While I was labbing Dynamic ARP Inspection I pinged a device (these systems had not communicated with each other yet) and was surprised to find the recipient generate an ARP broadcast to learn where to respond to the ping. I assumed that because the recipient would see where the ICMP message originated from based on the Ethernet header it would just add that to its ARP table and save itself the trouble and latency of an ARP.
This was between Linux and a Cisco ASA. Are there other host operating systems that behave differently?
No comments:
Post a Comment