the initial interference measurement result and an interference measurement parameter configured by the network device, and then report the determined CSI to the network device. The interference measurement parameter is configured by the network device, so the interference measurement parameter configured by the network device is capable of reflecting an actual situation of interference information, and interference information acquired by the terminal based on the interference measurement parameter matches interference information for actual transmission in a better manner.

Abstract: Techniques for bridging communication between multiple networks to facilitate distribution of multicast data are described herein. For example, the techniques may identify a particular node that is able to communicate with each network and instruct the particular node to act as a bridge between the networks. The particular node may forward data from either network to the other network. This may allow data for a multicast group to be distributed to members of the multicast group that are located in separate networks.

Abstract: A forwarding pipeline of a forwarding engine includes a mirror bit mask vector with one bit per supported independent mirror session. Each bit in the mirror bit mask vector can be set at any point in the forwarding pipeline when the forwarding engine determines that conditions for a corresponding mirror session are met. At the end of the forwarding pipeline, if any of the bits in the mirror bit mask vector is set, then a packet, the mirror bit mask vector and a pointer to the start of a mirror destination linked list are forwarded to the multicast replication engine. The mirror destination linked list typically defines a rule for mirroring. The multicast replication engine mirrors the packet according to the mirror destination linked list and the mirror bit mask vector.

Abstract: Communication apparatus includes a plurality of interfaces for receiving and transmitting data packets from and to a network and a memory, which receives and stores context data with respect to multicast groups. Packet processing circuitry establishes reliable connections over the network with the receiving nodes in the multicast groups, and upon receiving from a packet source on the network an incoming unicast packet containing multicast data and containing multicast metadata that identifies a multicast group, sends an acknowledgment of the incoming unicast packet to the packet source, reads the context data from the memory with respect to the identified multicast group, and transmits multiple outgoing unicast packets containing the multicast data via respective egress interfaces to the receiving nodes in the multicast group over the reliable connections.