When a link flaps, it could take a long time for LDP to reexchange labels, off course a network can use the FIB in the meanwhile, but this could present several problems with applications that leverage the use MPLS, line MPLS VPN to say at least one. With MPLS LDP Session Protection, we can provide faster LDP convergence when a link recovers from an outage, and this is done maintaining the LDP session for a period of time.
Now when a link fails, we know that in frame mode mpls LDP would store all the labels in the LIB, even if they are not used, this is because the IGP could decide to use another path, but the real problem here, comes into play when the link is recovered, when the IGP determines that the link is available could probably change the next hop is the path to reach the network is better. The problem here is the POP action used in the LFIB table of the router while the LDP tries to establish again the session, adding to our networks, more time to converge, since the LIB might not contain the label from the new next hop, by the time the IGP had converged.
We have 2 ways to solve the convergence issues that we are faced on flapping links, the first solution is to use MPLS LDP Session Protection and the second one is to use MPLS TE make before.
MPLS Fundamentals Book by Luc de Ghen CCIE 1897 states:
A common problem in networks is flapping links. The flapping of links can have several causes, but it is not the goal of this book to look deeper into this. Flapping links do have an important impact on the convergence of the network. Because the IGP adjacency and the LDP session are running across the link, they go down when the link goes down. This is unfortunate, especially because the link is usually not down for long. The impact is pretty severe though, because the routing protocol and LDP can take time to rebuild the neighborship. LDP has to rebuild the LDP session and must exchange the label bindings again. To avoid having to rebuild the LDP session altogether, you can protect it. When the LDP session between two directly connected LSRs is protected, a targeted LDP session is built between the two LSRs. When the directly connected link does go down between the two LSRs, the targeted LDP session is kept up as long as an alternative path exists between the two LSRs. The LDP link adjacency is removed when the link goes down, but the targeted adjacency keeps the LDP session up. When the link comes back up, the LSR does not need to re-establish the LDP session; therefore, the convergence is better. The global command to enable LDP Session Protection is this:
mpls ldp session protection [vrf vpn-name] [for acl] [duration seconds]