IPv6 Ignorance

Brian E Carpenter brian.e.carpenter at gmail.com
Fri Oct 5 22:25:26 CEST 2012

On 05/10/2012 20:55, Tim Densmore wrote:
> On 10/5/2012 11:49 AM, Brian E Carpenter wrote:
>>> Or is the assumption that we'd give them, say, a
>>> >v4 /29 and a v6 /48 and have them do their own nat64 or whatever?
>> Why would they need to do NAT64? If they are dual stacked, they are
>> home free.
>>     Brian
> Hi Brian,
> The proposed hypothetical customer needed a /24 according to their
> requirements.  A /29 falls pretty far short without some sort of
> NAT/PAT.  

If they need a /24 for client hosts (desktops etc.) then they will presumably
use RFC 1918 and NAT44 like anybody else today. If they need a /24 for publicly
accessible servers, tough luck - we ran out!

> I suppose it's possible that dual stack would make up for
> that, but I wouldn't expect it to, currently.  

No, but the day will come when service providers have no choice but
to offer v6 public services because they can't get v4 addresses.

> Anyway, this is well into
> the realms of straw men - my current "big blocks" concern isn't a
> collo'd server, it's a hotel that needs multiple IPs per room or the
> WISP that needs a /24 and a /25.  So far, IME, IPv6 doesn't help in
> either of these cases.  I'll be very glad when it does.

It will help more as the snowball rolls further. There is no 3 month return
on investment here - we just all have to do our bit if we want our industry
to continue growing. Putting IPv6 behind every connector as well as
RFC1918 plus NAT44 is each site's way of rolling the snowball.


More information about the ipv6-ops mailing list