Routing convergence method and device in multi-protocol label switching (MPLS) and border gateway protocol (BGP) networking

Abstract

The invention provides a routing convergence method and device in multi-protocol label switching (MPLS) and border gateway protocol (BGP) networking. According to the method, a provider edge (PE) enables a BGP route and related functions of a label-switch path (LSP) to be achieved and performs the steps of performing effectiveness detection on the LSP of next hop iteration of the BGP route on the PE periodically; when detecting that the LSP is invalid, enabling the BGP route to be marked as an invalid route, cancelling the BGP route when other BGP routes with identical prefixes as the BGP route do not exist on the PE; cancelling the BGP route when other BGP routes exist on the PE, and optimally selecting a BGP route meeting the following condition from other BGP routes to replace the BGP route marked as the invalid route so as to perform message forwarding: the LSP of the next hop iteration is different from the LSP of the next hop iteration of the BGP route marked as the invalid route. By adopting the routing convergence method and device, rapid convergence of the BGP route can be accelerated, and message forwarding performance is improved.

Description

technical field [0001] The present application relates to network communication technology, in particular to a route convergence method and device in multiprotocol label switching (MPLS: Multiprotocol Label Switching) and border gateway protocol (BGP: Border Gateway Protocol) networking. Background technique [0002] The networking using MPLS and BGP has good scalability and can easily support MPLS Quality of Service (QoS) and MPLS Traffic Engineering (TE: Traffic Engineering), and has been applied more and more. [0003] For ease of description, the networking using MPLS and BGP is referred to as MPLS and BGP networking, or BGP / MPLS networking for short. [0004] In a BGP / MPLS network, BGP acts as the upper-layer control plane to control the advertisement and withdrawal of BGP routes. As the lower layer data forwarding plane, MPLS uses the BGP routes advertised by BGP to forward packets. [0005] However, because BGP routes are forwarded hop by hop in the next hop mode, t...

AI Technical Summary

Problems solved by technology

[0005] However, because BGP routes are forwarded hop by hop in the next hop mode, the advertised BGP route will be marked as an invalid route only when the next hop of the advertised BGP route is detected to be invalid, so the following problems will occur : When MPLS Label-Switch Path (LSP: Label-Switch Path) forwarding is invalid in BGP / MPLS networking, for example, the link is unreachable or the MPLS label If it is not invalid, the BGP route will still be used for packet forwarding, which will cause packet forwarding failure, or even packet loss, and cannot achieve fast convergence of BGP routes

[0006] for example, figure 1 In the BGP / MPLS network shown, it is assumed that the next hop of the BGP route on the service provider network edge device (PE: Provider Edge) 1 is the IP address of PE2, and the BGP route is applied to the LSP between PE1 and PE2. The LSP forwarding between PE1 and PE2 is invalid. For example, the MPLS forwarding link between PE1 and P1 is faulty. At this time, PE1 has a published BGP route for the LSP between PE1 and PE2. If it is not marked as an invalid route, PE1 will not revoke the BGP route, and it will still use the BGP route to forward packets. Since the LSP forwarding between PE1 and PE2 is invalid, PE1 is using this BGP route. When the route forwards the message, the message will not be successfully forwarded, that is, a BGP route that cannot be forwarded has been applied to the forwarding of the message for a long time, and the fast convergence of the BGP route cannot be realized, which affects the forwarding of the message

Images