On Wed, 2005-04-13 at 11:01 +0200, Ronald.vanderPol@rvdp.org wrote: > On Wed, Apr 13, 2005 at 02:44:49 +0800, Fred Baker wrote: > > At that point we have to solve a rendezvous problem. When two of one > > kind of host need to communicate over a network of the other kind, we > > obviously want to tunnel. > > The 6bone ended up in a useless tunnel mess. Keep it simple has always > turned out to be the best approach. I think a dual-stack network is > easier to manage than a tunneled network. When users on an IPv4-only > network want to use IPv6-only services, they should convince the network > managers to upgrade to dual-stack. Otherwise, the tunnel mess will stay > forever, because there is no incentive to upgrade to dual-stack. Of course dual-stack is the way to go, but the biggest reason that people do not upgrade to dual-stack is the lack of support in their hardware. If they where able to dual-stack stuff, they would not even bother with tunneling. Also we are talking about tunneling inside the same AS here, not like the 6bone globally. A packet over a tunnel will thus have 90% the same path it would take when being native. See my traceroutes below, I'll post the diffs to the native path when it becomes native as my ISP indeed moved to their own ATM trunk and now can provide it natively, before they couldn't as the technology didn't allow for it. > Isn't the main reason for tunneling the fact that too few networks are > deploying IPv6? Trying to force IPv6 deployment by creating complicated > tunnel overlays seems to me like wrong engineering. No, the main reason seems to be support and stability. Also, they don't earn money selling IPv6 thus they are not allowed politically to deploy it. Greets, Jeroen -- IPv4: jeroen@purgatory:~$ traceroute www.bit.nl traceroute to http-unix.bip.bit.nl (213.136.12.232), 64 hops max, 40 byte packets 1 213-136-24-1.adsl.bit.nl (213.136.24.1) 12 ms 11 ms 11 ms 2 i49.ge-0-1-0.jun1.kelvin.network.bit.nl (213.136.31.33) 11 ms 22 ms 11 ms 3 http.lb.network.bit.nl (213.136.12.232) 12 ms 11 ms 11 ms IPv6: jeroen@purgatory:~$ traceroute6 www.bit.nl traceroute to http-unix.bip.bit.nl (2001:7b8:3:5::80:1) from 2001:7b8:300:0:290:27ff:fe24:c19f, 30 hops max, 16 byte packets 1 gw-1.ede-01.nl.sixxs.net (2001:7b8:2ff::1) 16.866 ms 13.277 ms 13.863 ms 2 sixxs-gw.ipv6.network.bit.nl (2001:7b8:3:4f:290:6900:4fc6:d81f) 13.697 ms 13.612 ms 19.932 ms 3 2001:7b8:3:5::80:1 (2001:7b8:3:5::80:1) 12.444 ms 19.834 ms 13.277 ms And just in case one is wondering, the tunnel box is found over this path: jeroen@purgatory:~$ traceroute nlede01.sixxs.net traceroute to nlede01.sixxs.net (193.109.122.244), 64 hops max, 40 byte packets 1 213-136-24-1.adsl.bit.nl (213.136.24.1) 12 ms 11 ms 11 ms 2 i49.ge-0-1-0.jun1.kelvin.network.bit.nl (213.136.31.33) 11 ms 11 ms 11 ms 3 nlede01.sixxs.net (193.109.122.244) 13 ms 11 ms 10 ms Which explains the 1ms difference in latency ;)
Attachment:
signature.asc
Description: This is a digitally signed message part