[AusNOG] IPv6: Who's dual stacked? Why don't I look stacked?

Mark Andrews marka at isc.org
Fri Mar 8 10:10:41 EST 2013


In message <CAOu9xNJ9_mkcAFaBkk40qhmu7gpDe1c581w87ck91kRk61h3hA at mail.gmail.com>, Robert Hudson writes:
> 
> (Apologies for the top-post - I am mobile).
> 
> I wondered something similar yesterday. I am an Internode customer at home,
> ans have a Fritzbox as my CPE, Win7 as my desktop OS, and Android 4.1 on my
> phone.
> 
> I know my Win7 box has full IPv6 connectivity (having run dozens of tests),
> but to my knowledge my Android phone doesn't yet support IPv6 (
> https://code.google.com/p/android/issues/detail?id=3389 - partial support
> available in 4.2 apparently).
> 
> What I don't know or have any visibility of is how much of my Internet
> traffic is IPv6 vs IPv4.  

For the most part one shouldn't need to know.  This sort of level
is really only of interest to techies and packet traces or stats
counters on routers can give this.

> For that matter, really of what apps do and do not
> support IPv6 transit and which do not, and even if an app (a mail client
> for example) supports IPv6, whether the server (Internode's mail server for
> example) supports IPv6. And finally, if everything supports IPv6, what
> priority is given to IPv6 over IPv4 (or vice versa).

IPv6 is given priority over IPv4 for most things.  The long term
plan is to be able to turn off IPv4 for wide aread traffic and not
have stuff tunneled.  This does mean that if you have a IPv6 tunnel
it will get the traffic if the remote end is dual stacked.

> I am sure all the answers to all of these questions are out there
> somewhere, but as someone who has about a billion other things on my plate
> right now, spending the time to find out and/or make any changes required
> to improve my IPv6 utilisation is right up there in my priority list along
> with things like booking my first moon flight and discovering the cure for
> broken fingernails, given things "just work" right now for me regardless of
> the version of IP being used for transit.

And when you introduce IPv6 they will continue to "just work".
 
> This leads me to believe we have two other problems to solve, and that
> we''re also using a potentially bad metric for judging success. The two
> problems are visibility and priority. The metric (what percent of global
> traffic is IPv6 is bad because although I am ready, right now on my home
> network for IPv4 exhaustion and an IPv6-only world, I doubt I am mail g use
> of that readiness - and I've highlighted another piece of the puzzle that
> needs tracking - application readiness.

You can also measure local traffic.  Eyeball networks get significant
amounts of IPv6 traffic once it is enabled.  My home net is around
50%.  This will only increase as more and more providers switch on
IPv6.

If you are a provider about 1-2% of your customers are already IPv6
enabled.  Turning on IPv6 will draw traffic from those 1-2%.  As
more eyeball networks turn on IPv6 for their customers this will
increase.

> So, does there exist:
> 
> a) A register of IPv6 capable/enabled/exclusive applications/operating
> systems/hardwRe? If not, I am happy to start one.
> b) A way for me to tell, on the fly, just how much of my personal IP
> traffic is actually IPv6?

 
-- 
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: marka at isc.org



More information about the AusNOG mailing list