- What is Label Switching?
- Why Use Label Switching?
- The ZIP Code Analogy
- Why A Label IS Not an Address
- How Label Switching is Implemented and How it Came About
- Clarification of Terms
- The Need for a QOS-based Internet
- Label Switching's Legacy: X.25 and Virtual Circuits
- MPLS: Status and Concepts
- Examples of Label and QOS Relationships
- Determination of the Physical Path Through the Network: The Label Switched Path (LSP)
Why A Label IS Not an Address
A label is not an address. It has no inherent topological significance. Moreover, until the label is correlated with an address, it has no routing significance. Therefore, a requirement still exists for conventional IP address advertising, as shown in Figure 13. Part of the job of a label switching network is to correlate the addresses and routes with labels.
Figure 13 Address advertisement.
The routes are discovered by the IP routing protocols and are based on IP addresses. In this example, the label switching routers are advertising address 18.104.22.168. In most situations, an address prefix is advertised (a prefix is the network and subnetwork part of the 32-bit address), but that need not concern us for this general example.3 This advertisement reaches the router on the left side of Figure 13. The router stores the routing information in its routing table. Thereafter, when the router receives a packet destined for address 22.214.171.124, it consults its routing table to find out how to reach this address.
In a label switching network, an important job is to choose a label value to place onto the packet header for use in the network and to inform the other label switching routers about the association of the label value to the address. How this operation is accomplished is shown in a general way in Figure 14. Router (node) A informs router (node) B that address 126.96.36.199 is to be associated with label 88888. This association is called a bind.
Figure 14 Label/address advertisement.
When router B receives this label/address advertisement, it consults its routing table and looks up the next node that is to receive traffic destined for 188.8.131.52. As we learned in Figure 13, that next node is router C. Therefore, router B builds an entry in another table (called by various names: label switching table, label mapping table, cross-connect table, as examples) specifying that an incoming label from node A with a value of 88888 is to be routed onto the outgoing link to node C. This process continues until the packet reaches the final destination.
You may have noticed that I did not show the operations between router B and router C in Figure 13. The reason for this exclusion is that there are some additional operations between the LSRs B and C that are explained later.
The operation in Figure 14 has the label assigned by LSR A after it has discovered the path to the address. Another approach is for the binding to occur at the same time the address is advertised. Consequently, in Figure 15, the process of binding begins at node C. The label switching networks can support both approaches, the pros and cons of which are explained in later chapters.
Figure 15 Advertising and binding at the same time.