Abstract: Techniques are disclosed for session-based routing of multipoint Open Systems Interconnection (OSI) Model Layer-2 (L2) frames of an L2 network extended over Layer-3 (L3) networks. In one example, L2 networks connect a source device to an ingress router and receiver devices to egress routers. An L3 network connects the ingress and egress routers. The ingress router receives, from the source device, a multipoint L2 frame destined for the receiver devices. The ingress router forms, for each egress router that is connected to at least one multipoint receiver device, a unicast L3 packet for the L2 frame and forwards the unicast L3 packet to the egress router. Each egress router generates, in response to receiving the unicast L3 packet, the multipoint L2 frame and forwards, to the receiver devices, the multipoint L2 frame.

Abstract: Techniques are described for learning an unknown virtual network information, such as an virtual Internet Protocol (IP) address, of a pod in a virtual network. In some examples, a virtual router executing at a computing device may receive an Address Resolution Protocol (ARP) packet from a virtual execution element in the virtual network, the virtual execution element executing at the computing device. The virtual router may determine, based at least in part on the ARP packet, whether virtual network information for the virtual execution element in a virtual network is known to the virtual router. The virtual router may, in response to determining that the virtual network information of the virtual execution element in the virtual network is not known to the virtual router, perform learning of the virtual network information for the virtual execution element.

Abstract: Techniques are described for performing latency-aware load balancing. In some examples, a computing device communicably coupled to a plurality of service endpoints that are in motion with respect to the computing device may receive data to be processed. The computing device may select, based at least in part on a communication latency of each of the plurality of service endpoints and a predicted compute latency of each of the plurality of service endpoints, a service endpoint out of the plurality of service endpoints to process the data. The computing device may send the data to the selected service endpoint for processing.

Abstract: A system that permits or otherwise facilitates assessment of operational state of a computing component in a computing environment. In one example, this disclosure describes a method that includes collecting, by a server device that is executing within a server device cluster, metric information indicative of an operational state of the server device, wherein the metric information is associated with a plurality of virtual computing instances executing on the server device; analyzing, by the server device and based on the metric information, whether a first condition associated with a first virtual computing instance is satisfied; analyzing, by the server device and based on the metric information, whether a second condition associated with a second virtual computing instance is satisfied; and updating control information characterizing the operational state of the server device executing within the server device cluster.

Abstract: One exemplary aspect describes systems and methods for determining normal SLE behavior, determining when a SLE exhibits abnormal deterioration, and determining whether to take an action to mitigate what appears to be an indication of an abnormal SLE.

Abstract: A mixed pitch method of placing pads in a ball grid array (BGA) package having a BGA substrate and a plurality of connectors arranged in an array and connected via the pads to the BGA substrate. Selected pairs of the pads are placed on the BGA substrate at a distance defined by a first pitch P1. Ground pads are placed on the BGA substrate at a distance from the selected pairs of pads defined by a second pitch P2, wherein P2=M*P1 and M is greater than one. The selected pairs of the pads on the BGA substrate are also placed at a distance from other selected pairs of the pads defined by the second pitch P2.

Abstract: Photonic and electronic integrated circuits can be cooled using variable conductance heat pipes containing a non-condensable gas in addition to a phase-changing working fluid. To package the heat pipe with a subassembly including the integrated circuits in a standard housing providing a heat sink contact area, the heat pipe is oriented, in some embodiments, with its axis between evaporator and condenser ends substantially perpendicular to the direction along which the integrated circuit subassembly is separated from the heat sink contact area, and a portion of the exterior surface of the heat pipe is thermally insulated, with a suitable thermal insulation structure, from the heat sink contact area.

Abstract: A method includes maintaining, by a controller device that manages a plurality of network devices, a data structure and receiving, by the controller device, an indication of a stateful intent. The method further includes generating, by the controller device, a logical packet based on the stateful intent and determining, by the controller device, a dependent intent of the data structure based on the logical packet. The method further includes determining that the dependent intent conflicts with the stateful intent. The method further includes, in response to determining that the dependent intent conflicts with the stateful intent, determining, by the controller device, a resolution that resolves the conflict between the dependent intent and the stateful intent based on the dependent intent and the stateful intent and outputting, by the controller device, an indication of the resolution.

Abstract: Techniques are described in which a centralized controller, such as a software defined networking (SDN) controller, constructs a service chain that includes a physical network function (PNF) between a bare metal server (BMS) and a virtual execution element (e.g., virtual machine or container), or in some instances a remote BMS, or vice-versa. In accordance with the techniques disclosed herein, the controller may construct an inter-network service chain that includes PNFs, or a combination of PNFs and virtualized network functions (VNFs). The controller may construct an inter-network service chain to steer traffic between a BMS and a virtual execution element or remote BMS through an inter-network service chain using Virtual Extensible Local Area Network (VXLAN) as an underlying transport technology through the service chain.

Abstract: This disclosure describes techniques for configuring software defined network (SDN) controllers within different cloud computing domains and, in particular, a multi-cluster controller that operates and presents, in some examples, a single interface for seamlessly controlling and configuring SDN controllers in different cloud computing domains. In one example, this disclosure describes a system that includes a plurality of clusters, each of the plurality of clusters including a plurality of configurable endpoints; a storage system; and processing circuitry having access to the storage system and capable of communicating with each of the plurality of configurable endpoints. In some examples, the processing circuitry is configured to receive a plurality of requests, each specifying a configuration operation, identify, for each of the requests, a configuration cluster and a configuration endpoint within the configuration cluster, and perform, for each of the requests, the specified configuration operation.

