Abstract: Techniques for dynamic routing based on application load are described herein. The techniques may include receiving load information associated with resources of an application orchestration system that are allocated to host an application, the resources associated with different geographical regions. Based at least in part on the load information, a network controller may determine that first resources of the application orchestration system are less constrained than second resources of the application orchestration system, the first resources associated with a first geographical region and the second resources associated with a second geographical region. Based at least in part on the first resources being less constrained than the second resources, application traffic may be routed through the network to the application hosted by the first resources in the first geographical region.

Abstract: Techniques for dynamic routing based on application load are described herein. The techniques may include receiving load information associated with resources of an application orchestration system that are allocated to host an application, the resources associated with different geographical regions. Based at least in part on the load information, a network controller may determine that first resources of the application orchestration system are less constrained than second resources of the application orchestration system, the first resources associated with a first geographical region and the second resources associated with a second geographical region. Based at least in part on the first resources being less constrained than the second resources, application traffic may be routed through the network to the application hosted by the first resources in the first geographical region.

Abstract: For a seamless and robust artificial intelligence-based assistant experience, an intent-based query and response router has been designed to operate as an intelligent layer between a user and multiple backend services that may respond to one or more queries over the course of a conversation with the user. The query router interacts with an intent classification service to obtain an intent classification for a prompt that is based on a user query. The query router uses the intent classification, which is used as an identifier of a backend service, to route the user query to an appropriate one (or more) of the backend services. When a response is detected, the query router determines a corresponding conversation and provides the response for the conversation.

Abstract: Systems and methods are provided for receiving bandwidth metrics from a plurality of routers on respective link routes in a network, compiling a link database including the bandwidth metrics of each respective link route in the network, selecting a first designated link path from the link database between a first router and a second router based on an application routing policy, the application routing policy being based on a routing metric, providing a first multiprotocol label switching label based on the first designated link path to the first router of the plurality of routers in the network, and restricting network traffic of the first router to the first designated link path provided in the first multiprotocol label switching label.

Abstract: Systems and methods are provided for receiving bandwidth metrics from a plurality of routers on respective link routes in a network, compiling a link database including the bandwidth metrics of each respective link route in the network, selecting a first designated link path from the link database between a first router and a second router based on an application routing policy, the application routing policy being based on a routing metric, providing a first multiprotocol label switching label based on the first designated link path to the first router of the plurality of routers in the network, and restricting network traffic of the first router to the first designated link path provided in the first multiprotocol label switching label.

Abstract: Techniques for dynamic routing based on application load are described herein. The techniques may include receiving load information associated with resources of an application orchestration system that are allocated to host an application, the resources associated with different geographical regions. Based at least in part on the load information, a network controller may determine that first resources of the application orchestration system are less constrained than second resources of the application orchestration system, the first resources associated with a first geographical region and the second resources associated with a second geographical region. Based at least in part on the first resources being less constrained than the second resources, application traffic may be routed through the network to the application hosted by the first resources in the first geographical region.

Abstract: Presented herein are embodiments for quickly identifying and recommending key performance indicators (KPIs) for network devices based on the type of network device and/or role of the device. The type or configuration of the network device may be obtained and compared to the capabilities of the network device. Operational or performance information of the network device, represented by strings, may be obtained based on the configuration information. Operational information that is not relevant to the configuration of the network device may be filtered out. The remaining operational information may be ranked as KPIs based on a relevance of the operational information with respect to the configuration information.

Abstract: Presented herein are methodologies for tracking a link state of a physical network connection and selectively reporting the link state to virtual machines that rely on the physical network connection. A method includes receiving an indication, at a hypervisor, which is running on a host computer and which instantiates a virtual switch, that a physical link interconnecting the host computer to a network has failed; determining that the physical link serves the virtual switch; determining whether link state tracking is enabled for the physical link; and when link state tracking is enabled for the physical link, notifying a virtual machine, which is running on the host computer and which is being served by the virtual switch, that a connection between the virtual machine and the network has failed.

Abstract: Systems and methods are provided for receiving bandwidth metrics from a plurality of routers on respective link routes in a network, compiling a link database including the bandwidth metrics of each respective link route in the network, selecting a first designated link path from the link database between a first router and a second router based on an application routing policy, the application routing policy being based on a routing metric, providing a first multiprotocol label switching label based on the first designated link path to the first router of the plurality of routers in the network, and restricting network traffic of the first router to the first designated link path provided in the first multiprotocol label switching label.

Abstract: Techniques for obtaining application network metadata from a service registry so that a network routing policy may be derived for traffic associated with the application are described herein. The techniques may include receiving, at a service registry, network metadata associated with traffic of an application hosted by a scalable application service platform. The techniques may also include obtaining, by a controller of a network and from the service registry, the network metadata associated with the traffic of the application. Based at least in part on the network metadata, the controller may determine a routing policy that is optimized for sending the traffic through the network. Additionally, the controller may send an indication of the routing policy to a node of the network or otherwise provision the network such that the traffic of the application is sent through the network according to the routing policy.

Abstract: In one embodiment, an edge device located at an edge of a local network provides connectivity between the local network and a cloud-based software as a service (SaaS) provider via one or more interfaces. The edge device obtains telemetry data associated with the edge device for a plurality of metrics. The edge device makes a determination that one or more of the plurality of metrics is anomalous. The edge device sends, based on the determination, an indication of the determination to the SaaS provider. The SaaS provider uses the indication to determine a root cause of an application served by the SaaS provider experiencing degraded application performance.

