Re: [mif] Route option for DHCPv6 - next steps?

[Alexandru Petrescu <[e-mail]>]

Le 25/04/2012 10:33, jouni korhonen a écrit :
>
> On Apr 24, 2012, at 3:20 PM, Alexandru Petrescu wrote:
>
> [snip]
>
>
>>> As for #3: RFC 6204 doesn't specify any specific technology, and it's
>>> certainly not limited to ADSL. Much of the behaviour it specifies is
>>> equally applicable to CPEs whose uplink is a cellular link.
>>
>> ?
>>
>> 'CPE' is not a term used for the cellular links, I think. RFC6204 says
>> 'residential or small-office router'. I think it is a stretch to claim
>> that RFC6204 applies to cellular links. E.g. few if any cellular
>
> RFC6204 is not necessarily the best fit for cellular (6204bis does it
> better)

Do you mean http://tools.ietf.org/html/draft-ietf-v6ops-6204bis-08 ?

> but it definitely does not preclude one making a compliant CE
> device with a cellular WAN link.
>
>> terminals use DHCP as of today, whereas RFC6204 would require them all
>> to. Also, RFC6204 requires all cellular terminals to use Ethernet
>
> So? You would most likely use your cellular just as a modem to connect
> to network and the rest of the system&  stack would be in the host side
> of the CE. This is sometimes referred as the "split-UE" in 3GPP circles.
>
>> encapsulation on their WAN interface whereas none actually does.
>
> All WLL-* requirements start with "If the WAN interface supports.."
> effectively making the following MUST conditional. Those MUSTs, like
> for ethernet, apply only when the WAN implements the said technology.

No cellular air interface to the terminal technology implements Ethernet 
encapsulation, so I don't understand the presence of that precluding If 
qualifier. (802.16 is not cellular).

>> There is another RFC - 3316 - "IPv6 for some 2G and 3G Cellular Hosts"
>> which relates more to cellular.
>
> Or a bit more recent RFC6459.

Thanks for the pointer to RFC6459 "IPv6 in 3GPP EPS".  I checked its 
Prefix Delegation section and I have comments on it, but I guess it's 
not here to be discussed about.

BAsically that section does not tell what a UE router receving a /64 can 
do with its Ethernet-compatible attached devices doing SLAAC.  In 
practice that means to either forbid SLAAC for the devices (use DHCP) or 
impossibility to use the /64 coming from the network.

This section together with 3GPP documents I looked at recently are pure 
forms of speculation with no implementation couterpart.

(in other document we propose a solution for it, which, if given way, 
makes further the case for using default routes also in DHCP).

Alex

>
> - Jouni
>
>
>>
>> Alex
>>
>> _______________________________________________
>> mif mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/mif
>
>