Re: [v6ops] DoS attacks (ICMPv6-based) resulting from IPv6 EH drops

[Tore Anderson <[e-mail]>]

Hello 神明達哉,

> I assumed that stateless translation always uses the well-known
> prefix.  Maybe I was incorrect about that?

Yes, that is incorrect. Per RFC 6052 the operator can use either the
Well-Known Prefix and a Network-Specific Prefix. There are quite few
reasons why one would choose the latter; for example it allows for the
translating device to be reached across the public internet. Another
reason is that only a single WKP exists, so if the operator wants to
deploy both SIIT and NAT64, and/or multiple instances of either
technology, all of his deployments (except for one), must necessarily
use an NSP.

Tore