[Olsr-dev] SmartGW: GW side

[Alina Friedrichsen]

> I told you that I would like some technical reasons why we need the back 
> tunnel.

The first technical reason is, that if we have the tunnel back, we don't
need "NIIT" (unstandardized 4to6 tunnels) in a IPv6 mesh, to go into the
IPv4 internet. A standardized 4in6 tunnel is enough. Without "NIIT" we
don't have the need for a global IPv4 address in the mesh. The IPv4
address could be gateway-local. So we have not the problem with the
autoconfiguration of this address. So that the IPv6 mesh with IPv4
internet access could be completely autoconfigured.

The second technical reason is, that build the back-tunnel is in the
most cases faster then advertise a now HNA in the mesh. Is would be nice
if we use dynamic a IPv6 address in the prefix of the current gateway.
(The alternative in IPv6 to avoid NAT. :)

Alina