Google and IPv6 inverse?

Hank_Nussbacher1 · June 6, 2011, 5:19am

Will Google have inverse working by June 8th?

[hank@noc ~]$ traceroute6 ipv6.google.com
traceroute to ipv6.l.google.com (2a00:1450:8001::68) from 2001:bf8:0:3:202:b3ff:feaf:f3fc, 30 hops max, 16 byte packets
  1 2001:bf8:0:3::1 (2001:bf8:0:3::1) 0.407 ms 221.144 ms 5.218 ms
  2 2001:bf8:0:b::1 (2001:bf8:0:b::1) 0.559 ms 0.54 ms 0.486 ms
  3 iucc-lb1.rt1.fra.de.geant2.net (2001:798:14:10aa::1d) 50.665 ms 50.611 ms 50.567 ms
  4 google-gw.rt1.fra.de.geant2.net (2001:798:14:10aa::e) 56.821 ms 50.525 ms 50.486 ms
  5 2001:4860::1:0:11 (2001:4860::1:0:11) 51.266 ms 51.106 ms 51.068 ms
  6 2001:4860::1:0:4b3 (2001:4860::1:0:4b3) 58.309 ms 58.078 ms 58.442 ms
  7 2001:4860::8:0:2db0 (2001:4860::8:0:2db0) 57.174 ms 57.339 ms 57.195 ms
  8 2001:4860::2:0:66f (2001:4860::2:0:66f) 72.496 ms 60.803 ms 72.381 ms
  9 2001:4860:0:1::1b (2001:4860:0:1::1b) 68.165 ms 62.21 ms 70.3 ms
10 2a00:1450:8001::68 (2a00:1450:8001::68) 61.21 ms 61.862 ms 61.331 ms

Thanks,
Hank

Christopher_Morrow · June 6, 2011, 12:45pm

Will Google have inverse working by June 8th?

poking the tiger some... 'why?'

Hank_Nussbacher1 · June 6, 2011, 12:53pm

> Will Google have inverse working by June 8th?
>

poking the tiger some... 'why?'

Just curious.

-Hank

STARNES_CURTIS · June 6, 2011, 1:12pm

It works from North Texas.

[cstarnes@tec-mgmt]~ host -6 ipv6.google.com
ipv6.google.com is an alias for ipv6.l.google.com.
ipv6.l.google.com has IPv6 address 2001:4860:800a::6a

[cstarnes@tec-mgmt]~ traceroute ipv6.google.com
traceroute to ipv6.google.com (2001:4860:800a::6a), 30 hops max, 40 byte packets1 6506-sup720.granbury.k12.tx.us (2620:101:3000:111::1) 0.731 ms 0.793 ms 0.872 ms
2 ipv6-rtr.granburyisd.org (2620:101:303f::1) 1.619 ms 1.662 ms 1.611 ms
3 tunnel144.tserv1.fmt.ipv6.he.net (2001:470:1f02:90::1) 53.564 ms 53.147 ms 53.540 ms
4 2001:470:0:1f::1 (2001:470:0:1f::1) 53.055 ms 53.021 ms 52.996 ms
5 10gigabitethernet1-2.core1.sjc2.he.net (2001:470:0:2f::2) 52.979 ms 53.115 ms 53.109 ms
6 2001:470:0:15e::2 (2001:470:0:15e::2) 53.405 ms 2001:4860:1:1:0:1b1b:0:9 (2001:4860:1:1:0:1b1b:0:9) 52.686 ms 2001:470:0:15e::2 (2001:470:0:15e::2) 52.511 ms
7 2001:4860::1:0:7ea (2001:4860::1:0:7ea) 66.009 ms 2001:4860::1:0:21 (2001:4860::1:0:21) 54.049 ms 2001:4860::1:0:7ea (2001:4860::1:0:7ea) 62.659 ms
8 2001:4860::8:0:2cb6 (2001:4860::8:0:2cb6) 85.601 ms 2001:4860::8:0:2cb7 (2001:4860::8:0:2cb7) 54.416 ms 2001:4860::8:0:2cb6 (2001:4860::8:0:2cb6) 53.600 ms
9 2001:4860::1:0:489 (2001:4860::1:0:489) 124.579 ms 2001:4860::1:0:5db (2001:4860::1:0:5db) 112.969 ms 113.006 ms
10 2001:4860::2:0:a7 (2001:4860::2:0:a7) 146.322 ms 147.518 ms 115.466 ms
11 2001:4860:0:1::10b (2001:4860:0:1::10b) 112.987 ms 112.185 ms 113.447 ms
12 yx-in-x6a.1e100.net (2001:4860:800a::6a) 113.500 ms 113.493 ms 113.881 ms

Curtis

Christopher_Morrow · June 6, 2011, 1:55pm

I think this is a case of some ips have this, some don't... I think
all are SUPPOSED to...

