Re: Automatic tunnels

[Tim Chown <tjc@ecs.soton.ac.uk>]

On Tue, Jul 29, 2003 at 12:24:26PM +0200, Gert Doering wrote:
> 
> Operational errors (like "wrong IP address configured on interface" or
> "routing protocol configuration things" or "correct v4 address in DNS but
> v6 address wrong" or so) hit in quite surprising ways - you test, you 
> run traceroute and ping, all looks fine, and then you realize that you 
> traceroute'd only v4 while v6 is broken, or vice versa.
> 
> Of course all this is solveable, but it means "more room for errors",
> and usually these do happen.
> 
> (We're in the process of converting the v4-only network to dual-stack,
> and it's quite a challenge)

There are many networks that have gone dual stack, so what you're saying
is we need more operational experience feedback of the problems, rather
than just the success stories.   The early adoptors are the US and European
research networks and backbones.   I think they've been instrumental in
helping Cisco, Juniper et al iron out many of the problems to make 
deployment less problematic (but by no means straightforward) for ISPs.

Tim