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

[Tim Chown <[e-mail]>]

Hi,

> On 3 Jun 2017, at 22:12, Brian E Carpenter <[e-mail]> wrote:
> 
> On 04/06/2017 05:32, Roger Jørgensen wrote:
>> On Fri, Jun 2, 2017 at 4:11 PM, Job Snijders <[e-mail]> wrote:
>> <snip>
>>> Abstract:
>>>   Over the history of IPv6, various classful address models have been
>>>   proposed, none of which has withstood the test of time.  The last
>>>   remnant of IPv6 classful addressing is a rigid network interface
>>>   identifier boundary at /64.  This document removes the fixed position
>>>   of that boundary for interface addressing.
>> 
>> what I find odd is that we over and over again are morphing IPv6 into IPv4
>> with just more IP adresses... 
> 
> That simply isn't true. The draft doesn't abolish the concept of 'interface
> identifier', which doesn't exist in IPv4. It doesn't attack SLAAC. It
> doesn't attack ILNP or draft-herbert-nvo3-ila. It doesn't attack the choice
> of /64 for all IPv6-over-foos to date.
> 
> It does say two things.
> 
> 1. BCP198
> 2. n in the addressing architecture is a parameter.

And it that sense, it’s fine, i.e. /64 is RECOMMENDED, and used for the various IPv6-over-foos that Brian mentioned, but also it sends a message to not hardcode /64.

The draft could cite RFC7421 and (for the point about address availability) RFC7934. 

Like David, at the university where we deployed IPv6, we didn’t try to lock down addresses to hosts; rather we used SNMP-based polling of network devices for accountability. We also had 802.1X (through eduroam) on WiFi, and I know of some universities deploying 802.1X for wired networks as well. There’s also the Cisco ND syslogging capability if you use their platform. 

Tim