Abstract: Dynamic buffer selection in ethernet controllers including determining, by an ethernet controller, based on a received header of a packet, a length of the packet; selecting, from a plurality of buffers, a buffer of a buffer size based on the length of the packet, wherein the plurality of buffers comprises a plurality of different buffer sizes; and transferring the packet to the selected buffer.

Abstract: A method includes receiving, at an unmanaged switch, a link-aggregation control protocol (“LACP”) protocol data unit (“PDU”) on each port of two or more connections to be in a link-aggregation group (“LAG”). The ports are in the unmanaged switch, which is unconfigured for LACP and is connected over the connections to a managed switch configured for LACP. The method includes, in response to the ports that received an LACP PDU being unconfigured for LACP, configuring each port receiving an LACP PDU for LACP, creating a LAG that includes the connections of the ports that received the LACP PDUs, and starting an LACP timer. The method includes, in response to determining that the LACP timer has expired, clearing the LACP configuration of the ports configured for LACP. The LACP timer expires in response to the ports in the LAG not receiving additional LACP PDUs prior to expiration of a timer period.

Abstract: Secure updating of strong passwords in a composable system, includes: during deployment of the composable system, requesting, by a central controller in the composable system, device information from a device vendor, the device information specifying a number of expected devices for the composable system; checking, by the central controller, a number of discovered devices against the number of expected devices; only upon discovering all the expected devices, updating, by the central controller, a password for all the devices in the composable system; and upon discovering fewer than the total number of expected devices, halting, by the central controller, deployment of the composable system.

Abstract: A method for transmitting a packet vector across compute nodes implementing a packet processing graph on a vector packet processor is disclosed. The method includes determining that a packet vector processed by a previous graph node in a first compute node is ready to be processed by a next graph node in a second compute node. The packet vector includes a plurality of data packets and the previous and next graph nodes are graph nodes of a packet processing graph implemented as a directed acyclic graph (“DAG”) that extends across the first and second compute nodes. The first and second compute nodes each run an instance of a vector packet processor. The method includes transmitting the packet vector from the first compute node to the second compute node using remote direct memory access (“RDMA”).

Abstract: A method for reducing the number of IP addresses used by VRRP includes creating a VRRP control packet at a router of a plurality of physical routers configured in parallel in a data pathway. The physical routers are configured as a virtual router and the VRRP control packet is configured as an IPv6 control packet. The method includes inserting an IP address of the virtual router in a router address field of a control packet header. The IP address is formatted as an IPv4 address. The method includes inserting, in a reserved field of the control packet, a first value. The first value differs from a default value for the reserved field signifying that IP addresses in the router address field are formatted as IPv6 addresses for IPv6 control packets and the first value signifies that the IP addresses in the router address field are formatted as IPv4 addresses.

Abstract: Optimizing network load in multicast communications including receiving, at a first interface of a first router from a second router, a multicast packet for a multicast group; determining whether a time to live (TTL) value of the multicast packet is equal to zero; and sending, in response to determining that the TTL value of the multicast packet is equal to zero, a Protocol Independent Multicast prune message on the first interface of the first router, wherein the Protocol Independent Multicast prune message comprises a request to stop receiving multicast traffic for the multicast group.

Abstract: Dynamic buffer selection in ethernet controllers including determining, by an ethernet controller, based on a received header of a packet, a length of the packet; selecting, from a plurality of buffers, a buffer of a buffer size based on the length of the packet, wherein the plurality of buffers comprises a plurality of different buffer sizes; and transferring the packet to the selected buffer.

Abstract: A method for detecting an open circuit on a cable for 1GBASE-T and/or 10GBASE-T Ethernet communications includes transmitting one or more autonegotiation signals during an attempt to identify an open cable. The autonegotiation signals are sent from an autonegotiation-capable Ethernet transceiver connected to a cable with a twisted pair. The method includes receiving one or more waveforms from the cable, analyzing the received one or more waveforms, and transmitting an open cable alert in response to determining that the received one or more waveforms are indicative of a reflected autonegotiation signal.

Abstract: Methods, apparatuses, and computer program products for fast resumption of dormant sessions on a client device are disclosed. A client device receives a push notification from a push notification server, the push notification having a payload that includes a datagram packet from an application server that initiated the push notification. A push notification process on the client device extracts the datagram packet from the push notification and injects the datagram packet into a communications protocol stack for at least one application on the client device. The application is awakened from a dormant state and reads the datagram packet. The application may then send a response to the datagram packet to the application server.

Abstract: A method for reducing transmission of telemetry packets includes receiving, at a network controller, telemetry packets from network nodes in a communication pathway of data packets being transmitted from a sending host to a receiving host and determining from the telemetry packets a number of network nodes in the communication pathway. The method includes setting a network node threshold value equal to the determined number of network nodes in the communication pathway, and transmitting, from the network controller, the network node threshold value to each network node in the communication pathway.

Abstract: A method for in-band telemetry rate limiting is disclosed. An apparatus and computer program product also perform the functions of the method. Embodiments of the method include receiving a telemetry packet at an egress node of a data pathway in a data network. The data pathway includes an ingress node, transit nodes and the egress node and the telemetry packet includes a telemetry parameter where the telemetry parameter is for a node in the data pathway. The method includes comparing the telemetry parameter with a previous telemetry parameter from a previous telemetry packet received at the egress node prior to receiving the telemetry packet, and transmitting telemetry data from the telemetry packet to a network controller in response to a telemetry difference exceeding a telemetry parameter threshold. The telemetry difference includes a difference between the telemetry parameter and the previous telemetry parameter.

