And now, for your regularly scheduled dose of broken ipv6 routing...

Sun Sep 24 19:51:11 CEST 2006

Iljitsch van Beijnum wrote:
> On 22-sep-2006, at 10:16, Carlos Friacas wrote:

>> For what i've seen so far, the measurements i get with a smaller
>> difference than 1ms between v4 & v6 are from other NRENs (to be more
>> accurate... 13 of them to the date). Could be better, could be worse...
[..]
> I was going to show you that from my system to www.fccn.pt IPv6 is worse
> than IPv4, but I don't get any ping replies over IPv4 so never mind...

It's always funny to hear folks complain who are in academic networks.
See my traceroute to www.fccn.pt, going through level3 -> to GEANT
Italy, Geneva, mad.ch (dunno which that one is), an unnamed hop and then
finally to portugal. Nice detour we have there ;)

Fortunately there are also NREN's which actually have some decent
connectivity, see the second trace to SWITCH, which only takes 21ms to
get in Zurich (v4 is about 20ms ;), while it takes the former 100ms+ to
get in Italy. One would almost think that l3 is routing over the US ;)
But one can't tell as the reverses don't exist.

The v4 path to www.fccn.pt is only 45ms to their first host which drops
traces, while it is 150+ for the v6 one, ouch!

ISC pings at 150ms v4 versus 171ms v6 btw, note that that is almost as
good as riding GEANT to Portugal from the Netherlands ;)

To not make this a complete rant against NREN's/GEANT/fccn (note that I
said nice things about SWITCH :), I would like to propose that *any*
location who has significant differences in latency between IPv4 and IPv6
 a) sign up to GRH, this so we can can help them out
    in seeing where the bgp tables for them are not
    favourable for them.
 b) maybe install a RIPE TTM box.
    This will give them all the immeasurably superior(*)
    they will need to fix up their networks. Politics
    seem to be a huge problem in that area too though,
    but as operators, instead of screaming and yelling
    most can't help out there unfortunately.
    TTM already has a number of NREN players but also quite a lot
    of commercial ones. And we do want to be able to talk between
    those two don't we? Or don't we need google.com as .edu or don't
    we need the very large ftp.heanet.ie archives as .com? :)

But that is just my proposal of an end user of the various networks
around this nice and bright planet, oh no not anymore, the sun just set
behind the mountains ;)

Greets,
 Jeroen

* This is a term taken from http://www.he.net/colo_ipv6.htm "Currently,
our IPv6 routing is immeasurably superior to other existing companies.
Our routing table has more prefixes (routes) and more paths to each
prefix (ways to get to a destination address block) than most other IPv6
providers."

--
Le traces:

$ traceroute6 www.fccn.pt
traceroute to www.fccn.pt (2001:690:a00:40aa:2c0:9fff:fe20:e261) from
2001:838:1:1:210:dcff:fe20:7c7c, 30 hops max, 16 byte packets
 1  fe0.breda.ipv6.concepts-ict.net (2001:838:1:1::1)  0.408 ms  0.458
ms  0.671 ms
 2  se2.ams-ix.ipv6.concepts-ict.net (2001:838:0:10::1)  2.006 ms  1.958
ms  1.956 ms
 3  nl-ams04a-re1-fe-0-0.ipv6.aorta.net (2001:7f8:1::a500:6830:1)  2.635
ms  2.451 ms  2.581 ms
 4  2001:1900:5:3::1d (2001:1900:5:3::1d)  85.499 ms  85.542 ms  85.396 ms
 5  2001:1900:5:2::6 (2001:1900:5:2::6)  103.504 ms  103.4 ms  117.002 ms
 6  so-3-0-0.rt1.mil.it.geant2.net (2001:798:cc:1001:1e01::2)  115.895
ms  115.97 ms  116.035 ms
 7  so-6-3-0.rt1.gen.ch.geant2.net (2001:798:cc:1201:1e01::1)  123.046
ms  122.98 ms  122.946 ms
 8  so-7-0-0.rt1.mad.ch.geant2.net (2001:798:cc:1201:1701::2)  145.389
ms  145.352 ms  145.259 ms
 9  2001:798:cc:1701:2401::2 (2001:798:cc:1701:2401::2)  157.925 ms
159.905 ms  157.728 ms
10  fccn-gw.pt1.pt.geant.net (2001:798:2024:10aa::2)  157.981 ms  157.92
ms  157.9 ms
11  2001:690:800:1::7 (2001:690:800:1::7)  158.216 ms  158.059 ms
158.067 ms
12  2001:690:800:1::7 (2001:690:800:1::7)  157.703 ms !S  158.3 ms !S
157.819 ms !S

