Re: [v6ops] I-D Action: draft-ietf-v6ops-host-addr-availability-04.txt

[Alexandre Petrescu <[e-mail]>]

Le 06/01/2016 23:55, Templin, Fred L a écrit :
> Hi Gert,
>
>> -----Original Message-----
>> From: Gert Doering [mailto:[e-mail]]
>> Sent: Wednesday, January 06, 2016 1:21 PM
>> To: Templin, Fred L
>> Cc: Gert Doering; Lorenzo Colitti; [e-mail] WG
>> Subject: Re: [v6ops] I-D Action: draft-ietf-v6ops-host-addr-availability-04.txt
>>
>> Hi,
>>
>> On Wed, Jan 06, 2016 at 07:57:14PM +0000, Templin, Fred L wrote:
>>> As we have been saying, this could result in lots of multicast ND messaging
>>> on the link over which the IA_NAs are received when clients configure lots
>>> of addresses. It is also incompatible withclient mobility,.
>>
>> Supposedly multicast is one of the strong points of IPv6,
>
> Many people are saying that it is too harsh for some wireless access technologies.
>
>> and for mobility, there is mobile IPv6.
>
> Mobile IPv6 is about associating a mobile host's home address with a
> care-of address through the insertion of a mobility extension header
> (i.e., an encapsulation). It is about a mobile host that goes wandering
> away from its home link yet wants to be able to still communicate as
> though it was still at home. It is uni-addressing centric, whereas we
> want to support mobile devices that need multi-addressing.

Yes, it is a total network-layer centric solution smoothly hiding all 
address change events from applications.

Mobile IPv6 is also about Mobile Routers (NEMO - Network Mobility, RFC3963).

Uni-address centric?  No, because in NEMO there is also use of DHCPv6-PD 
between the Mobile Router and the Home Agent.

The only aspects that could be considered as disadvantages to Mobile 
Routers implementing Mobile IPv6/NEMO are:
- this totality coverage of address change disruption (some apps dont
   really need, some deal with it themselves).
- lack of coverage of other mobility events (e.g. lack of deal with
   intermittent reconnection to precisely same network).
- lack of securely established optimal routes (the famous RO).
- lack of multi-tier edge IP address architectures (a Mobile
   Router is in theory one CPU with several IP interfaces, whereas in
   practice Mobile Routers are themselves separated in multiple CPUs
   tightly bound by some specific interfaces some being IP-enabled
   others not).

>> If we think that these are all historic errors, and the future is IPv6
>> over GRE (or LISP), maybe we should make this more clear.
>
> Did I mention that AERO works over GRE:
>
> https://datatracker.ietf.org/doc/draft-templin-aeromin/
>
> And, my understanding of LISP is that it requires an online mapping
> database facility. This is no more or no less of a critical infrastructure
> requirement than for AERO.

I agree with this point. Many mobility solutions actually require
infrastructure support - LISP, MIP, AERO. Because one of the fundamental
principles of the Internet architecture is that one address is valid at
one single particular point in the topology (it's not like e.g. the
POBox address architecture which says that I can keep my POBox number
wherever in the postal address system I am situated).

Alex

>
> Thanks - Fred
> [e-mail]
>
>> Gert Doering
>>          -- NetMaster
>> --
>> have you enabled IPv6 on something today...?
>>
>> SpaceNet AG                        Vorstand: Sebastian v. Bomhard
>> Joseph-Dollinger-Bogen 14          Aufsichtsratsvors.: A. Grundner-Culemann
>> D-80807 Muenchen                   HRB: 136055 (AG Muenchen)
>> Tel: [phone]           USt-IdNr.: DE813185279
>
> _______________________________________________
> v6ops mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/v6ops
>