This disclosure is related to devices, systems, and techniques for controlling a flow of network traffic between two or more devices. For example, a network device includes a control unit, a forwarding unit including a forwarding manager including a server configured to store a tunnel hierarchy structure, and a packet processor. The packet processor is configured to receive, via an interface card of a set of interface cards, a packet including a packet header. The forwarding manager is configured to parse the packet header in order to identify a service corresponding to the packet, wherein the service is associated with a service color, identify, based on the tunnel hierarchy structure, a set of tunnels, wherein each tunnel of the set of tunnels is associated with a tunnel color, and determine whether at least one tunnel of a set of tunnels associated with the tunnel color is in an online state.

A method supporting distributed resilient network interconnect (DRNI) in a link aggregation group at a network device is disclosed. The method starts with encapsulating a distributed relay control protocol data unit (DRCPDU) in a frame, wherein the DRCPDU includes a protocol data unit (PDU) structure. The PDU structure includes a type field indicating that the DRCPDU is for DRCP, a version field indicating a version number of the DRCP, and a set of type/length/values (TLVs) including: a terminator TLV indicating an end of the PDU structure, a portal information TLV indicating characteristics of the first portal, a portal configuration information TLV indicating configuration information of the first portal, a DRCP state TLV indicating variables associated with an intra-portal link (IPP), a home ports information TLV and a neighbor ports information TLV. The method continues with transmitting the frame encapsulating the DRCPDU from the network device to a neighbor network device.

Systems and methods for supporting inter subnet control plane protocol for consistent multicast membership and connectivity across multiple subnets in a high performance computing environment. In accordance with an embodiment, by associating a multicast group with an inter-subnet partition, and enforcing a dedicated router port for the multicast group, multicast loop avoidance can be provided for between connected subnets. Because only a single router port is selected as being capable of handling the MC packet, no other router port in the subnet can then pass a multicast packet back to the originating subnet.

Embodiments of a device and method are disclosed. In an embodiment, a method of communications involves allocating communications devices of a wired communications network to clusters, assigning addresses to the clusters, where each communications device within one of the clusters has an identical address, and conducting communications between the communications devices based on the addresses assigned to the clusters.

Systems and methods for supporting SMA level abstractions at router ports for inter-subnet exchange of management information in a high performance computing environment. In accordance with an embodiment, a subnet manager in a local subnet is responsible for establishing and configuring a remote attribute a switch having a switch port configured as a router port. This remote attribute can comprise certain information about the local subnet, including connectivity information and port status information. On receiving a query from a remote subnet manager, via a SMP (or a vendor specific SMP), information contained in the remote attribute can be communicated back to the remote subnet manager.

A switch architecture enables ports to stash packets in unused buffers on other ports, exploiting excess internal bandwidth that may exist, for example, in a tiled switch. This architecture leverages unused port buffer memory to improve features such as congestion handling and error recovery.

A packet processing device includes a first unit, a second unit, and a switching unit. The first unit counts the number of arrived packets in a first period that is from the time slot present after a priority section up to the end of the initial time slot in the subsequently-arriving priority section. When the counted number of arrived packets is positive, the first unit determines that forward mismatch has occurred in an observation cycle. The second unit counts the number of arrived packets in a second period which is from the time slot present immediately after the priority section in the first period of time up to the end of the initial time slot of burst sections in the subsequently-arriving priority section. When the counted number of arrived packets is “0”, the second unit determines that backward mismatch has occurred in the observation cycle.

Example methods, apparatus, systems and articles of manufacture (e.g., physical storage media) to change a time sensitive networking schedule implemented by a softswitch are disclosed. Example apparatus disclosed herein to change a time sensitive networking schedule implemented by a first softswitch on a compute node include a network node configurator to deploy a second softswitch on the compute node based on a first configuration specification associated with the first softswitch, configure the second softswitch to implement an updated time sensitive networking schedule different from the time sensitive networking schedule implemented by the first softswitch, and replace the first softswitch with the second softswitch in response to a determination that a first set of constraints is met for simulated network traffic processed by the second softswitch based on the updated time sensitive networking schedule. Disclosed example apparatus also include a simulator to apply the simulated network traffic to the second softswitch.

In general, this disclosure describes a network device to determine a cause of packets being dropped within a network. An example method includes generating, by a traffic monitor operating on a network device, an exception packet that includes a unique exception code that identifies a cause for a component in the network device to discard a transit packet, and a nexthop index identifying a forwarding path being taken by the transit packet experiencing the exception. The method also includes forwarding the exception packet to a collector to be processed.

[Problem] Efficiently utilizing physical resources in a communication system that builds a virtual network based on various requirements.

[Solution] A communication system (10) includes a Spine switch group (12) consisting of a plurality of Spine switches (102), a Leaf switch group (14) consisting of a plurality of Leaf switches (104), a plurality of servers (106) connected to any one of the plurality of Leaf switches (104), and a controller (110) configured to build a virtual network on the physical resources. At least one of the Spine switch group (12) and the Leaf switch group (14) is constituted by a mix of switch devices having different performance. The controller (110) selects physical resources to be used for building the virtual network based on the desired performance of the virtual network.