IPv4 can eventually do it via APIPA and make it’s own link local addresses, but only as a fallback after DHCP times out. With IPv6, link-local is the first step of SLAAC. The interface comes up, you instantly get a fe80:: address, and you have immediate connectivity on the switch without waiting for DHCP to fail. When deploying unprovisioned (not having to set a static IP for each host) embedded Linux images, I prefer IPv6’s native design over IPv4’s error state.
IPv4 can eventually do it via APIPA and make it’s own link local addresses, but only as a fallback after DHCP times out. With IPv6, link-local is the first step of SLAAC. The interface comes up, you instantly get a fe80:: address, and you have immediate connectivity on the switch without waiting for DHCP to fail. When deploying unprovisioned (not having to set a static IP for each host) embedded Linux images, I prefer IPv6’s native design over IPv4’s error state.