[AusNOG] My Predictions for the ISP Industry

Mark Newton newton at atdot.dotat.org
Fri Mar 16 14:42:25 EST 2012


On Fri, Mar 16, 2012 at 02:23:05PM +1100, Paul Brooks wrote:

 > > Yeah, but a counterpoint is that 6to4 often doesn't work.
 > > Check Geoff's presso at APNIC31 for the lowdown.
 > 
 > Yeah, but a counterpoint is that 6to4 works better than nothing at
 > all, which is what most users have while their providers have
 > native IPv6 disabled

I'm not totally sure I agree with that.

In an environment with no IPv6 at all, "the internet" works fine.

In an environment with dual-stack, "the internet" also works fine.

In an environment with working 6to4, "the internet" also works fine.

In an environment with broken 6to4, "the internet" comes with
random failures which end users will find almost impossible to 
diagnose (and, depending on the failure mode, their helpdesk
support staff will find it almost impossible to duplicate their
bugs!)

So I reckon no IPv6 at all is actually superior to IPv6 delivered
over busted 6to4, if your expectation is that the internet should
work.

Moreover, for any service provider which isn't currently providing
dual-stack access, random autotunnel breakage is to be expected 
/right now/.  The only effective way to improve service to your
end users is to (ta-da!) deploy dual-stack native IPv6, because
you have no useful way of controlling whether or not their auto-
tunnelling works, so your perceived service quality is at the
mercy of stuff that fails all the time, until you provide your
users with an environment that doesn't need auto-tunneling.

Remember: For any given failure, most users will blame their ISP
first. They don't know what 6to4 gateways are;  they probably
don't even know that they're using IPv6 in the first place.  They
just know that whatever they're trying to do sucks, but when
their mates at some other ISP do the same thing it works fine.

  - mark




More information about the AusNOG mailing list