MAP-T in production

Mon Jul 27 19:51:49 UTC 2020

While I believe I in time will get at least 90% of residential users on
IPv6, the track record for commercial customers is close to 0%. Also the
number of websites and other Internet services with ipv6 is abysmal. We are
somewhat saved by the American internet gigants which means that by traffic
volume we are already halfway there.

My take is that the so called transition is going to last forever or so
long that it could as well be. By volume and by end users we might get to
90% but no more. We need to plan for that and accept that is how it will
be.

IPv6 is already saving money by reducing the need for NAT444 or NAT64. That
50% IPv6 traffic just doubled the number of users per NAT server. If we can
get that to 90% we will have 10 times the users per server. Huge thanks to
Netflix, Google, Apple, Akamai, Facebook etc for that.

With that said I would love to one day deploy MAP or LW4o6 simply to get
rid of that NAT server. We need IPv4 as a service and stateless in the
network. If we need to keep ipv4 around for a long time, we can as well
make it easy and cheap.

Regards

Baldur

man. 27. jul. 2020 21.05 skrev Brian Johnson <brian.johnson at netgeek.us>:

> NAT444 CGN does NOT solve an IPv6 problem at all. It solves an IPv4
> shortage problem at best and is not designed as a long-term solution. I
> cannot force customers to buy new equipment to make them IPv6 compliant.
> The best option is to support, fully and unabashedly, IPv6 and help with
> the transition from IPv4 using techniques to solve for the
> problems/corner-cases.
>
> All transition technologies are band-aids. We are talking about ways to
> bridge a gap. Anyone looking at any of these techniques as an end design
> goal has missed the IPv6 point all together and is not serving their
> users/customers well. In the end, we will have everyone on IPv6 and this
> entire conversation is mute.
>
>
> BTW… name a transition technology that is supported by all legacy
> equipment…. I’ll enjoy the silence. ;)
>
> > On Jul 26, 2020, at 7:23 AM, Mark Tinka <mark.tinka at seacom.com> wrote:
> >
> >
> >
> > On 25/Jul/20 03:24, Randy Bush wrote:
> >
> >> a great path.  fork lift all cpe and cgn in the core.  the vendors'
> >> dream
> >
> > All major vendors are shipping IPv6. Some even 464XLAT. And yet they
> > will not put those forward as long term solutions.
> >
> > As Randy points out, CG-NAT sells plenty in license fees. And you need
> > more and more, every year. Not less.
> >
> > So, go ahead and enrich the vendors, at the expense of your business.
> > And we shall all wonder why they keep saying "But no one is asking for
> > IPv6".
> >
> > Mark.
> >
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.nanog.org/pipermail/nanog/attachments/20200727/db0551fc/attachment.html>