Multi-Protocol label switching (MPLS)
In traditional 3-tier forwarding, each router extracts forwarding information from a 3-tier file header as packets are transmitted over the network. This file header analysis is repeated for each router that the packet passes through.
In MPLS networks, packets are forwarded according to the label. Each IP network that can be reached through an interface is assigned a unique label. A mapping is to be established between the network label and the outbound label. This mapping is stored in the tag forwarding information base (LFIB) table. Each node check the network label, do a table check, the network label Exchange into a network label, and then forward the packet to the network interface. The application of this form allows the MPLS network to establish a label Switching channel (LSP) for the entire network.
Figure 1 shows the details of the MPLS file header. It is located between the 3-layer (IP) file header and the 2-tier file header. The test bit (exp-bits) and TTL (Live time) fields of the Mpls file headers can be copied from the IP file header. The S-byte indicates whether there is more than one MPLS tag in the packet.
Figure 1:mpls Label file header
Routers in an MPLS network use a protocol to assign labels to IP networks and Exchange tag information with other routers. The most common protocol currently used is the label Allocation Protocol (the LDP port number is 646). This protocol is based on the TCP protocol and runs on the MPLS label Switching router (LSR).
The concept of LSR is often used to illustrate MPLS devices. LSR runs a routing protocol and therefore has 3 layers of intelligence. However, once the LSP is established, the MPLS device performs only the check label work when forwarding traffic. Therefore, they combine the intelligence of routing with the speed of exchange. Figure 2 describes the basic operations performed by an MPLS network.
Figure 2: Overview of label Exchange