Re: [v6ops] I-D Action: draft-ietf-v6ops-mobile-device-profile-17.txt - C_REC#9 464XLAT

[Kossut Tomasz - Hurt <[e-mail]>]

Hi,
IPv6-only + CLAT+NAT64(no dns64) is used in Orange Poland - somewhere~ 30% mobile users. You won't see this in worldwide IPv6 stats {"No." of IPv6 users are measured form traffic -ipv4/ipv6 - typical mobile user (smartphone) generates up to 6 times lower traffic comparing to lines users..} 

Assuming our experience:
1. The best for mobile nowadays is CLAT+PLAT+DNS 
*One path for IPv4 traffic (always via CLAT)
*ALG’s treated as NAT44
*IPv4 literal & domain use same path
*One path for IPv6 traffic (native IPv6)
*Motivation for native IPv6 content
*Application address family independent
*65% - traffic from subscribers is native IPv6
*Google Nexus, Sony,HTC, Samung,LG, WindowsPhone, supports CLAT.

Cheers,
tk


-----Original Message-----
From: [e-mail] [mailto:[e-mail]] 
Sent: Monday, February 16, 2015 7:43 AM
To: Alexandru Petrescu; Heatley, Nick
Cc: [e-mail]
Subject: Re: [v6ops] I-D Action: draft-ietf-v6ops-mobile-device-profile-17.txt - C_REC#9 464XLAT

Hi Alex,

Please see inline.

Cheers,
Med

-----Message d'origine-----
De : Alexandru Petrescu [mailto:[e-mail]]
Envoyé : vendredi 13 février 2015 17:13
À : Heatley, Nick; BOUCADAIR Mohamed IMT/OLN Cc : [e-mail] Objet : Re: [v6ops] I-D Action: draft-ietf-v6ops-mobile-device-profile-17.txt - C_REC#9 464XLAT

Le 13/02/2015 16:33, Heatley, Nick a écrit :
>
>
> -----Original Message----- From: Alexandru Petrescu 
> [mailto:[e-mail]] Sent: 13 February 2015 14:35
> To: Heatley, Nick; [e-mail] Cc: [e-mail]
> Subject: Re: [v6ops] I-D Action:
> draft-ietf-v6ops-mobile-device-profile-17.txt - C_REC#9 464XLAT
>
> Le 13/02/2015 14:14, Heatley, Nick a écrit :
>> Hi Alex, Yes, that is wrong. You can ping any IPv4 literal from a 
>> 464xlat device. My handset only has an IPv6 address + CLAT and I am 
>> pinging 8.8.8.8 :-))
>
> Ok, I see.  I didnt know CLAT-on-device was required when using 
> v6-only APNs.
>
> [Heatley, Nick] For service continuity with IPv4, on an IPv6-only 
> bearer use CLAT. A slightly pedantic note,  if you have a "dual stack 
> APN" but the device only requests IPv6 then the device will receive an 
> IPv6 bearer and requires CLAT. You will hear from Ross and me talking 
> about having a single APN supporting all modes, IPv4, IPv4v6 and IPv6 
> bearers. There may be business reasons for this, as well as avoiding 
> consumers having to change APNs.

In my understanding 464xlat/clat are trials these days.

[Med] No, it is in use in live networks, including within our Group.

  They could be
improved before going live.  If the 464xlat/clat happened elsewhere than on the user terminal (e.g. BAse Station) I think more of these terminals could be accommodated, more business.

[Med] If all applications are AF-independent (including IPv4 referrals are not in use or address synthesis a la RFC6052 is supported), then CLAT can be avoided. Some operators want to adopt an IPv6-only connectivity mode but with a condition to provide the same level of service as IPv4; For those CLAT is even a must. It might happen that other operators can judge that the broken applications under an IPv6-only mode + NAT64 are not important; for those CLAT is not needed. These reasons, among others, are why we set the language to 'should'. 

> That means that the device vendors MUST implement CLAT in the 
> smartphones which connect to a v6-only APN.  It is a too strong 
> requirement.
>
> [Heatley, Nick] Today's reality is that operators cannot take devices 
> to IPv6-only mode of operation without CLAT. In the future will other 
> alternatives appear?

Alternatives there are very many.

E.g. 464lat/clat to run on an operator-controlled network device, not on end-user terminal.

E.g. use a v4-APN and v6-in-v4 encapsulation on the end-user device.

3GPP also designed a DS-MIPv6, not sure whether you are aware of it.  It had the same goal of terminal on v6-only link.

[Med] All those are no options in the mobile context. IPv6-only + NAT64 is the mode that is currently adopted by several operators.