Re: [v6ops] Need ref to 3GPP spec for the use of DHCPv6 Prefix Delegation, e.g. for smartphone tethering

[Alexandru Petrescu <[e-mail]>]

Le 07/10/2013 14:19, Vízdal Aleš a écrit :
>> -----Original Message----- From: [e-mail]
>> [mailto:[e-mail]] On Behalf Of Alexandru Petrescu
>> Sent: Monday, October 07, 2013 1:44 PM To: Mikael Abrahamsson Cc:
>> [e-mail] Subject: Re: [v6ops] Need ref to 3GPP spec for the
>> use of DHCPv6 Prefix Delegation, e.g. for smartphone tethering
>>
>> Le 07/10/2013 10:33, Mikael Abrahamsson a écrit :
>>> On Mon, 7 Oct 2013, Alexandru Petrescu wrote:
>>>
>>>> Hello,
>>>>
>>>> I need a reference to a 3GPP spec which requires or explains
>>>> the use of DHCPv6 Prefix Delegation for smartphone tethering
>>>> (mobile hotspot).
>>>>
>>>> This is because we write a problem statement draft in DHC WG
>>>> ""Route Problem at Relay during DHCPv6 Prefix Delegation".
>>>>
>>>> I would like that draft to reflect the problem not only in the
>>>>  homenet-kind of deployments (DSL, broadband forum) but also
>>>> smartphone deployments.
>>>
>>> http://www.3gpp.org/ftp/Specs/html-info/23402.htm
>>>
>>> If you download the 10.8.0 spec you will find:
>>>
>>> "4.7.5    IPv6 Prefix Delegation using S2c Optionally a single
>>> network prefix shorter than a /64 prefix may be assigned to a PDN
>>> connection (TS 23.401 [4]). When S2c is used to access a PDN, the
>>> UE acting as a Mobile Router may request delegation of one or
>>> more IPv6 prefix(es) via DHCPv6 Prefix Delegation signalling as
>>> described in RFC 6267 [56]. The UE does not need to explicitly
>>> register these additional prefixes using S2c signaling as
>>> implicit mode registration is used."
>>
>> Thanks for the reference.  This is the kind of use at 3GPP that I
>> am looking for.  I will use it right away.
>>
>> (for info: there are some errors in the above text (I guess they
>> mean RFC 6276 bit 6267; second, I doubt Prefix Delegation with
>> NEMO is needed in this context - but rather simple PD without
>> tunnels; RFC 6276 assumes the Relay is on the mobile router,
>> whereas the cellular deployment would host the Relay in the fixed
>> network).)
>
> I guess that the right RFCs shall be 3633 & 6603.

I think yes, sure, that's better, that's the pure PD and the PD-exclude
RFCs.  Not sure about the necessity of PD-exclude in a 3GPP specific
scenario, but there may exist a reason for that -exclude.

Remark though that in that spec I couldnt find enunciating a problem of
routing at Relay, during the Prefix Delegation process.

It's strange, because the broadband forum _has_ identified such a problem.

Either there is no problem, in which case 3gpp would rather advise bf to
not worry  about it; or there is, and 3gpp would need to ack it.

Alex
PS: the BroadBand's problem at Relay during Prefix Delegation is
described in technical report titled "IPv6 in the context of TR-101",
dated November 2010 as:
> "hosts receiving IPv6 addresses from the RR are not known to the
> BNG, i.e. the BNG is not aware of what addresses/prefixes are
> assigned to hosts attached to the RG acting as RR."