Le 25/10/2019 à 22:08, John Levine a écrit :
In article
<CAPTMOt+N6d+0Ucsf_9iNSDmZ8i741eLMtNnYvy-WW_dKLW2xZg@xxxxxxxxxxxxxx>
you write:
-=-=-=-=-=-
Fernando Gont had said:
No. RFC7217 simply says that the algorithm works with any IID
lengths. But it doesn't change the requirement of /64 for SLAAC.
SLAAC not having a /64 requirement makes think that nothing is actually
changed by RFC7217.
Yet OpenBSD does change some things when it uses IPv6 addresses with
IPv6 addresses of non-64 prefix lenghts.
Well, my only intention was to show that assignment of 64 bits IP
addresses per subnet is wasteful.
It might be, but since there is no chance whatsoever that the IPv6
address size will change, why are we even arguing about it?
The IPv6 address size will not chage.
However, experience with addressing system variations may help IPv6
design within that address size (128 bit).
Variations of addressing systems are, for example: LISP protocol, DHCPv6
Prefix Delegation, and (non-std) variable-length Interface IDs.
Other new variations of addressing systems may include new paradigms
like point-distance: my IP address is 3 units away from that one. This
could help in ISP deployments at home where houses are not numbered in
traditional way, but more like in point-distance way. Or it could help
new header compression schemes.
We should not stop working on addressing systems, just because the size
128bit is frozen.
As others have noted, the large addresses make it easier to configure
IPv6 networks, and to leave plenty of slack so if the network grows,
you won't have to renumber or reconfigure.
I agree.
Alex
R's, John