Abstract: Presented herein are methodologies for tracking a link state of a physical network connection and selectively reporting the link state to virtual machines that rely on the physical network connection. A method includes receiving an indication, at a hypervisor, which is running on a host computer and which instantiates a virtual switch, that a physical link interconnecting the host computer to a network has failed; determining that the physical link serves the virtual switch; determining whether link state tracking is enabled for the physical link; and when link state tracking is enabled for the physical link, notifying a virtual machine, which is running on the host computer and which is being served by the virtual switch, that a connection between the virtual machine and the network has failed.

Abstract: A method for managing routing tables and data packet forwarding is disclosed. The method comprises obtaining, at a networking device, a first outgoing label associated with one or more output port identifiers of the networking device. The first outgoing label identifies a first destination node. The one or more output port identifiers identify one or more of a plurality of output ports. The method further comprises determining whether or not the one or more output port identifiers are also associated with a second outgoing label. The second outgoing label identifies a second destination node different from the first destination node. The method further comprises merging, into a next hop table allocated in a non-transitory memory, the first outgoing label with the second outgoing label in response to determining that the one or more output port identifiers are also associated with the second outgoing label.

Abstract: Local switching may be provided over a flexible cross-connect VLAN-aware based service. First, a failure of a first segment link of a first segment may be detected by a first network device. The first segment may have a first segment identifier and the first segment link may be connected to the first network device. Next, a route withdraw indicating the first segment identifier may be sent by the first network device in response to detecting the failure of the first segment link of the first segment. A second network device may then receive the route withdraw. Then, the second network device may forward, in response to receiving the route withdraw, traffic received from the first network device to a second segment link of the first segment connected to the second network device. The traffic may be received from the first network device over a backup connection over an Ethernet Virtual Private Network (EVPN) core.

Abstract: A method is implemented by a node of a network domain having a plurality of nodes, where the node functions as an ingress node of the network domain for a data flow. The method enables data flow analysis across the network domain. The method includes receiving a data packet belonging to a data flow at an ingress node of the network domain, determining whether the data packet is to be marked for the data flow analysis, determining whether an egress node of the network domain for the data flow supports data packet marking, marking the data packet with a marking label indicating to supporting nodes in the network domain that the data packet is to be processed for data flow analysis, and forwarding the data packet toward the egress node of the network domain.

Abstract: A method is implemented by a network device to establish a two-way active measurement protocol (TWAMP) test session that supports use of a Network Time Protocol (NTP) timestamp format and a Precision Time Protocol Version 2 (PTPv2) timestamp format. The network device acts as a server that communicates with a control-client to establish a TWAMP test session between a sender and a reflector. The method includes sending a server greeting message to the control-client indicating timestamp formats that the reflector can set, receiving a set-up-response message from the control-client indicating timestamp formats that the sender can interpret, checking whether the sender supports the multiple timestamp format extensions to TWAMP, configuring the reflector to set timestamps in a format that the reflector can set and the sender can interpret if the sender supports multiple timestamp format extensions to TWAMP, and sending a server-start message to the control-client.

Abstract: Local switching may be provided over a flexible cross-connect VLAN-aware based service. First, a failure of a first segment link of a first segment may be detected by a first network device. The first segment may have a first segment identifier and the first segment link may be connected to the first network device. Next, a route withdraw indicating the first segment identifier may be sent by the first network device in response to detecting the failure of the first segment link of the first segment. A second network device may then receive the route withdraw. Then, the second network device may forward, in response to receiving the route withdraw, traffic received from the first network device to a second segment link of the first segment connected to the second network device. The traffic may be received from the first network device over a backup connection over an Ethernet Virtual Private Network (EVPN) core.

Abstract: In one embodiment, a device in a network assigns two multi-protocol label switching (MPLS) labels to a single flexible cross-connection maintained by the device for a plurality of endpoint. The device establishes an Ethernet virtual private network (EVPN) backup path to a peer router for the flexible cross-connection in part by exchanging the MPLS labels with the peer router. The device detects a communication failure that resulted from attempting to send a packet between a first one of the endpoints and a second one of the endpoints using local switching on the device and via the flexible cross-connection. The device sends the packet towards the second endpoint using the EVPN backup path to the peer router, after detecting the communication failure and without first using Border Gateway Protocol (BGP) signaling to signal the communication failure to the peer router.

Abstract: A method is implemented by a node for enabling shortest path bridging in a network that is scalable to support sixteen million virtual local area network (VLAN) identifiers using multiprotocol label switching (MPLS) encapsulation. The method comprises allocating a tunnel label using a distributed tunnel label allocation algorithm, allocating a source label using a distributed source label allocation algorithm, assigning a bridge instance a service label, and distributing the tunnel label, source label and service label to other nodes in the MPLS network.

Abstract: A method for managing routing tables and data packet forwarding is disclosed. The method comprises obtaining, at a networking device, a first outgoing label associated with one or more output port identifiers of the networking device. The first outgoing label identifies a first destination node. The one or more output port identifiers identify one or more of a plurality of output ports. The method further comprises determining whether or not the one or more output port identifiers are also associated with a second outgoing label. The second outgoing label identifies a second destination node different from the first destination node. The method further comprises merging, into a next hop table allocated in a non-transitory memory, the first outgoing label with the second outgoing label in response to determining that the one or more output port identifiers are also associated with the second outgoing label.