Implementation

Figure 1 illustrates the MPLS TE framework.

Figure 1 MPLS TE framework

MPLS TE is implemented based on four functions:

IGP-based information advertisement for TE information collection
Path calculation using the collected information
Path setup through signaling packet exchange between upstream and downstream nodes
Traffic forwarding over an established MPLS TE tunnel

Table 1 describes the four functions.

**Table 1** Functions for MPLS TE implementation
No.	Function	Description
1	Information advertisement	Collects network load information in addition to routing information. MPLS TE extends an IGP to advertise TE information, including the maximum link bandwidth, maximum reservable bandwidth, reserved bandwidth, and link colors. Every node collects TE information about all links in a local area and generates a traffic engineering database (TEDB).
2	Path calculation	Uses the Constrained Shortest Path First (CSPF) algorithm and data in the TEDB to calculate a path that satisfies specific constraints. CSPF evolves from the Shortest Path First (SPF) algorithm. It excludes nodes and links that do not satisfy specific constraints and uses the SPF algorithm to calculate a path.
3	Path setup	Sets up a static or dynamic CR-LSP. Static CR-LSP Forwarding and resource information is manually configured for a CR-LSP without the need of a signaling protocol or path calculation. Setting up a static CR-LSP consumes few resources because no MPLS control packets are exchanged between the two ends of the CR-LSP. Static CR-LSPs cannot be adjusted dynamically; therefore, static CR-LSP setup applies only to small networks with simple topologies. Dynamic CR-LSP Nodes on a network use the Resource Reservation Protocol (RSVP) TE signaling protocol to set up CR-LSP tunnels. RSVP-TE messages carry constraints for a CR-LSP, such as the bandwidth, explicit path, and affinity attribute. There is no need to manually configure each hop along a dynamic CR-LSP. Dynamic CR-LSP setup applies to large-scale networks. RSVP authentication can be used to enhance security and reliability of CR-LSPs.
4	Traffic forwarding	Directs traffic to an MPLS TE tunnel and forwards traffic over the MPLS TE tunnel. The first three functions set up an MPLS TE tunnel, and the traffic forwarding function directs traffic arriving at a node to the MPLS-TE tunnel.

A static CR-LSP is manually established and does not require information advertisement or path calculation.
A dynamic CR-LSP is set up using a signaling protocol and involves all the four functions listed in the table.

To deploy MPLS TE on a network, you must configure link and tunnel attributes. Then MPLS TE sets up tunnels automatically. After a tunnel is set up, traffic is directed to the tunnel for forwarding.