Static versus Dynamic MPLS Label-Switched Paths (LSPs)
Multi-Protocol Label Switching (MPLS) converts your routed network to something closer to a switched network and offers transport efficiencies that simply aren’t available in a traditional IP-routed network. Instead of forwarding packets on a hop-by-hop basis, paths are established for particular source-destination pairs. The predetermined paths that make MPLS work are called label-switched paths (LSPs).
You can create MPLS LSPs in one of two ways:
Static configuration: Static LSPs are a lot like static routes. You basically have to explicitly configure every LSR in an LSP manually. Because no protocols dynamically signal the LSP for you, the load on the LSRs is reduced. However, if you have changes in the topology, the paths can’t adapt to the new network. As a result, topology changes create routing black holes.
The lack of dynamic update is a significant drawback and one that should not be overlooked. Dynamic LSPs should be used wherever feasible.
Dynamic setup: Dynamic LSPs use signaling protocols to establish themselves and propagate LSP information to other LSRs in the network. You configure the ingress router with LSP information that is transmitted throughout the network when you enable the signaling protocols across the LSRs.
Note that you have to configure the signaling protocols on all of the LSRs. If only a subset of routers is able to exchange information, the LSP isn’t established.
Because the LSRs must exchange and process signaling packets and instructions, dynamic LSPs consume more resources than static LSPs. However, dynamic LSPs can avoid the network black holes of static LSPs by detecting topology changes and outages and dynamically establishing new LSPs to move around the failure.