Abstract: A controller device manages a plurality of network devices. The controller device includes one or more processing units configured to receive an upgrade request; obtain a redundancy model indicating network device redundancies or service redundancies; determine, based on the upgrade request and the redundancy model, an update graph having nodes each representing one of the network devices or one of the services, the update graph also having at least one edge that connects two of the nodes and that indicates the network redundancy or the service redundancy; determine, based on the update graph including the at least one edge, an update schedule for performing an update of software for the network devices that ensures availability for the first network device or second network device or availability for the service; and update the software of each of the network devices according to the update schedule.

Abstract: A fan tray, for a fan module of a network device chassis, may include an inner assembly that includes an inner cassette, one or more fans connected to the inner cassette, a first latch connected to the inner cassette and configured to removably connect to an outer assembly of the fan tray, and a fan controller connected to the inner cassette and configured to control operation of the one or more fans. The outer assembly may be configured to receive and retain the inner assembly, and may include an outer cassette with one or more openings configured to communicate with the one or more fans, a second latch connected to the outer cassette and configured to removably connect to a rear portion of the network device chassis, and an adaptor connected to the outer cassette and configured to connect and provide power to the fan controller.

Abstract: An example method includes receiving, by an SD-WAN system, WAN link characterization data for a plurality of WAN links of the SD-WAN system over a time period; and for each site of a plurality of sites of the SD-WAN system, generating, by the SD-WAN system, a local policy for the site, wherein generating the local policy is based on a machine learning model trained with the WAN link characterization data for the plurality of WAN links, and providing the local policy to an SD-WAN edge device of the site.

Abstract: A network device may receive configuration data identifying resource subscription thresholds associated with a plurality of respective endpoint devices and may receive traffic from the plurality of endpoint devices. The network device may process the traffic and the configuration data, with a resource allocation model, to determine that processing traffic associated with a first endpoint device requires allocating a resource of the network device, and may process the configuration data, with the resource allocation model, to identify the resource of the network device from a particular resource of the network device that is currently allocated to traffic associated with a second endpoint device. The network device may allocate the particular resource of the network device to the traffic associated with the first endpoint device, and may process the traffic associated with the first endpoint device with the particular resource to generate processed traffic.

Abstract: Example network devices, systems, and methods are disclosed. In an example, a network device includes memory configured to store information associated with one or more service level agreements (SLAs) for applications in a software-defined wide area network (SD-WAN) and an application-based multipath routing (AMR) module including processing circuitry. The AMR module is configured to identify, based on criteria, one or more of the applications for AMR, wherein each criterion of the criteria is associated with a corresponding property of an application. The AMR module is configured to determine a breach of one of the SLAs on each WAN link associated with a first application of the identified one or more applications. The AMR module is configured to apply, in response to determining the breach, AMR for the first application.

Abstract: A first device may receive content from a second device based on a request for the content. The first device may be located between the second device and a third device. The first device may determine a value for a portion of the content using a function, where the value is to be used to analyze the content. The value may uniquely identify the portion of the content. The first device may determine whether a classification of the content can be determined. The first device may selectively determine the classification of the content by providing the value or the portion of the content corresponding to the value, to a fourth device when the classification cannot be determined, or determine the classification of the content using a data store when the classification can be determined. The first device may perform an action with respect to the content.

Abstract: A first device may determine a prime integer for generating public keys, wherein the prime integer is based on a set of distinct prime integers. The first device may generate a set of multiple public keys based on a set of secret keys. The first device may combine the set of multiple public keys into a single value based on the prime integer and the set of distinct prime integers. The first device may transmit the single value to a second device.

Abstract: An ingress network device may receive a core domain network segment identifier associated with a core domain network of the multi-domain network. The ingress network device may receive location data of an egress network device associated with a second leaf domain network of the multi-domain network, wherein the location data may include data identifying the core domain network segment identifier, a second leaf domain network segment identifier associated with the second leaf domain network, and an egress network device segment identifier associated with the egress network device. The ingress network device may store the core domain network segment identifier and the location data, and may utilize the core domain segment identifier and the location data to route traffic to the egress network device.

Abstract: Compute nodes can execute virtual routers to implement a forwarding plane for one or more virtual networks having virtual network destinations hosted by the compute nodes. In one example, a method includes generating, by a software-defined networking (SDN) controller that manages a plurality of compute nodes, based on a unique identifier of a virtual network, a route target value for the virtual network, wherein the virtual network comprises virtual network endpoints executing on the compute nodes; and outputting, by the SDN controller and to a routing protocol peer device, a virtual private network (VPN) route that includes the route target value for the virtual network and a virtual network prefix associated with the virtual network, the VPN route for routing to the plurality of compute nodes executing the virtual network endpoints of the virtual network.

Abstract: In general, techniques are described for signaling IP path tunnels for traffic engineering using constraints in an IP network. For example, network devices, e.g., routers, of an IP network may compute an IP path using constraint information and establish the IP path using, for example, Resource Reservation Protocol, to signal the IP path without using MPLS. As one example, the egress router generates a path reservation signaling message that includes an egress IP address that is assigned for use by the routers on the IP path to send traffic of the data flow by encapsulating the traffic with the egress IP address and forwarding toward the egress router. As each router in the IP path receives the path reservation signaling message, the router configures a forwarding state to forward traffic encapsulated with the egress IP address to a next hop along the IP path toward the egress router.