Abstract: In one embodiment, a routing table of a router is populated with a plurality of prefixes. One or more performance characteristics of the plurality of prefixes are monitored. The router determines that a need exists to split a selected prefix of the plurality of prefixes. The router determines one or more boundaries upon which to split the selected prefix based on the monitored performance characteristics. The router then injects one or more more-specific prefixes into the routing table, each more-specific prefix referring to a smaller subset of nodes than the selected prefix, to split the selected prefix on the one or more boundaries to create a plurality of split prefixes. The split prefixes are controlled to optimize distribution of traffic.

Abstract: A technique performs route optimization of services provided by one or more service providers (SPs) for communication links that are combined in a computer network. According to the novel technique, performance characteristics and statistics (“link data”) for two or more combined links are merged to form link data for a single “virtual combined link.” Route optimization techniques (e.g., priority-based route optimization) may then be applied to one or more virtual combined links based on the combined link data, in addition to link data for zero or more conventional single links (physical or virtual). Illustratively, a route optimization technique may be used in accordance with the present invention to efficiently and accurately minimize costs associated with services provided by a plurality of SPs, wherein each SP charges for utilization of its service based on a tiered pricing structure.

Abstract: In one embodiment, probe messages may be transmitted from a local domain to a set of one or more responding probe targets located beyond a plurality of probed first-hop domains of the local domain, the probe messages transmitted via each of a plurality of probed first-hop domains. A reliability of each probed first-hop domain may then be determined based on responsiveness of probe messages transmitted via each respective probed first-hop domain, such that traffic destined beyond the probed first-hop domains may be load balanced across the probed first-hop domains based on the respective reliability, regardless of a traffic-class and a destination of the traffic.

Abstract: A technique dynamically increases the granularity of prefix control in a computer network. According to the novel technique, a routing table is populated with one or more reachable prefixes, e.g., learned or configured prefixes, and performance characteristics (e.g., bandwidth, cost, delay, etc.) of the prefixes are monitored. Upon determining that a need exists for more granular prefix control (i.e., smaller prefixes), one or more boundaries may be determined upon which to divide one or more of the prefixes based on the monitored characteristics and/or need. Once the one or more boundaries are determined, the one or more prefixes may be “split” accordingly, e.g., by injecting more specific prefixes into the routing table, thus providing more granular prefix control (i.e., by controlling the split prefixes).

Abstract: An Optimized Edge Routing (OER) technique provides efficiently data routing at the edge of a network or subnetwork. The technique employs a Master node that manages a set of border nodes located at the edge of the network or subnetwork. The Master node may be a stand-alone network management node or may be incorporated into a network node, such as a border node. Unlike prior implementations, the Master node instructs the border nodes to dynamically acquire (“learn”) prefixes of incoming and outgoing data flows and to selectively filter a set of learned address prefixes whose corresponding data flows match a predetermined set of criteria. The criteria may be based on routing metrics other than, or in addition to, conventional cost-based or distance-based metrics. Further, the criteria may include a set of filtering parameters that may be reconfigured, e.g., by the Master node, from time to time.

Abstract: In one embodiment, a first set of one or more control policies at a control server of a network domain may be transmitted to a routing master controller of the network domain, which uses a second set of one or more traffic policies to determine optimal paths for directing traffic through the domain. The routing master controller may then generate a third set of one or more integrated policies based on the first and second sets of policies, e.g., based on the knowledge and policies of both the control server and the routing master controller.

Abstract: In one embodiment, performance parameters may be determined for each of a plurality of network address prefixes in a computer network. Based on the respective performance parameters, the prefixes may be resized through at least one of consolidation of adjacent prefixes and splitting of prefixes, and traffic may then be routed in the computer network based on the resized prefixes.

Abstract: In one embodiment, performance measurements may be obtained at a local domain from a set of one or more responding targets of a particular remote domain. The performance measurements may then be applied to prefix-specific policies of any corresponding destination address prefix within the particular remote domain, such that traffic destined to a particular prefix of the particular remote domain may be routed from the local domain based on the corresponding prefix-specific policies and applied performance measurements.

Abstract: A technique dynamically creates and utilizes a plurality of multi-homed Virtual Private Network (VPN) tunnels from a client node of one spoke network to a client node of another spoke network in a computer network. According to the technique, a VPN client node, e.g., a “spoke,” creates at least one VPN tunnel with an enterprise network, e.g., a “hub.” Once the spoke-to-hub tunnel is established, the spoke may dynamically create a plurality of VPN tunnels with a peer spoke network, e.g., a “peer spoke.” The spoke designates (e.g., for a prefix) one of the tunnels as a primary tunnel and the other tunnels as secondary tunnels, and monitors the quality (e.g., loss, delay, reachability, etc.) of all of the dynamic tunnels, such as, e.g., by an Optimized Edge Routing (OER) process. The spoke may then dynamically re-designate any one of the secondary tunnels as the primary tunnel for a prefix based on the quality of the tunnels to the peer spoke.

