In Chapter 18 and Chapter 19, you discovered different ways to protect the traffic in case of transit node or transit link failure. All the methods are based on preinstalling a backup next hop in the hardware Forwarding Information Base (HW FIB). Upon network failure, the primary next hop is removed from the HW FIB and traffic uses the preinstalled backup next hop. The difference between each method lies in how the backup next hop is determined and how the backup path is established.
This chapter explores additional FIB optimization techniques that you can implement on the ingress PE to improve failover convergence. PE nodes typically hold a large number of service prefixes. A prefix in this context can be a typical L3 VPN prefix, but it can also be information required to forward traffic using other types of services, such as pseudowires (PWs) signaled by using BGP or LDP. When it comes to optimization of failover times on the ingress PE, there are two main areas requiring special attention:
Optimization of next-hop structures in the hardware FIB
Preinstallation of the next hop associated with the backup egress PE
These optimization techniques are explored in the next two sections.
If you go back and reexamine Figure 18-1 carefully, you probably realize that failure case number 3 is not exactly the same as failure case numbers 5 or 7. Similarly, failure case number 4 is not quite the same as failure case ...