IPv4 literals and IPv6-only hosts

Cameron Byrne cb.list6 at gmail.com
Fri Sep 17 16:59:58 UTC 2010


In planning for IPv4 exhaust, i have been tinkering with NAT64/DNS64
in preparation to launch it with customers.  In my experience it works
well, my phone has been Ipv6-only + NAT64 for over 6 months.... no
major roadblocks.    There have also been other documented NAT64
deployments that work well, especially on mobile devices
http://tools.ietf.org/html/draft-arkko-ipv6-only-experience-01

I do not believe that there is much risk in deploying NAT64, the
customer experience can be engineered on mobile to be very good, but
there is some minor risk that does not need to be there because of
content providers using IPv4 literals, which are IPv4 address embedded
in content (HTML, XML, apps...).  I think across the  industry people
believe using names (FQDNs) is a better practice than embedding IPv4
addresses, nonetheless it happens.  So, in an effort to help content
providers understand that IPv6-only customers will not be able to
access their IPv4 literal laden content, i have created this group
http://groups.google.com/group/ipv4literals

The group is not about name and shame or fixing every instance of
literals, it is about making sure that the risks of using IPv4
literals are known by the content providers and that they have a good
opportunity to fix it, start using FQDNs or deploy IPv6.

Also, if anyone has a good nytimes.com or Amazon video on demand
contact, you may want to forward this on to them.

Thanks,

Cameron
http://groups.google.com/group/ipv4literals
http://groups.google.com/group/tmoipv6beta




More information about the NANOG mailing list