Abstract: In one embodiment, probe messages may be transmitted from a local domain to a set of one or more responding probe targets located beyond a plurality of probed first-hop domains of the local domain, the probe messages transmitted via each of a plurality of probed first-hop domains. A reliability of each probed first-hop domain may then be determined based on responsiveness of probe messages transmitted via each respective probed first-hop domain, such that traffic destined beyond the probed first-hop domains may be load balanced across the probed first-hop domains based on the respective reliability, regardless of a traffic-class and a destination of the traffic.

Abstract: In one embodiment, performance measurements may be obtained at a local domain from a set of one or more responding targets of a particular remote domain. The performance measurements may then be applied to prefix-specific policies of any corresponding destination address prefix within the particular remote domain, such that traffic destined to a particular prefix of the particular remote domain may be routed from the local domain based on the corresponding prefix-specific policies and applied performance measurements.

Abstract: In one embodiment, a first set of one or more control policies at a control server of a network domain may be transmitted to a routing master controller of the network domain, which uses a second set of one or more traffic policies to determine optimal paths for directing traffic through the domain. The routing master controller may then generate a third set of one or more integrated policies based on the first and second sets of policies, e.g., based on the knowledge and policies of both the control server and the routing master controller.

Abstract: Techniques are provided for updating best path based on real-time congestion feedback. A method comprises monitoring packets received from an internetworked system, wherein the packets are received on one of a plurality of external interfaces of a networking device; detecting that a received packet includes real-time information that signals a present or pending congestion condition on a path from the external interfaces of the networking device to the internetworked system; notifying a control logic of the real-time information; receiving from the control logic control information defining a change in one or more paths from the external interfaces to the internetworked system; and changing the one or more paths from the external interfaces to the internetworked system. Examining ingress traffic on external interfaces of an internetworked system can cause changes to routes, routing policies and PBRs in routers of the first internetworked system in response to real-time congestion.

Abstract: In one embodiment, a network device (e.g., a master controller) may detect an event on a current path in a computer network from a local network domain to a destination address prefix of a remote domain. In response, the device may dynamically (e.g., intelligently) determine a trace target destination address within the destination address prefix, and may perform (or request performance of) a Traceroute of the current path and a selection of alternate paths in the network from the local network domain to the trace target, where the Traceroute is adapted to obtain per-hop measurements along the respective traced path. The measurements may then be stored, for example, to be used for optimal path selection, fault identification reporting, etc.

Abstract: In one embodiment, a method includes receiving at a network device information on paths to a destination node, determining a number of common nodes or links between at least two of the paths, calculating a path diversity index based on the number of common nodes or links, and selecting a route to the destination node based on the path diversity index. An apparatus for calculating path diversity index for use in route selection is also disclosed.

Abstract: A method for detecting a suspicious packet flow in a packet-switched network comprises the computer-implemented step of receiving a first packet in which the SYN bit but not the ACK or RST bit of the packet's TCP header is set. If a specified first time has elapsed, a packet counter associated with the destination address of the flow is incremented. A determination as to whether the packet counter is greater than a specified threshold values is made. If the packet counter is greater than the threshold value, a notification message is generated. In one embodiment, information identifying a packet flow is aggregated to an aggregation cache based on the destination address of the flow.

Abstract: A technique performs route optimization of services provided by one or more service providers (SPs) for communication links that are combined in a computer network. According to the novel technique, performance characteristics and statistics (“link data”) for two or more combined links are merged to form link data for a single “virtual combined link.” Route optimization techniques (e.g., priority-based route optimization) may then be applied to one or more virtual combined links based on the combined link data, in addition to link data for zero or more conventional single links (physical or virtual). Illustratively, a route optimization technique may be used in accordance with the present invention to efficiently and accurately minimize costs associated with services provided by a plurality of SPs, wherein each SP charges for utilization of its service based on a tiered pricing structure.

Abstract: Techniques are provided for updating best path based on real-time congestion feedback. A method comprises monitoring packets received from an internetworked system, wherein the packets are received on one of a plurality of external interfaces of a networking device; detecting that a received packet includes real-time information that signals a present or pending congestion condition on a path from the external interfaces of the networking device to the internetworked system; notifying a control logic of the real-time information; receiving from the control logic control information defining a change in one or more paths from the external interfaces to the internetworked system; and changing the one or more paths from the external interfaces to the internetworked system. Examining ingress traffic on external interfaces of an internetworked system can cause changes to routes, routing policies and PBRs in routers of the first internetworked system in response to real-time congestion.

Abstract: A method for detecting a suspicious packet flow in a packet-switched network comprises the computer-implemented step of receiving a first packet in which the SYN bit but not the ACK or RST bit of the packet's TCP header is set. If a specified first time has elapsed, a packet counter associated with the destination address of the flow is incremented. A determination as to whether the packet counter is greater than a specified threshold values is made. If the packet counter is greater than the threshold value, a notification message is generated. In one embodiment, information identifying a packet flow is aggregated to an aggregation cache based on the destination address of the flow.