Toerless Eckert wrote:
Brian reminded us that we have ample RFCs to elaborate on the fact that you can not reasonably expect for connections to an anycast address to work when persistently using the anycast address.
But, the reality is that no practical protocols need persistent stability.
Christian pointed out how QUIC does the right thing.
Wrong. It only makes an anycast server victim of DDOS by the alternative unicast address. Remember how IPv6 was improved to be useless with so many fancy ideas. Masataka Ohta
