Dedicated Network Path
When creating LSPs, a crucial consideration is the path that will be used. In the case of LSP123 in Figure 1, the path comprises the routers R2-R3-R4-R5. After we have the path (and some other items not important to us at the moment), we can then ask the router R2 to signal or create the LSP. The signaling process consists of running a protocol procedure on R2 that asks R3 whether it can supply part of an LSP. R3 then asks R4 whether it can supply the next part of the LSP. Finally, R4 asks R5 whether it can supply the last part of the LSP
If the signaling is successful, the LSP123 is created, ready for traffic forwarding. At this point, traffic landing at R2 can be pushed either in the direction of R6 or into LSP123. If traffic is pushed in the direction of R6, a normal IP lookup occurs at R6.
If traffic enters LSP123, however, that traffic is switched along the path of LSP123 until it reaches R5—where a normal IP lookup occurs. In other words, at R5, the destination IP address of the traffic is consulted to decide the next hop. (For example, should it go to the PBX or on to R7 in Figure 1?).
So what’s the point of the LSP? Well, it allows for constraining the traffic to pass in a given direction along a specific path. Remember traffic engineering? This is what it’s all about: moving traffic into regions of the network in which resources exist to handle it without congestion or packet loss.
There’s a huge world-wide effort under way to push legacy traffic (such as telephony and other technologies) into IP networks. Traffic engineering is one of the major challenges that have to be overcome during what has become known as network convergence. (For more on this and related topics, please see "Additional Reading" at the end of the chapter.)