$ traceroute www.fccn.pt
traceroute to www.fccn.pt (193.136.2.218), 64 hops max, 40 byte packets
 1  ge-1-3-0.colo.breda.concepts-ict.net (213.197.29.1)  0 ms  0 ms  0 ms
 2  at-0-3-1.nikhef.concepts-ict.net (213.197.27.126)  2 ms  2 ms  2 ms
 3  Gi0-0-203.ams-koo-access-2.interoute.net (212.23.59.161)  3 ms  3 ms
 3 ms
 4  PO6-1.ams-koo-core-1.interoute.net (212.23.41.137) [MPLS: Label 58
Exp 0]  110 ms  95 ms  34 ms
 5  PO2-0.fra-006-core-2.interoute.net (84.233.190.22) [MPLS: Label 48
Exp 0]  10 ms  10 ms  10 ms
 6  PO6-0.fra-012-inter-1.interoute.net (212.23.42.166)  10 ms  10 ms  9 ms
 7  ffm-b3-geth1-2-1-102.telia.net (213.248.103.53)  10 ms  10 ms  10 ms
 8  ffm-bb2-link.telia.net (80.91.249.142)  10 ms  10 ms  10 ms
 9  prs-bb2-pos7-0-0.telia.net (213.248.65.117)  20 ms  20 ms  21 ms
10  ldn-bb2-pos7-0-0.telia.net (213.248.65.113)  18 ms  19 ms  18 ms
11  ldn-b2-link.telia.net (80.91.249.189)  18 ms  18 ms  18 ms
12  dante-01233-ldn-b2.c.telia.net (213.248.75.110)  18 ms  19 ms  20 ms
13  so-5-0-0.rt1.lis.pt.geant2.net (62.40.112.145)  45 ms (TOS=128!)  45
ms  45 ms
14  fccn-gw.pt1.pt.geant.net (62.40.103.178)  46 ms  45 ms  45 ms
15  Router3.GE.Lisboa.fccn.pt (193.137.0.10)  45 ms  45 ms  45 ms
16  * *

$ traceroute6 www.switch.ch
traceroute to oreius.switch.ch (2001:620:0:1b::b) from
2001:838:1:1:210:dcff:fe20:7c7c, 30 hops max, 16 byte packets
 1  fe0.breda.ipv6.concepts-ict.net (2001:838:1:1::1)  0.401 ms  0.361
ms  0.343 ms
 2  se2.ams-ix.ipv6.concepts-ict.net (2001:838:0:10::1)  2.021 ms  2.167
ms  1.949 ms
 3  eth10-0-0.xr1.ams1.gblx.net (2001:7f8:1::a500:3549:1)  2.91 ms
2.543 ms  2.768 ms
 4  2001:450:2001:1000:0:670:1708:128
(2001:450:2001:1000:0:670:1708:128)  11.512 ms  11.732 ms  11.764 ms
 5  2001:450:2002:10::2 (2001:450:2002:10::2)  22.106 ms  21.051 ms
21.453 ms
 6  swiEL2-10GE-1-3.switch.ch (2001:620:0:c06a::2)  21.637 ms  21.68 ms
 21.326 ms
 7  swiCP2-10GE-2-0-1.switch.ch (2001:620:0:c049::1)  21.774 ms  21.675
ms  21.582 ms
 8  oreius.switch.ch (2001:620:0:1b::b)  21.292 ms  21.942 ms  21.344 ms

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 311 bytes
Desc: OpenPGP digital signature
URL: <https://lists.cluenet.de/pipermail/ipv6-ops/attachments/20060924/95fa8144/attachment.sig>