Re: draft-bourbaki-6man-classless-ipv6-00

[Erik Kline <[e-mail]>]

On 14 June 2017 at 22:06, Mark Andrews <[e-mail]> wrote:

>
> In message <[e-mail]>, Peter Hessler writes:
> > On 2017 Jun 14 (Wed) at 11:52:27 +0200 (+0200), [e-mail]
> wrote:
> > :Peter,
> > :
> > :> leaving a RECOMMENDED for /64 subnets is fine.  forcing it to be a
> MUST
> > :> is not fine to me.  While the arguments in favour of /64 subnets are
> > :> valid, there are many differently valid reasons to have non-/64
> subnets,
> > :> as mentioned many times in this (and other recent) threads.
> > :
> > :Which arguments matter to you?
> > :What _problem_ does changing the 64 bit boundary solve for _you_?
> > :
> > :Ole
> >
> > The problem is wasting space (I acknowledge your arguments, and reject
> > them), uglyness of the addresses, and the fact that _my_ network is not
> > _your_ network.
> >
> > I'm already running non-/64 subnets on my personal networks.  I'm also
> > running non-/64 subnets on my $work networks.
> >
> > mandating /64 subnets in the _architectural specification_ is a bug,
> > period.
> >
> > IPv4 got rid of classful subnets in 1993, IPv6 can join the last century
> > as well.
>
> We got rid of classful subnets in 1993 *because* there were not
> enough subnets to go around with them.  We didn't get rid of them
> because as a concept they were bad.
>
> Do you say that 18446744073709551616 subnets is too few to go
> around?
>
> 2^64 address would have been enough to support the world using
> variable length subnets.  IPng evaluated this and decided that it
> was not a good idea as variable length subnets really are too
> complicated.  Instead IPng went initially with 128 bit addresses
> and, initially, /80 subnets so that we didn't have to deal with
> variable sized subnets.  This was later changed to /64 bit subnets
> to handle 64 bit mac addresses.
>

Thank you.