Re: [Hipsec] Teredo and HIP mobility/NAT

[Miika Komu <[e-mail]>]

Hi,

On 05/23/2014 02:54 PM, Robert Moskowitz wrote:
> I have thought a lot about this and generally it works out bad no matter
> how you slice it.  Well, if I was writing the network kernel, I would
> incorporate Teredo so that all interfaces presented an IPv6 address at
> all times and if it had a 'native' IPv6 would not use Teredo.  Basically
> tying Teredo right into the interface handling?
>
> We have probably all thought long and hard about this.  Multiple
> interfaces, most of them mobile.  They are suppose to be changing their
> priority based on something or other (IEEE 802.21?)  IPv6 should be IPv6
> publicly routable.  But IPv4 will change from public, to good NAT, to
> bad NAT, and bounce around.  Because of this bad mix of reality we go to
> the lowest common denominator and do everything as if there is a bad NAT
> in the way.  We have no effective method of intelligently switching.
>
> HIP everywhere does not fix bad NATs.
>
> Networking reality basically xxxxx, well I do try and control my
> language in public.

yes, draft-keranen-hip-native-nat-traversal-01 would allow HIP a better 
control of the NAT traversal.