Re: [v6ops] I-D Action: draft-ietf-v6ops-ipv6-roaming-analysis-02.txt

[Tore Anderson <[e-mail]>]

* Ross Chandler

>> It might be that when the UE visits a network that doesn’t allow
>> IPv6 in the user-plane Android behaviour is kicking in.
>> 
>> I’ve been told that Android treats all APNs as “roughly equal”.
>> 
>> If the new IPv6-only-without-IPv4-fallback APN fails then Android
>> might be trying the IPv4 one.
> 
> If this is what’s going on in Telenor case the plus is it maximises
> the number of roaming partners that come up with IPv6 but at the
> possible cost of indeterminism on which APN the UE settles on using
> when it gets back home.

I'm fairly certain there is no such indeterminism.

Basically the UEs get set up to request IPv6 PDP type both home and when
roaming. So when they establish a data bearer they get IPv6 (+464XLAT)
no matter where in the world they are, or they get no data bearer at all
(if there's some sort of problem somewhere).

That Telenor feels confident in doing it in this manner is very good
news, I think. It reminds me of the World IPv6 Launch, actually - where
it was decided that the bugs and issues that had previously made it
unsafe to deploy IPv6 on a global scale (as opposed to a select few
known good networks, corresponding well to "the home PLMN only" in the
mobile case), has been sufficiently solved so that IPv6 can be made
generally available instead.

I guess now it's up to Cameron and Michał to follow suit. ;-)

Tore