[menog] Rapid IPv6 deployment for World IPv6 Day

[Ahmed Abu-Abed]

Comments in italic below.

-Ahmed


From: Brian Candler 
Sent: Wednesday, June 08, 2011 5:53 PM
To: Ahmed Abu-Abed 
Cc: 'menog at menog. net' ; IPv6jordan 
Subject: Re: [menog] Rapid IPv6 deployment for World IPv6 Day


On Wed, Jun 08, 2011 at 12:56:39PM +0300, Ahmed Abu-Abed wrote:
>    World IPv6 Day ends at 3am KSA/Jordan time June 9th.
> 
>    To setup a computer with IPv6 public address over any IPv4 connection
>    (3G, ADSL, dial-up, etc. and even behind nested NATs) for World IPv6
>    Day testing then I suggest downloading the Freenet6 client and running
>    it with default settings. Then use it to access Google, Youtube,
>    Facebook, CNN, etc. or ping their sites.

This is arguably missing the point of IPv6 day.

If you want to play with a v6 tunnel client, you can do this any day.  You
simply point your browser at http://ipv6.google.com/ or
http://www.v6.facebook.com/ to see if it works.

>> The purpose of World IPv6 Day (W6D) was for everyone to participate and test the behavior of standard web portal URLs on both IPv4 and IPv6 connections. 99% of internet users do not have access to native dual stack connections, hence tunneling to IPv6 servers is the only solution for users to access the websites over a v6 connection.

However, the Internet migration strategy is dual-stack. There will never be
any requirement for users to install tunnel clients on their machines.

>> Until the whole internet AND web content AND networks AND applications move to IPv6 ONLY then there will be a need for tunneling. Dual-stacking is needed but it doesn't solve the IPv4 depletion issue, and multi-level NATing (CGN and LSN, etc) creates its own set of problems. Forward tunneling, IPv6-in-IPv4 is in use now for rapid IPv6 access, while reverse tunneling IPv4-in-IPv6 will be needed in the future when networks IPv6-only with legacy IPv4 content.


The fundamental reason for v6 day is to encourage *content providers* to
enable both v4 and v6 concurrently on their *well-known* URL (i.e.  return
both AAAA and A records), and then see how many of their (non-v6) users are
broken by doing this.  That would be users whose local v6 stack is broken,
or who are on a network which announces v6 connectivity when it doesn't
actually have it.

>> See my first comment above.

So, this is what's different for v6 day:

$ dig www.google.com aaaa
...
;; ANSWER SECTION:
www.google.com. 86094 IN CNAME www.l.google.com.
www.l.google.com. 218 IN AAAA 2a00:1450:400c:c01::93

Previously, you'd have got an empty response (and the client would then
look for an A record instead).

The hope is that this will give confidence to those websites to run both v4
and v6 permanently on their main URL.

Regards,

Brian.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: /pipermail/attachments/20110609/10a8292f/attachment-0001.html