This document reserves the IPv6 prefix 64:ff9b:1::/48 for local use
within domains that enable IPv4/IPv6 translation mechanisms.
Working Group Summary
The fundamental issue was raised by the author, who is solving a problem
he observes in his network. He uses SIIT-DC to translate between IPv4
clients and IPv6-only services in his data center, and separately uses
stateless NAT64 to certain IPv4-only services. The current specification
permits the use of exactly one prefix for "the" translator,
64:ff9b::/96. If he has two or more translators (in this case, one for
SIIT-DC and one for stateless NAT64) facing different networks, he needs
to be able to distinguish them, using different prefixes within his
network. The use of 64:ff9b:1::/48 enables him to do so.
As noted in
document was proposed as an individual submission in May 2016, discussed
in the working group, and adopted in March 2017. Discussion was not
about the validity of the requirement or alternate solutions, but about
address scope, "why this prefix" (checksum neutrality), and deployment
Joe Clarke provided a helpful early OpsDir review, and the
document was updated (with assistance / input from the WG) to address
It also used to Update: RFC6890, but it was only adding to a
registry created by RFC6890, so this was removed.
Fred Baker is the Document Shepherd
Warren Kumari is RAD!