Cisco Technology, Inc.

Robert Marcin Raszuk

Jean-Philippe Vasseur

Clarence Filsfils

Jim Guichard

David Delano Ward

Stefano Benedetto Previdi

Keyur Patel

Christian Cassar

In one embodiment, a network device determines identities of each peer device in a second routing domain attached to edge devices in a first routing domain. The network device associates each address prefix reachable in the second routing domain with an identity of each peer device in the second routing domain that advertised the address prefix and with an identity of one or more edge devices in the first routing domain to which that peer device is attached. The network device determines an address prefix is associated with a same identity of a peer device in the second routing domain but with different edge devices in the first routing domain. The network device assigns the different edge devices in the first routing domain associated with the determined address prefix to a shared risk node group (SRNG).

Dynamic shared risk node group (SRNG) membership discovery

In one embodiment, an angular distance calculation technique may enable an intermediate node, such as a route reflector, to render customized best path selection decisions to destination address prefixes and advertise those decisions to its intermediate peering nodes, i.e., route reflector clients, in an autonomous system of a computer network. That is, the route reflector may execute a routing protocol, such as BGP, to select and advertise a next hop (e.g., an address prefix of an egress point of the autonomous system) associated with a customized best path to a route reflector client based on an angular distance between the route reflector client and the egress point. In this manner, the route reflector client may route data traffic destined to a destination address prefix utilizing the egress point associated with the customized best path to that prefix.

Angular distance calculation for BGP best path selection

In one embodiment, a loss of communication is detected between a first edge device of a computer network and a neighboring routing domain. A data packet is received at the first edge device, where the received data packet contains a destination address that is reachable via the neighboring routing domain. A determination is made whether a service label is located in a Multi-Protocol Label Switching (MPLS) label stack included in the received data packet. A service label in the MPLS label stack indicates that the received data packet was previously rerouted in accordance with fast reroute (FRR) operations. In response to a determination that the received data packet does not include a service label in the MPLS label stack, the received data packet is rerouted to a second edge device of the computer network for forwarding to the neighboring routing domain.

Loop prevention technique for MPLS using service labels

A technique institutes backup Border Gateway Protocol (BGP) paths for non-multipath BGP Fast Convergence in a computer network. According to the novel technique, a protecting node, e.g., a border router, calculates a best BGP path to a destination prefix, and also calculates one or more backup BGP paths (i.e., non-multipath paths) to the destination prefix. The best path is selected as a utilized path for the destination prefix, e.g., the path over which to send traffic destined for the prefix from the border router. Upon determining that the best BGP path has failed, one of the backup BGP paths is selected as the utilized path for the destination prefix.

Backup BGP paths for non-multipath BGP fast convergence

Techniques for automatically creating a BGP mesh, and more specifically, an iBGP mesh, are provided. Peering information is flooded to routers within a network AS. The peering information is analyzed and a list of neighbors is created. BGP sessions are then performed with the neighbors to establish the BGP mesh.

Growing community of inventors

San Jose, CA, United States of America

Robert Marcin Raszuk