Re: Informed regulator about the shorter-than-64 necessity on 3G/4G/5G

Le 19/01/2021 à 04:19, Mark Andrews a écrit :
> 
> 
>> On 19 Jan 2021, at 09:04, Fernando Gont <fgont@si6networks.com> 
>> wrote:
>> 
>> Hi, Alex,
>> 
>> On 18/1/21 05:45, Alexandre Petrescu wrote:
>>> Hi, 6MANners, Today I formulated a brief personal email to some 
>>> person(s) at the regulator agency suggesting the allocation of 
>>> shorter-than-64 prefixes (e.g. /56) to end users. I explained 
>>> that the current situation where each of the mobile operators 
>>> deliver a /64 and not shorter to smartphones is not usable for 
>>> networks such as mobile hotspots or in-car multi-subnet
>>> networks. I informed that there are some I-D proposals of
>>> Variable SLAAC that is relatively refused at IETF because of the
>>> IETF necessity of 64bit (and not shorter) IIDs.
>> 
>> This seems to be confusing things.
>> 
>> Proposals for "classless IPv6" (e.g.: 
>> draft-bourbaki-6man-classless-ipv6) are meant to allow SLAAC to 
>> work with *longer* prefixes (i.e., shorter IIDs)
>> 
>> Shorter prefixes (i.e., bigger address blocks leased to users) are
>>  already supported by DHCPv6-PD, without the need of any additional
>>  protocol action.
> 
> I think the intent of the message was to get regulators to force 
> operators to *use* DHCPv6-PD with short prefixes.

Could be an idea.

I first formulated it as a need of other-than-64 prefix lengths, and not
in terms of a protocol to require.

We might see whether or not the regulator considers the necessity to
require a particular protocol such as DHCPv6 and DHCPv6-PD.

In the context, this particular regulator was able to formulate a
intentionally brief requirement of deploying 'IPv6 compatible routing'
when allocating spectrum for 5G to mobile operators; it means operators
get 5G spectrum only if they commit on a simple IPv6 commitment too
(among other non-tech commitments).  However, the regulator has
reluctance in expanding that brief 'IPv6 compatible routing'
requirement, because of having experienced too complex requirements that
were too hard check compliance with.

This is one of the reasons I suspect that formulating a simple req only
on the prefix length (and not on protocol compliance) might have some
chance to get through all the way to operators.

Another reason is the following: I suspect a non-64 plen requirement
might be dealt with between the regulator and the operator, whereas a
DHCPv6 protocol requirement might need to involve the modem manufacturer
as well.  It would be a longer path between 3 parties, rather than 2.

Alex

> 
> There is too big a power differential between the cellular user that
>  want to be able to have multiple subnets and the telcos for free 
> market economics to work.
> 
>> Thanks, -- Fernando Gont SI6 Networks e-mail:
>> fgont@si6networks.com PGP Fingerprint: 6666 31C6 D484 63B2 8FB1
>> E3C4 AE25 0D55 1D4E 7492
>> 
>> 
>> 
>> 
>> --------------------------------------------------------------------
>>
>>
>>
>> 
IETF IPv6 working group mailing list
>> ipv6@ietf.org Administrative Requests: 
>> https://www.ietf.org/mailman/listinfo/ipv6 
>> --------------------------------------------------------------------
>