Re: [v6ops] Implementation Status of PREF64

[Owen DeLong <[e-mail]>]

> On Oct 11, 2021, at 02:05 , Alexandre Petrescu <[e-mail]> wrote:
> 
> 
> 
> Le 11/10/2021 à 02:14, Owen DeLong a écrit :
>> Alexandre,
>> You’re fighting a different battle than the one most people in this thread care about.
>> I’m not saying it’s not a good fight, but fighting uphill with 3GPP and Lorenzo on the 3GPP interface(s) is not what is delaying
>> enterprise IPv6 rollout.
> 
> Thanks for the clarification.
> 
>> Support for IA_NA on the WLAN interface(s) is the key issue here.
> 
> Thanks for clarifying it.
> 
> To me, IA_NA on WLAN for Android is easy to deal with.  We already built, installed and configured several Android smartphones with multiple DHCPv6 clients on Android WiFi interfaces.  It is at research scale, not at industry-wide deployment scale, but it worked fine.

I suspect that most enterprises aren’t going to drive their android users that way. Native support for IA_NA on WLAN in the android OS is what is creating an additional barrier to enterprise adoption.

> An enterprise network might deploy its own DHCPv6 .apk's into employee's Android smartphones, for that particular enterprise network.

Most enterprises expect this feature to be built into the OS. I suspect that any IA_NA implementation that became popular would end up in an arms race with Lorenzo disabling them in updates.

> Beyond that, maybe what IA_NA-WLAN-Android effort wants is to have these .apk's built into the Android smartphones and tablets?  Or maybe try to have these DHCPv6 .apk's accepted in the respective Store?  Has somebody tried to submit them?  What was the reply from the store owner?

I’m not up on such efforts, but I think it is safe to say that most enterprises expect the OS to come native with proper DHCP support for each applicable IP stack.

> As a side note, that problem - have DHCPv6 .apk's accepted into the respective store - is probably as hard as persuading github to put their servers on IPv6.

Well… There is progress in that area… GitHub just put some of their resources on IPv6, though GitHub itself still isn’t. As a result, I’ve moved most of the projects I have
any authority over onto Bitbucket because it does have good IPv6 support.

Owen

> 
> Alex
> 
> 
>> To the best of my knowledge, everyone EXCEPT android supports that.
>> Owen
>>> On Oct 7, 2021, at 06:18 , Alexandre Petrescu <[e-mail]> wrote:
>>> 
>>> 
>>> 
>>> Le 30/09/2021 à 08:20, [e-mail] a écrit :
>>>>>> As a matter of fact, I know of at least one deployment in a "fortune 500"
>>>>>> company that was seriously impaired due to lack of DHCPv6 support in
>>>>>> Android.  They want control over address assignment, tracking of address
>>>>>> assignments, and DHCP is the machinery they use for it (plus NAC, making
>>>>>> sure that only assigned IPv6 addresses can be used).
>>>>>> 
>>>>>> But they want to support Android devices.
>>>>>> 
>>>>>> So, still no IPv6 today...
>>>>>> 
>>>>> 
>>>>> Is there any other way to break this logjam than to implement DHCPv6 IA_NA
>>>>> and accept one IPv6 address per device? What about DHCPv6 PD or
>>>>> /64-per-host? What about resurrecting draft-ietf-dhc-addr-registration
>>>>> <https://datatracker.ietf.org/doc/draft-ietf-dhc-addr-registration/>, so
>>>>> the device can inform the network of addresses it has created?
>>>> We would love to use IA_PD only, and no IA_NA. Having DHCPv6 PD for
>>>> Android would be a significant step in that direction.
>>> 
>>> I agree, having even a limited DHCP (maybe only IA_PD, or only address registering, or other limitations) working on Android on a cellular modem would be a significant achievement.
>>> 
>>> One such limitation is 'stateless DHCP'.  It is a 'may' in 3GPP documents.  However, even that simple limited DHCPv6 is prohibited in 3GPP modems.
>>> 
>>> The problematic text in 3GPP TS 23.501 is the following:
>>>> /64 IPv6 prefix allocation shall be supported via IPv6 Stateless
>>>> Auto-configuration according to RFC 4862 [10], if IPv6 is supported.
>>>> The details of Stateless IPv6 Address Autoconfiguration are described
>>>> in clause 5.8.2.2.3. IPv6 parameter configuration via Stateless
>>>> DHCPv6 (according to RFC 3736 [14]) may also be supported.
>>> 
>>> Remark the 'may' in the last line.
>>> 
>>> In practice there is no DHCPv6 allowed by any 3GPP modem out there.
>>> 
>>> Alex
>>> 
>>>> Steinar Haug, AS2116
>>>> _______________________________________________
>>>> v6ops mailing list
>>>> [e-mail]
>>>> https://www.ietf.org/mailman/listinfo/v6ops
>>> 
>>> _______________________________________________
>>> v6ops mailing list
>>> [e-mail]
>>> https://www.ietf.org/mailman/listinfo/v6ops