Mpls/bgp Vpn Multicast Research And Improvement Of The Relevant Agreements

With the development of Multi-Protocol Label Switching(MPLS) and the advent of Multi-Protocol Extensions for BGP4(MP-BGP4), the technology of MPLS/BGP VPN keeps developing and has become one of the mainstream L3VPN technologies; at the same time, the demands of network customs for P2MP and MP2MP applications or services such as distributed applications, multimedia applications also dramatically increase. When P2MP or MP2MP applications need to be carried out in MPLS/BGP VPN, the demands for multicast in MPLS/BGP VPN arises. Although the tow kinds of technologies--"MPLS/BGP VPN" and "Multicast" are already relatively formed, there are still many problems in the implement of multicast in MPLS/BGP VPN. The combination of different two technologies needs to be researched, therefore bringing about this dissertation.This dissertation begins in the study of mechanisms and implements of MPLS/BGP VPN and RSVP-TE P2MP, studying and analyzing the problems needs to be solved related to the implement of multicast in MPLS/BGP VPN, then doing the following:1. Analyzed the architecture of the system of multicast in MPLS/BGP VPN, then designed an overall architecture and interactions among different modules, further implemented the critical routing modules in forwarding plane.2. Researched the mechanisms of the signaling module—RSVP-TE P2MP, analyzed the limits of backup and grafting in multicast in MPLS/BGP VPN, and respectively proposed two more compatible schemes, which are based on the standardized RSVP-TE P2MP. By defining a new RSVP object, the proposed backup scheme can be realized on provides’edge routers, including node protection, link protection and LSP protection. And by defining a new kind of RSVP message, an improved grafting method was designed.3. In order to verify the proposed RSVP-TE P2MP backup and grafting schemes, first designed and implemented the function of establishing LSPs by RSVP-TE P2MP messages—Path and Resv, referring to RFC4875, by UNIX network programming methods. By initiating and sending a Path message from PE router, then checking the contents of Path and Resv messages among related PE and P routers, as well as the Label Forwarding Information Base, confirmed the correctness and reliability of the designed processes of LSPs establishment.4. On those basis, implemented the proposed backup and grafting schemes, respectively by initiating and sending Path messages with backup function and Join messages on an established P2MP LSP,, then checking the contents of Path, Resv and Join messages among related PE and P routers, as well as the Label Forwarding Information Base, confirmed the merits and reliability of the designed processes of P2MP LSP backup and sub-LSP grafting.