On Sun, Nov 6, 2011 at 8:32 AM, Tom Hill <tom@ninjabadger.net> wrote:

On Fri, 2011-11-04 at 15:04 -0700, Cameron Byrne wrote:

...

FYI.

T-Mobile USA now has opt-in beta support for an Android phone on IPv6, more info here https://sites.google.com/site/tmoipv6/lg-mytouch

Very, very good. I hope T-Mobile UK (and elsewhere in the world) take heed.

I have to wonder why they've chosen to go with IPv6-only & DNS/NAT64 instead of a dual-stack approach. Is there a particular restriction that prevents this?

There are a variety of reasons. Most prominent is that if the issue is lack of IPv4 addresses (public and private), dual-stack does not solve this problem, each device still gets an IPv4 address. Another major issue is that in GSM/UMTS (3GPP pre-release 9), having dual-stack means having 2 attachments to the network, one for v4 and one for v6. Most mobile providers pay for most of their network kit in terms of these attachments known as PDP. Consequently, dual-stack doubles the of the packet-core network. If we take the licensing and contractual parts out of the equations, double the attachments means double the signalling and mobility events ... resulting in double the CPU / Memory / blah ... LTE does not have the dual attachment problem since there is the concept of having v4 and v6 in one attachment, but it does not change the fact that there are not enough IPv4 addresses to go around, especially from a strategic planning perspective (let's design this once for 5 to 10+ year life ...)

...

As far as i know, this is the first Android phone that support IPv6 on the GSM/UMTS mobile interface.  Previous version of Android phones supported IPv6 on WiFi and LTE.

Indeed, the 'Network Info II' application will show you the IPv6 addresses gained on WiFi interfaces (if anyone's interested). My Galaxy S (unlocked/orig.) does this very well.

I wonder if it's possible to provide IPv6 support for UMTS/GSM via firmware and/or software updates from Samsung?

That's a good question for Samsung. Most vendors would rather have you buy a new device :( CB

Tom

LTE does not have the dual attachment problem since there is the concept of having v4 and v6 in one attachment, but it does not change the fact that there are not enough IPv4 addresses to go around, especially from a strategic planning perspective (let's design this once for 5 to 10+ year life ...)

Most networks seem to dish out address space behind a LSN box these days. I have three dongle things from three networks in the UK, none of them give me a public address. -- Leigh Porter ______________________________________________________________________ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email ______________________________________________________________________

Hi Cameron, On Sun, 2011-11-06 at 21:31 -0800, Cameron Byrne wrote:

There are a variety of reasons. Most prominent is that if the issue is lack of IPv4 addresses (public and private), dual-stack does not solve this problem, each device still gets an IPv4 address. Another major issue is that in GSM/UMTS (3GPP pre-release 9), having dual-stack means having 2 attachments to the network, one for v4 and one for v6. Most mobile providers pay for most of their network kit in terms of these attachments known as PDP. Consequently, dual-stack doubles the of the packet-core network. If we take the licensing and contractual parts out of the equations, double the attachments means double the signalling and mobility events ... resulting in double the CPU / Memory / blah ...

That'll probably explain it... Thanks. :)

LTE does not have the dual attachment problem since there is the concept of having v4 and v6 in one attachment, but it does not change the fact that there are not enough IPv4 addresses to go around, especially from a strategic planning perspective (let's design this once for 5 to 10+ year life ...)

If only the UK was as far ahead on LTE as the US! Tom