IPv6 content experiment

Colm MacCarthaigh colm at stdlib.net
Mon Apr 9 13:03:26 CEST 2007


On Mon, Apr 09, 2007 at 05:09:55AM -0500, Kevin Day wrote:
> When we first tried deploying IPv6, we thought we were going to be  
> doing a service to the end users by doing so. Some non-technical  
> users had IPv6 turned on without actually having IPv6 connectivity,  
> and their browsers failed to do the right thing by falling back to  
> IPv4. We got complaints from the small number of people who were  
> using IPv6 that their connection to us over IPv6 was far worse than  
> it was with IPv4, and they were disappointed that the only way they  
> could fall back to IPv4 for substantially faster downloads was to  
> turn off IPv6 completely.

When we enabled IPv6 on ftp.heanet.ie (5 years ago), I added protocol
version specific records which work to this day;

	ftp.ipv4.heanet.ie
	ftp.ipv6.heanet.ie

We also added clear instructions to that effect;

	http://ftp.ipv4.heanet.ie/

On the rare ocasions people have suffered because of sheer network
brokenness, they've been able to use the version specific records.

Afaik, HEAnet has nothing but positive responses about that approach,
Personally I share the concerns of content providers (I work for one!)
and anyone who did experience suckiness over IPv6 was extremely thankful
when we pointed them at the v4-only record.

Now that I'm working for a major content provider (I'll be requesting PI
from ripe the day I can btw), I can see how it's a very tough
proposition. With implementors like the Debian/Ubuntu folk making such
basic errors of implementation (and they are hardly alone) and the
prevalance of broken home routers with terrifying DNS implementations,
it's just not a production-quality service path.

Nor are content owners going to provide IPv6 only content, I mean what
is the point of that? What on earth is in it for them? 

Where IPv6 is starting to have an edge is in p2p applications, getting
rid of NAT and allowing in-bound connections again solves a lot of
problems for the p2p providers. Microsoft is already taking this
approach and using it for their live-meeting stuff. We'll probably see
some more of this.

I think at this stage it's relatively clear that we are not going to
enjoy a graceful transition to IPv6. Personally I think it will still
happen, we are going to run out of IPv4 addresses soon, there are no
other tennable proposals, and a point will come when IPv4 connectivity
will actually fare worse for enough users than their IPv6 connectivity
that the market will respond. But it will take that pain for it happen.

I think we should all look forward to it, because we've been using IPv6
for longer than anything else and know it sufficient detail that it will
differentiate our services.

-- 
Colm MacCárthaigh                        Public Key: colm+pgp at stdlib.net


More information about the ipv6-ops mailing list