摘要:
A method for improving LDP (Label Distribution Protocol) convergence time in an MPLS (Multi-Protocol Label Switching) network is described. An hierarchical LSP is established to transport packets belonging to a FEC attached to an egress LSR. The hierarchical LSP includes an egress LSR LSP that is common for each of the FECs attached to the egress LSR and forms a path from the ingress LSR through one or more intermediate LSRs to the egress LSR. The egress LSR LSP is used when label switching packets destined for the FECs attached to the egress LSR. The hierarchical LSP also includes a unique FEC LSP for each FEC that is used by the egress LSR to identify and forward packets to that FEC. Responsive to a topology change that changes a next-hop of the ingress LSR to reach the egress LSR, the ingress LSR modifies an entry in a forwarding structure to change the next-hop for the egress LSR LSP and does not modify substantially any forwarding structure entities for the FEC LSPs. LDP convergence time is reduced through a reduction of forwarding structure modifications after the topology change.
摘要:
Signaling a Label Switched Path (LSP) tunneling model is described. In one embodiment, a network element that is acting as an egress network element in a Multiprotocol Label Switching (MPLS) network signals multiple LSPs for multiple disparate applications provided in the MPLS network that each require a different type of tunneling model. The network element transmits a first label mapping message for one of the LSPs that includes an indication of a first tunneling model type applicable for that LSP. That LSP is to be used to transport traffic in the MPLS network for a first one of the disparate applications that requires the indicated first tunneling model type. The network element further transits a second label mapping message for a second one of the LSPs that includes an indication of a second tunneling model type applicable for that LSP. That LSP is to be used to transport traffic in the MPLS network for a second one of the disparate applications that requires the indicated second tunneling model type. The first tunneling model type and the second tunneling model type are different. In one embodiment, the first tunneling model type is a uniform tunneling model and the second tunneling model type is a pipe tunneling model (regular pipe or short-pipe).
摘要:
An edge router runs a Multipath Transmission Control Protocol (MPTCP) proxy to allow for a host that implements TCP (Transmission Control Protocol) to operate normally yet reap the benefits of an MPTCP connection. An upgrade of a TCPIP stack on the host is not necessary. The edge router demultiplexes packets received from the host over a TCP connection to an MPTCP connection and multiplexes packets sent to the host over an MPTCP connection to a TCP connection. As a result, higher throughput of packet communication can be realized, for example, for improved video support.
摘要:
Improving efficiency of encapsulation for packets of a first set of one or more protocols on a packet-pseudowire over a tunnel in a Packet Switched Network (PSN) is described. A first provider edge (PE) network element is coupled with a customer edge (CE) network element over an attachment circuit and is coupled with a second PE network element over the packet-pseudowire. The first provider edge network element receives a frame from the CE network element over the attachment circuit. Responsive to the first PE network element determining that the frame encapsulates a packet of the first set of protocols, the first PE network element encapsulates the packet into a protocol data unit for transmission over the packet-pseudowire without including substantially all of the plurality of fields of the data link layer header. The first PE network element transmits the protocol data unit over the packet-pseudowire over the PSN tunnel to the second PE network element. Efficiency is improved through a reduction of overhead on the packet-pseudowire by removing substantially all the fields of the data link layer header of the first frame.
摘要:
A method and apparatus that operates two bridging protocols in a hybrid bridging node (204D) is described. The operation of the two bridging protocols in the hybrid node allows for an incremental transition of a provider bridging network from operating a legacy bridging protocol that shares MAC addresses to a bridging network that operates VPLS and or PBB bridging protocols. The hybrid bridging node selectively broadcasts address resolution packets and unicast packets with unknown MAC addresses from the nodes operating VPLS and or PBB to nodes operating a legacy bridging protocol.
摘要:
MPLS (Multiprotocol Label Switching) fast re-routing using LDP (Label Distribution Protocol) is described. A first network element in a MPLS network receives a first label advertised from a second network element in the MPLS network. The first network element computes a shortest path tree (SPT) to reach a destination network element under a potential failure condition. The second network element is a nexthop of the first network element in the computed SPT and is not upstream from the potential failure condition. The first network element advertises a second label to one or more third network elements for a backup LDP LSP (Label Switched Path) that will serve as a backup when the potential failure condition is realized. The third network element(s) are upstream neighbors on the computed SPT with respect to the first network element. The first network element installs a swap action from the second label to the first label.
摘要:
Computing a constraint-based label switched path (LSP) that spans multiple areas is described. In one embodiment, a router in a first one of the multiple areas computes a path segment that meets a set of one or more constraints to at least one border router of the first area that lies in a path necessary to reach the destination. The router transmits a path computation request message to a path computation element (PCE) in a second one of the areas, which includes a set of one or more attributes for each computed path segment that are used by the PCE to compute one or more path segments towards the destination of the constraint-based LSP. The router receives a path computation reply message from the PCE that specifies a set of one or more computed path segments that meet the set of constraints and that were computed by one or more PCEs downstream from the router. The router stitches at least one of the set of computed path segments that are specified in the path computation reply message with the path segment computed by the router to form at least part of the constraint-based LSP.