What is generally used for an iBGP session to form?

• A. A loopback address is generally used in the BGP neighbor command
• B. A physical interface address
• C. A multicast address
• D. A randomly chosen address

Answer: (A)

Using a loopback address for an iBGP session is common because it provides a stable, always-up endpoint for the BGP TCP connection. Loopbacks are logical, not tied to a single physical interface, so they tend to stay up even if interfaces flap or IPs on those interfaces change. This makes peering more reliable over time.

To form the session with loopbacks, you configure the neighbor to be the loopback IP of the remote router and you also set the local source of the BGP session to the loopback as well (update-source loopback). You must ensure there is reachable connectivity to the loopback addresses, typically via an IGP or static routes. On some platforms you may also need to account for multi-hop reachability if the loopbacks are not directly connected, but the core idea is using stable loopback addresses for the peering endpoint rather than a potentially transient interface IP.

Using a physical interface address is less robust because that IP can change with the interface, link states, or reboot, which can disrupt the session. A multicast address isn’t used for BGP because BGP relies on unicast TCP, and a randomly chosen address wouldn’t guarantee reachability or consistency.