When designing with BFD, which consideration is true?

BFD (Bidirectional Forwarding Detection) is a lightweight, low-overhead protocol designed to detect failures in the forwarding path between two systems in a very short time, often sub-second. It achieves this by sending rapid hello packets. This rapid failure detection is crucial for dynamic routing protocols like OSPF, EIGRP, and BGP, allowing them to quickly adapt to topology changes and reroute traffic, thus providing subsecond convergence for the routing plane. BFD is not typically used directly with first-hop routing protocols (FHRPs) like HSRP or VRRP for subsecond convergence; FHRPs have their own hello mechanisms. While BFD can complement NSF (Non-Stop Forwarding) and Graceful Restart by providing faster failure detection, it's not used with them in the same way it integrates with dynamic routing protocols; NSF/Graceful Restart are control plane resilience mechanisms, while BFD is a data plane failure detection mechanism. Finally, BFD is generally less CPU-intensive than drastically reducing routing protocol hold timers because BFD is specifically optimized for rapid, low-impact detection, whereas very aggressive routing protocol timers can significantly increase CPU load due to frequent updates and recalculations.