MPLS-Enabled Applications: Emerging Developments and New Technologies

11.6. CONTROL PLANE OPERATION

Let us see how the control plane for Layer 2 transport operates. We will examine the LDP-based scheme [MRT-TRS] and the BGP-based scheme [KOM-BGP]. Both approaches have the following characteristics in common:

1. A means for a PE, when forwarding traffic from a local CE via a remote PE to a remote CE, to know the value of the VPN label (inner label) that the remote PE expects.

2. A means for signaling characteristics of the pseudowire, such as media type and MTU. This provides a means to detect whether each end of a pseudowire are configured in a consistent manner or not.

3. An assumption that the pseudowire formed is bidirectional. Hence, if there is a problem with transport in one direction, forwarding is not allowed to occur in the opposite direction.

4. A means for a PE to indicate to remote PE(s) that there is a problem with connectivity, e.g. if the link to a CE goes down.

The two schemes differ significantly in the way in which a PE knows which remote PE(s) it needs to build pseudowires to. In the original LDP-based scheme, this information had to be manually configured on the PEs. The BGP scheme, in contrast, has in-built autodiscovery properties, so this manual configuration is not required. The original LDP scheme was later modified in order to also avoid this manual configuration, by using information discovered by some means external to LDP. One option for the autodiscovery aspect is to use BGP. In the following sections, we will discuss the three cases ...