[6bone] separating IPv6 experimental from production traffic

[Ronald van der Pol]

On Wed, Aug 21, 2002 at 10:29:15 -0700, Bill Manning wrote:

> 	Tell me how you propose to do this?

I think we first have to decide what to do: cleanup or separation.

> 	If we want things to work, we need to use them.

Most of my traffic goes over v6.

> 	For myself, I find that the term "production"
> 	is vague at best.  I don't think we (as users
> 	of v6 protocols) can dictate such a change.
> 	A majority of my v6 peering/transit providers
> 	are -UNWILLING- to run infrastructure in dual-stack
> 	mode due to code/feature stability. 

Why should they? If I understand correctly, IIJ is offering
commercial IPv6 service over a separate infrastructure.

I agree "production" is vague. For me it means something like:
recent software, no dozens of tunnels all over the world, no
transit by small sites/ISPs, about the same number of transit
upstreams as you have for v4, monitoring and management
similar to v4.

> So there is
> 	a v4 suite of routers and a separate v6 suite of
> 	routers.  Until vendor code is -MUCH- more stable
> 	and the features are integrated, we will have 
> 	this appearence of a "production" v4 network and
> 	an "experiemental" v6 network, at least as seen 
> 	by a significant portion of the IP community (save JP).

At my previous job at SURFnet in the Netherlands we build a 10 Gbps
dual stack network with GSR routers. As of October last year it's
used daily by the Dutch research and university community. Although
most traffic is v4, the network supports both v4 and v6.  Both v4
and v6 are "production".

> 	If we don't use(and fix) what we create, it will never
> 	be used. 

More and more OSes run IPv6 out of the box. End users start using
IPv6 and notice problems because there are too many routing problems
on the 6bone. So they switch off IPv6 again. We are scaring people
away from IPv6 because of the tunnel mess, old router software
and sites that don't monitor the transit they are giving to everyone.

	rvdp