Le 16/02/2017 à 21:52, Mark Smith a écrit :
On 17 February 2017 at 07:31, Randy Bush <randy@xxxxxxx> wrote:
What I think we need is to make it clear that there are real
exceptions to 64, and it is therefore not acceptable to embed 64
in code.
close. 64 is the exception; so slaac can work.
hiding the steaming classful pile in cute places around the
document only gets folk more annoyed.
So what is your objection to having a well-known addressing
structure?
Objecting to it, calling it "classful", is implying that it is
exactly the same as IPv4 classful addressing. I don't think IPv6 is.
Perhaps not precisely the same, because there are no 'IPv6 classes'.
But, an IPv4 Class C is quite the same as an IPv6 prefix/64. The limit
is 24bit in IPv4, and 64bit in IPv6. True, the forwarding in IPv6 is on
bit-boundary where classful IPv4 is on byte-boundary, yet an otherwise
bit-boundary IPv6 _>/64_ route can not reach computers relying on
SLAAC/Ethernet /64.
That makes it a limit, which in turn generates what appears to be an
'IPv6 class' - the leftmost 64bits.
Alex
My memory of classful IPv4 addressing is that it not only specified
an addressing structure, it also specified a forwarding method based
on that structure. (e.g., the default route was only used if there
weren't any routes in the table that had the class matching the class
of the destination address.)
IPv6 doesn't. The addressing structure doesn't specify how different
bits are used during forwarding - all 128 are, using a longest match
per BCP 198.
So IPv6 specifically separates the addressing structure from the
forwarding method. I think calling that "classful" is
mischaracterising it.
Regards, Mark.
randy
--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@xxxxxxxx Administrative Requests:
https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------