[6bone] separating IPv6 experimental from production traffic

[Bill Manning]

% Bill Manning <bmanning@ISI.EDU> writes:
% 
% > 	One of the promises of v6 was that -AS A TRANSITION-
% > 	folks could run dual stack.  Not to pick, but
% > 	the router vendors I've looked at do not have 
% > 	similar code quality or feature sets between v4 and v6.
% 
% Juniper does have very high code quality for IPv6. In fact we haven't
% experienced any operational problem with running it in parallel to
% IPv4, in contrast to The Router Vendor on dedicated routers.

	If only they would have told me about it before I made the
	last purchase of hardware.  They refused to even ack. v6
	support until after I had taken delivery of ~USD500k
	of routing gear.  Two weeks later they announced v6 support.
	No Juniper in our shop for another 24 months.  Thats a
	-LONG- time for code to mature.

% Non-availability of high-quality IPv6 connectivity in US is not a good
% reason to mess up European and Asian progress in this area.

	No reason to.  Route as you see fit.  Just don't dictate
	routing policy to me unless/until we directly peer and
	then we can have a discussion.


% > 	Same was true with inital IPv4 deployment.
% 
% You are really implying that today's user community has similar
% requirements to that at those times?

	Nope. I'm saying that code stability/interoperability
	was as bad or worse than in todays environment and circuit
	stability was much worse.  However the clue density of
	network users was -MUCH- higher.

% Today you cannot tell the users "use IPv6, and by the way, don't worry
% if you don't have connectivity, it will propably come back in a couple
% of days".

	Sure you can.  What is your business model?

% Robert


--bill