[6bone] separating IPv6 experimental from production traffic

[John Fraizer]

On Fri, 23 Aug 2002, Pekka Savola wrote:

> > Of course.  But if you have *enough* peerings, you'll be able to reach
> > most networks in a maximum of 2 hops - and if you then apply some 
> > MED fiddling to mark "slow" tunnels, you can achieve pretty good results.
> 
> This leads to an "arms race"; everyone will need to get even more tunnels, 
> leading to even tighter spaghetti.

No.  People who want tighter peering will seek it.  People who don't will
settle for what they have.  High peer counts does NOT mean bad
connectivity.  Peering with some joker on a 56K modem who then leaks a
full table to you (and you accept it) leads to bad connectivity.

> 
> I don't think having to do manual fiddling should be necessary, but that's 
> life I suppose.

I guess you're not the v4 peering person for your network then.  In v4
land, manual tuning is rather commonplace.



---
John Fraizer              | High-Security Datacenter Services |
EnterZone, Inc            | Dedicated circuits 64k - 155M OC3 |
http://www.enterzone.net/ | Virtual, Dedicated, Colocation    |