Re: [ipv6-wg] V6 connectivity

[Bernhard Schmidt berni@localhost]

Max Tulyev wrote:

Hi,

> I see v6 connectivity is much worse than v4:
>
> traceroute to kite-www.ripe.net (2001:610:240:0:a843::8) from 
> 2a01:d8:0:99:211:11ff:fe1f:bfc9, 30 hops max, 16 byte packets
>  1  2a01:d8:0:99::1 (2a01:d8:0:99::1)  0.782 ms  0.719 ms  2.227 ms
>  2  2a01:d8:0:1::1 (2a01:d8:0:1::1)  40.904 ms  2.346 ms  2.754 ms
>  3  2001:b08:0:2::1 (2001:b08:0:2::1)  35.849 ms  7.533 ms  2.576 ms
>  4  2001:b08:b08:3::2 (2001:b08:b08:3::2)  35.205 ms  25.399 ms  25.649 ms
>  5  se-kth.nordu.net (2001:948:0:f02c::1)  29.649 ms  25.579 ms  25.545 ms
>  6  se-ov.nordu.net (2001:948:0:f027::2)  33.705 ms  25.539 ms  25.927 ms
>  7  dk-gw2.nordu.net (2001:948:0:f03f::2)  39.785 ms  33.901 ms  33.688 ms
>  8  * * *
>  9  * * *
> 10  * * *
> 11  * * *
> 12  nikipv6.ripe.net (2001:7f8:1::a500:3333:1)  214.93 ms  226.989 ms 
> 215.676 ms
> 13  kite-www.ripe.net (2001:610:240:0:a843::8)  214.484 ms  215.588 ms 
> 221.891 ms
I don't know whether 2a01:d8::/32 is supposed to be routed through GEANT 
(going through nordu.net suggests so), but currently I see it only 
through Sprint. The Sprint upstream of Nordunet has been a great joy for 
several occasions now (I have an IPv6-in-IPv6 tunnel to UNINETT for 
Multicast which is routed back to me via Sprint).
 2:  rt4.muc1.m-online.net                     46.408ms
 3:  2001:a60:0:101:89:1:1:9                  asymm  4  48.232ms
 4:  ge-1-0-0-132.rt7.muc3.m-online.net       asymm  5  80.360ms
 5:  ge-0-1-0-302-gar1.muc.cw.net             asymm  6  47.831ms
 6:  so-4-3-0-dcr2.fra.cw.net                 asymm  7  57.416ms
 7:  so-4-0-0-dcr1.amd.cw.net                 asymm  8  71.665ms
 8:  so-3-0-0-zcr1.amt.cw.net                 asymm  9  70.195ms
 9:  nl-ams04a-re1-fe-0-0.ipv6.aorta.net       72.908ms
10:  be-bru01a-re1-t-2.ipv6.aorta.net          77.627ms
11:  uk-lon01a-re1-t-2.ipv6.aorta.net         asymm 12  83.444ms
12:  us-nyc02a-re2-t-0.ipv6.aorta.net         asymm 13 151.593ms
13:  us-nyc02a-re1-fe-0-0.ipv6.aorta.net      asymm 14 154.725ms
14:  sprarpa-gw2.us.ipv6.aorta.net            asymm 15 169.199ms
15:  sl-s1v6-nyc-t-1000.sprintv6.net          asymm 16 184.698ms
16:  sl-bb1v6-sto-t-101.sprintv6.net          asymm 17 281.296ms
17:  2001:7f8:d:fb::24                        221. 74ms
18:  se-kth.nordu.net                         219.620ms
19:  SE-gw.RUN.net                            asymm 18 223.267ms
20:  2001:b08:b08:3::1                        asymm 19 239.308ms
21:  2001:b08:0:2::2                          asymm 20 244.649ms
22:  2a01:d8:0:1::4                           asymm 21 246. 81ms
23:  2a01:d8:0:99:211:11ff:fe1f:bfc9          asymm 22 246.726ms reached
     Resume: pmtu 1280 hops 23 back 22

I suggest you kick Nordunet/GEANT to transit this prefix (if it is 
academic) or try to get a better upstream.
> And some parts of v6 Internet is completly unreachable. Is it my bug
> or real situation?

It is a real problem, but not a common one in IPv6.


Regards,
Bernhard

>