Abstract: A method for tunning high speed link parameters for cables with a transceiver at each end includes transmitting, in response to establishing a link, an eye pattern bit sequence from a first computing device to a second computing device over a cable with a transceiver connected to each end of the cable. Establishing the link includes establishing a communication link over the cable between the first computing device and the second computing device. The method includes receiving an eye pattern quality message from the second computing device using a device discovery protocol. The eye pattern quality message includes quality information of a received waveform corresponding to the transmitted eye pattern bit sequence. The method includes adjusting transmission parameters for transmission over the cable in response to the quality information, and retransmitting the eye pattern bit sequence using the adjusted transmission parameters.

Abstract: A method for reducing the number of IP addresses used by VRRP includes creating a VRRP control packet at a router of a plurality of physical routers configured in parallel in a data pathway. The physical routers are configured as a virtual router and the VRRP control packet is configured as an IPv6 control packet. The method includes inserting an IP address of the virtual router in a router address field of a control packet header. The IP address is formatted as an IPv4 address. The method includes inserting, in a reserved field of the control packet, a first value. The first value differs from a default value for the reserved field signifying that IP addresses in the router address field are formatted as IPv6 addresses for IPv6 control packets and the first value signifies that the IP addresses in the router address field are formatted as IPv4 addresses.

Abstract: A method for a VXLAN extension header for segment routing includes receiving a data packet at a node of a data pathway between an ingress node and an egress node. The data packet includes a tunneling protocol header followed by a segment routing header ahead of a payload and the segment routing header includes a list of identifiers of nodes in the data pathway. The method includes processing the segment routing header to address the data packet to be transmitted to a next node listed in the segment routing header and to update a pointer in the segment routing header, and transmitting the data packet to a next node based on the address of the next node.

Abstract: Methods, apparatuses, and computer program products for edge device assisted mitigation of publish-subscribe denial of service (DoS) attacks are disclosed. An edge device hosts a virtualized copy of an Internet-of-Things (IoT) device subscribed to one or more publish-subscribe topics. When the edge device receives an indication to activate the virtualized copy of the IoT device, for example, during a DoS attack on the IoT device, the edge device activates the virtualized copy of the IoT device, which receives traffic from the publish-subscribe topic. The virtualized copy of the IoT device applies security policies to incoming traffic received from the subscription topics and transmits to the IoT device sanitized traffic obtained from the received incoming subscription content traffic.

Abstract: A method for link layer authentication includes receiving, at an edge network access node, a link layer authentication packet from a client, seeking network access, using a remote NAS agent running on the edge network access node. The method transmits, using a tunneling connection, the link layer authentication packet to a remote NAS in a link layer authentication process. The link layer authentication process exchanges the link layer authentication packet with an authentication server to authenticate the client. The method includes receiving a link layer authentication packet from the remote NAS over the tunneling connection. The received link layer authentication packet includes a response from the authentication server regarding the transmitted link layer authentication packet.

Abstract: A method for storage device compression/deduplication layer congestion control includes comparing a level of congestion at a compression/deduplication layer of a storage device with a lower congestion limit. The compression/deduplication layer is configured for data compression and/or data deduplication of a data packet prior to storing data of the data packet. The data packet is received from a computing device using transmission control protocol (“TCP”). The method includes, in response to the level of congestion exceeding the lower congestion limit, marking as congested, in a flow table, a data flow between the computing device and the storage device. The flow table is shared between the compression/deduplication layer and a TCP layer of the storage device.

Abstract: Providing a hybrid virtual network, includes: receiving from a source VM, by a hypervisor of a first stack, a packet to be transmitted to a target VM within a virtual network includes multiple VMs spanning a multiple stacks, where each stack includes an aggregation of compute, storage, and network resources and separate stacks are coupled for data communications via a network level protocol; if the target VM is located in the first stack, overwriting a MAC address of in the packet with a replacement MAC address for transmission via a data link layer protocol and transmitting the packet with the data link layer protocol; and if the target VM is not located in the first stack, encapsulating the packet for tunneling via a network layer protocol and transmitting the encapsulated packet with the network layer protocol.

Abstract: A method includes receiving, at an unmanaged switch, a link-aggregation control protocol (“LACP”) protocol data unit (“PDU”) on each port of two or more connections to be in a link-aggregation group (“LAG”). The ports are in the unmanaged switch, which is unconfigured for LACP and is connected over the connections to a managed switch configured for LACP. The method includes, in response to the ports that received an LACP PDU being unconfigured for LACP, configuring each port receiving an LACP PDU for LACP, creating a LAG that includes the connections of the ports that received the LACP PDUs, and starting an LACP timer. The method includes, in response to determining that the LACP timer has expired, clearing the LACP configuration of the ports configured for LACP. The LACP timer expires in response to the ports in the LAG not receiving additional LACP PDUs prior to expiration of a timer period.

Abstract: Secure updating of strong passwords in a composable system, includes: during deployment of the composable system, requesting, by a central controller in the composable system, device information from a device vendor, the device information specifying a number of expected devices for the composable system; checking, by the central controller, a number of discovered devices against the number of expected devices; only upon discovering all the expected devices, updating, by the central controller, a password for all the devices in the composable system; and upon discovering fewer than the total number of expected devices, halting, by the central controller, deployment of the composable system.