Abstract: A method is performed at a source node in a network of nodes configured with a link state protocol, and in which at least some of the nodes are enabled for multiprotocol label switching (MPLS). The node discovers and stores a link state topology representing the nodes of the network, links between the nodes, path-costs for the links, and whether each link is enabled or not enabled for MPLS. The node determines one or more shortest paths from the source node to a destination node among the nodes based on traversing the link state topology and, while the node traverses the link state topology, detects whether each shortest path supports or does not support MPLS end-to-end dataplane continuity. The node programs an IP dataplane with each shortest path, and programs an MPLS dataplane with ones of the one or more shortest paths that support the end-to-end MPLS continuity.

Abstract: Content distribution system cache management may be provided. First, a sync packet may be received by a cache server from a first server. The sync packet may include a list indicating a cache server where a chunk is to be stored and the address for the chunk. Next, an address for the chunk may be obtained by the cache server by parsing the sync packet. The cache server may then determine that the chunk is not stored on the cache server by using the address for the chunk. Next, in response to determining that the chunk is not stored on the cache server, a connection may be opened between the first server and the cache server. The cache server may then receive the chunk over the connection and cache the chunk on the cache server.

Abstract: A method is performed at a source node in a network of nodes configured with a link state protocol, and in which at least some of the nodes are enabled for multiprotocol label switching (MPLS). The node discovers and stores a link state topology representing the nodes of the network, links between the nodes, path-costs for the links, and whether each link is enabled or not enabled for MPLS. The node determines one or more shortest paths from the source node to a destination node among the nodes based on traversing the link state topology and, while the node traverses the link state topology, detects whether each shortest path supports or does not support MPLS end-to-end dataplane continuity. The node programs an IP dataplane with each shortest path, and programs an MPLS dataplane with ones of the one or more shortest paths that support the end-to-end MPLS continuity.

Abstract: In one embodiment, a device in a network receives a packet that includes a forwarding label for a service in a service chain. The packet encapsulates a service chain header for the service chain. The device swaps the forwarding label for the service in the packet for a reserved label that identifies the packet as encapsulating the service chain header. The device forwards the packet with the reserved label to the service.

Abstract: One embodiment performs segment routing network processing of packets including segment routing packets having a multiple segment routing header packet structure that provides processing and/or memory efficiencies. In one embodiment, a particular packet is received by a particular router in a network. In response to the particular router data plane ascertaining based on the particular packet a particular segment routing (SR) policy identifying one or more ordered SR identifiers, the particular router adding one or more SR headers to the particular packet resulting in the particular packet including multiple ordered SR headers instead of the particular packet having a packet structure with a single SR header, with each of the one or more SR headers including at least one segment identifier of said one or more ordered SR identifiers. The packet with the multiple ordered SR headers is sent from the particular router.

Abstract: In one embodiment, packets are forwarded in a network according to a Segment Routing-based (SR-based) multicast distribution tree identified by a Tree Segment Identifier (Tree-SID). This packet forwarding includes packet replication to cause multiple copies of a same packet to be forwarded to different nodes of the SR-based multicast distribution tree. The Tree-SID is typically a same global value used within the network to identify the SR-based multicast distribution tree. As each packet is being routed through the network according to the SR-based multicast distribution tree, the packet includes the Tree-SID in a Segment List of the packet, with the Segment List being an ordered list of SID's identifying information for forwarding the packet in the network. The Tree-SID provides a lookup key for efficient forwarding of packets by packet switching devices making forwarding decisions, which may including forwarding multiple copies of the packet.

Abstract: In one embodiment, a device in a network receives a packet that includes a forwarding label for a service in a service chain. The packet encapsulates a service chain header for the service chain. The device swaps the forwarding label for the service in the packet for a reserved label that identifies the packet as encapsulating the service chain header. The device forwards the packet with the reserved label to the service.

Abstract: In one embodiment, a device in a network receives a packet that includes a forwarding label for a service in a service chain. The packet encapsulates a service chain header for the service chain. The device swaps the forwarding label for the service in the packet for a reserved label that identifies the packet as encapsulating the service chain header. The device forwards the packet with the reserved label to the service.

Abstract: Content distribution system cache management may be provided. First, a sync packet may be received by a cache server from a first server. The sync packet may include a list indicating a cache server where a chunk is to be stored and the address for the chunk. Next, an address for the chunk may be obtained by the cache server by parsing the sync packet. The cache server may then determine that the chunk is not stored on the cache server by using the address for the chunk. Next, in response to determining that the chunk is not stored on the cache server, a connection may be opened between the first server and the cache server. The cache server may then receive the chunk over the connection and cache the chunk on the cache server.

Abstract: In one embodiment, a device in a network receives a packet that includes a forwarding label for a service in a service chain. The packet encapsulates a service chain header for the service chain. The device swaps the forwarding label for the service in the packet for a reserved label that identifies the packet as encapsulating the service chain header. The device forwards the packet with the reserved label to the service.

