[6bone] separating IPv6 experimental from production traffic

[Pekka Savola]

On Wed, 21 Aug 2002, Ronald van der Pol wrote:
> Comments?

These have been my worries, as I've been experiencing the same problems 
for some time now.

I've been trying to advocate two things:

 1) "don't do transit" -- yeah it's cool to give transit to everyone, but
you really shouldn't on your rotten Cisco 2600 or 4500 behind a 2 Mbit/s
line!

 2) "not everybody needs a pTLA" -- if we give pTLA to everyone that asks, 
the number of folks that are (practically) able to do 1) increases.  Also, 
this will lead to people filtering whole 6bone space if we want to enforce 
strict aggregation.

The most important thing for stability is starting to unentangle the
tunnel mess, or at least preventing it from getting worse.  Major _real_
transit providers or major ISP's could be in key position here (I'm not
sure how much they're currently), by providing either tunneled or native
v6 transits (traffic which would mostly run through their lines anyway).

But it's a huge task..

-- 
Pekka Savola                 "Tell me of difficulties surmounted,
Netcore Oy                   not those you stumble over and fall"
Systems. Networks. Security.  -- Robert Jordan: A Crown of Swords