Christian,
=> the draft is supposed to address both issues. For the first one
we can add more cautionary words if one believes there are not yet
enough. For the second one we already merged two very different
experiments into one request.
Well, if that is what we want, then we can write a much simpler draft,
"reserving a prefix for non routable identifiers". It will have
essentially one or two paragraph of text, plus the usual ten pages of
boiler text.
I would almost agree with you. However, as we do discuss in the
draft, we also want to use a hash function to securely bind
information into the identifier. Consequently, there is the tension
of having as many bits as possible in the hash, and spending as few
bits as possible in the prefix.
Consequently, if we want to share the prefix with multiple
experiments, we either
a) need to define some bits to define which experiment is going
on, or
b) define a means of encoding this information in the hash
We have chosen b), with the motivation of thereby getting a longer
hash. No explicit bits needed to differentiate the experiments,
resulting in a slightly longer hash.
--Pekka
--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
