The Research On The Key Technologies Of ISP Backbone Network Bases On MPLS

At present the network has been a stage for daily work, study and entertainments. As ISPs, they hope network provide them more benefit. In order to achieve this object, they need to hold more service with less investment. With the reference point of the above, we here to study from two points of view to cater for this purpose. Firstly, we make the original network become several backbone networks through dividing logically in order to reduce the input from the ISPs. Secondly, it fully uses the existing bandwidth through Traffic Engineering (TE) to avoid the continually updating for backbone networks due to the increase of the data traffic by ISPs.On the study of the logical division of backbone networks, the Virtual Private Network (VPN) based on Multi-Protocol Label Switch (MPLS) is that MPLS technology transmits packets through the backbone networks and Border Gateway Protocol (BGP) distribute route through the backbone networks. Each VPN based on MPLS actually is a dedicated IP network, which can compile dedicated IP address for every ISP's PE and connect the client node rapidly. BGP is used to distribute route to every node in the VPN. This research mainly comes down to three aspects which are setting up private route table isolated with backbone network entirely, inducting BGP in existence and expanding the protocol according to the demand of VPN based on its inherent attribute, and realizing the requirement of using the same address segment among different VPNs .Finally, the simulation is established based on the practice. And we erect corresponding test environment in order to test the main function and the compatible quality with existing VPN.The second aspect of this paper is setting up Traffic Engineering (TE) in backbone networks. TE is used to arrange the traffic in the network in reason to avoid using resource imbalance. Because the network resource is costliness and Internet business competition is drastic, TE has been a necessary function of large-scale autonomous system.The existing Internal Gateway Protocol (IGP) has no capability to provide TE. IGP computes through topology. It only bases on a simple affixation measurement, such as hop number and some administer value and doesn't issue information such as bandwidth usability and service character. This presumes that when the IGP compute its transfer table, it didn't consider the traffic load in the network. It maybe induce that the services are not assigned averagely in connection, which results in part resource not using sufficiently.Label Switch Router (LSP) is build according to existing Label Distribute Protocol (LDP) and Resource Reservation Protocol (RSVP). At present this LSP uses handworked configure manner. That is, when finding traffic is congestion in some route, it can make the traffic transfer to another route artificially. Consequently it takes full advantage of the resource and improves the utilization efficiency in the hole networks.