Re: Deprecating IPv6 (Re: draft-bourbaki-6man-classless-ipv6-00)

[Lorenzo Colitti <[e-mail]>]

On Tue, Jun 6, 2017 at 3:47 AM, james woodyatt <[e-mail]> wrote:

> Now is as good a time as any to repeat that I’m working on delivering
> *basically* this now. Not in ten years. Now.
>

Here's a solution that works now: ND proxying.

OpenWRT has had code to do it for a few years now. IIRC they're on to their
second implementation, the first one was here
<https://wiki.openwrt.org/doc/uci/6relayd>.

IIRC when I wrote my own proof-of-concept implementation it wasn't too
hard. It's particularly easy when you don't need to support autoconf on the
southbound interface, which you presumably don't need to support since
6LoWPAN has IPv6 addresses that are known in advance. AIUI this is similar
to what the Android on ChromeOS does: the Android code runs in a separate
namespace (equivalent role to the 6LoWPAN network), and gets an IPv6
address. There's a daemon on the root namespace (equivalent role to the BR)
runs ND proxy for that address. That code may be open sourced under a
license that would allow you to use it as is.

Running ND proxying on Linux is pretty easy even without writing any packet
handling code using the IPV6_JOIN_ANYCAST socket option. Apache licensed
code (running on production Android devices) is here
<https://android.googlesource.com/platform/external/android-clat/+/master/setif.c>,
look for do_anycast_setsockopt.

We are already on the path to IPv6/NAT w/ address amplification. We had a
> chance to stop it, and we blew it. It’s time to move on from that mistake.
>

You may be on that path at the moment, but I don't think you were obliged
to take it. You could have (and maybe still can) implement ND proxying.