Bringing Traffic Engineering and Resiliency to LDP Provisioned MPLS Forwarding Planes

In today's converged IP networks we gradually move away from the model where traffic forwarding closely follows routing. In many modern backbone networks traffic flows over a MPLS based forwarding plane that consists of a set of MPLS tunnels. In most cases these tunnels are provisioned using RSVP-TE, are computed and optimized off-line, failures are repaired by bypass tunnels, and the routing protocol (IGP) has a rather limited role in how traffic is routed. Compared to RSVP-TE provisioned MPLS forwarding planes, LDP provisioned MPLS forwarding planes have significant advantages in terms of scalability and simplicity, but they are still tied to the IGP routing and do not provide mechanisms for traffic engineering or traffic protection. In this work we build on our previous work in multi-topology routing and IGP link weight optimization and show how with simple extensions to the LDP protocol we can build MPLS forwarding planes that match the resilience and traffic engineering capabilities of the RSVP-TE provisioned planes without excessive label consumption.