Wes Hardaker wrote:
Protocols and implementations should generally respond using the
address to which the request packet was sent. That being said, there
are sometimes protocol reasons not to do this and sometimes
implementations don't necessarily handle things properly internally.
But, I doubt most protocols ever specify the "legality" of what
address is used to respond to a request.
RFC1035 says it a bug. So, it should be illegal.
- Some name servers send their responses from different
addresses than the one used to receive the query. That is, a
resolver cannot rely that a response will come from the same
address which it sent the corresponding query to. This name
server bug is typically encountered in UNIX systems.
However, the "bug" is overcome by ID field provided in DNS.
SNMP is seemingly doing the same thing with request ID.
Ietf mailing list