Abstract: In one embodiment, a method comprises generating, by a first provider edge router associated with a first segment identifier, a primary label for reaching a destination, and repair information for reaching the destination if a second provider edge router is unavailable to reach the destination; allocating, by the first provider edge router, a first protected next-hop address associated with the first segment identifier for protected reachability to at least the destination; and sending via a core network, by the first provider edge router, an advertisement specifying the label and the repair information, enabling an ingress provider edge router to insert, into a data packet destined for the destination, the labels from the first provider edge router and the second provider edge router based on the repair information, for fast rerouting to the destination via one of the first or second provider edge router if the other is unavailable.

Abstract: In one embodiment, a method comprises generating, by a first provider edge router associated with a first segment identifier, a primary label for reaching a destination, and repair information for reaching the destination if a second provider edge router is unavailable to reach the destination; allocating, by the first provider edge router, a first protected next-hop address associated with the first segment identifier for protected reachability to at least the destination; and sending via a core network, by the first provider edge router, an advertisement specifying the label and the repair information, enabling an ingress provider edge router to insert, into a data packet destined for the destination, the labels from the first provider edge router and the second provider edge router based on the repair information, for fast rerouting to the destination via one of the first or second provider edge router if the other is unavailable.

Abstract: In an example embodiment, a method is provided that receives a broadcast of available bandwidth from a first routing device. A congestion of traffic is detected along a downstream path to a second routing device. This second routing device is an immediate downstream neighbor. As such, an alternate path is established to the second routing device by way of the first routing device based on the available bandwidth in the network and a portion of the traffic is transmitted along the alternate path.

Abstract: According to one embodiment, an apparatus that makes improved routing decisions is provided that includes a receiver, a retriever, a proximity engine, and a transmitter. The receiver may be configured to receive a request from a source. The retriever may be configured to retrieve a plurality of providers capable of servicing the request. The proximity engine may be configured to rank the plurality of providers based on weights indicative of a network distance. The weights may correspond to a first Border Gateway Protocol (BGP) community attribute associated with the source and at least a second BGP community attribute associated with at least one of the providers in the plurality of providers. The transmitter may be configured to transmit the request to the provider in the plurality of providers with the highest ranking.

Abstract: In one embodiment, a packet switching device is configured to convert an Internet Protocol Version 6 (IPv6) destination address, of a received particular IPv6 packet, to a second, shorter destination address. This second destination address is then used to determine forwarding information for the received IPv6 packet, which is forwarded accordingly. In one embodiment, this second address is a 32-bit address, and in particular, an Internet Protocol Version 4 (IPv4) address. Thus, one embodiment can use the IPv4 forwarding infrastructure of a packet switching device for determining how to forward IPv6 packets. In a network according to one embodiment, packets are encapsulated in an IPv6 packet using an IPv6 destination address (that can be converted to an IPv4 address) of an egress edge packet switching device. Thus, core packet switching devices can forward IPv6 packets using IPv4 lookup operations.

Abstract: In one embodiment, a packet switching device is configured to convert an Internet Protocol Version 6 (IPv6) destination address, of a received particular IPv6 packet, to a second, shorter destination address. This second destination address is then used to determine forwarding information for the received IPv6 packet, which is forwarded accordingly. In one embodiment, this second address is a 32-bit address, and in particular, an Internet Protocol Version 4 (IPv4) address. Thus, one embodiment can use the IPv4 forwarding infrastructure of a packet switching device for determining how to forward IPv6 packets. In a network according to one embodiment, packets are encapsulated in an IPv6 packet using an IPv6 destination address (that can be converted to an IPv4 address) of an egress edge packet switching device. Thus, core packet switching devices can forward IPv6 packets using IPv4 lookup operations.

Abstract: In an example embodiment, a method is provided that receives a broadcast of available bandwidth from a first routing device. A congestion of traffic is detected along a downstream path to a second routing device. This second routing device is an immediate downstream neighbor. As such, an alternate path is established to the second routing device by way of the first routing device based on the available bandwidth in the network and a portion of the traffic is transmitted along the alternate path.

Abstract: In an example embodiment, a method is provided that receives a broadcast of available bandwidth from a first routing device. A congestion of traffic is detected along a downstream path to a second routing device. This second routing device is an immediate downstream neighbor. As such, an alternate path is established to the second routing device by way of the first routing device based on the available bandwidth in the network and a portion of the traffic is transmitted along the alternate path.

Abstract: According to one embodiment, an apparatus that makes improved routing decisions is provided that includes a receiver, a retriever, a proximity engine, and a transmitter. The receiver may be configured to receive a request from a source. The retriever may be configured to retrieve a plurality of providers capable of servicing the request. The proximity engine may be configured to rank the plurality of providers based on weights indicative of a network distance. The weights may correspond to a first Border Gateway Protocol (BGP) community attribute associated with the source and at least a second BGP community attribute associated with at least one of the providers in the plurality of providers. The transmitter may be configured to transmit the request to the provider in the plurality of providers with the highest ranking.

Abstract: In an example embodiment, a multicast tree is accessed. The multicast tree defines one or more destination label switch routers and paths from a source label switch router to the destination label switch routers. Multicast addresses are then transmitted to the destination label switch routers. In an example embodiment, upon receipt of the multicast addresses, a request to update the multicast tree is transmitted. The request includes the identifier of the label switch router that originated the request.