-chris

Christopher_Morrow · June 6, 2011, 1:55pm

> Will Google have inverse working by June 8th?
>

poking the tiger some... 'why?'

Just curious.

it occurs to me that I should have also asked: "router hops or end-systems" ?
(in v4 it seems google does end-hosts, but not routed hops)

Leo_Bicknell1 · June 6, 2011, 2:05pm

In a message written on Mon, Jun 06, 2011 at 08:45:31AM -0400, Christopher Morrow wrote:

> Will Google have inverse working by June 8th?

poking the tiger some... 'why?'

It's the network equivilent of holding open the door for someone
or saying please and thank you. Civilized network operators do it
to be polite.

Picking on the name in Hank's e-mail:

% traceroute -n efes.iucc.ac.il
traceroute to efes.iucc.ac.il (128.139.202.17), 64 hops max, 52 byte packets
1 149.20.48.1 1.515 ms 0.783 ms 0.706 ms
2 149.20.65.9 2.756 ms 0.972 ms 3.156 ms
3 64.215.195.21 50.607 ms 50.524 ms 54.705 ms
4 207.138.144.46 166.234 ms 166.235 ms 166.240 ms
5 62.40.125.122 230.584 ms 230.520 ms 237.217 ms
6 128.139.188.1 230.609 ms 443.760 ms 230.629 ms
7 * * *

Quick question, which network providers were involved in that trace?
Have fun hitting up whois to find out!

% traceroute efes.iucc.ac.il
traceroute to efes.iucc.ac.il (128.139.202.17), 64 hops max, 52 byte packets
1 exit.blue.sql1.isc.org (149.20.48.1) 2.410 ms 2.275 ms 0.782 ms
2 int-0-4-0-0.r1.pao1.isc.org (149.20.65.9) 3.493 ms 2.452 ms 0.958 ms
3 ge-9-15-1G.ar1.PAO2.gblx.net (64.215.195.21) 50.488 ms 50.540 ms 54.440 ms
4 DANTE.TenGigabitEthernet7-3.ar1.FRA4.gblx.net (207.138.144.46) 166.227 ms 166.231 ms 166.217 ms
5 iucc-lb1-gw.rt1.fra.de.geant2.net (62.40.125.122) 234.928 ms 245.142 ms 244.414 ms
6 gp1-gp0-te.ilan.net.il (128.139.188.1) 230.536 ms 230.618 ms 230.655 ms
7 *^C

Ah, ISC->Global Crossing->GEANT2.

See, rDNS just saved me about 2 minutes of whois pain. Which is
what being polite is all about, making a very minor effort on your
part to save someone else a minor amount of pain.

Or, if you want another view on it:

Start at slide 12.

Bjorn_Mork · June 6, 2011, 2:38pm

Leo Bicknell <bicknell@ufp.org> writes:

Quick question, which network providers were involved in that trace?
Have fun hitting up whois to find out!

You can convince your traceroute to do that for you:

  -A --as-path-lookups Perform AS path lookups in routing registries and
                              print results directly after the corresponding
                              addresses

traceroute to ipv6.google.com (2a00:1450:4008:c00::93), 30 hops max, 80 byte packets
1 canardo-br0-7.ipv6.mork.no (2001:4620:9:2::1) [AS2119] 1.174 ms 1.349 ms 1.373 ms
2 2001:4600:10:101::1 (2001:4600:10:101::1) [AS2119] 14.623 ms 14.800 ms 14.956 ms
3 * * *
4 * * *
5 * * *
6 2001:4860::1:0:60d (2001:4860::1:0:60d) [AS15169] 60.938 ms 42.619 ms 42.614 ms
7 2001:4860:1:1:0:847:: (2001:4860:1:1:0:847: [AS15169] 27.633 ms 27.732 ms 27.886 ms
8 2001:4860::1:0:26ec (2001:4860::1:0:26ec) [AS15169] 30.820 ms 30.873 ms 34.980 ms
9 2001:4860::1:0:60d (2001:4860::1:0:60d) [AS15169] 54.459 ms 39.116 ms 43.212 ms
10 2001:4860:0:1::217 (2001:4860:0:1::217) [AS15169] 47.002 ms 44.100 ms 44.174 ms
11 2a00:1450:4008:c00::93 (2a00:1450:4008:c00::93) [AS15169] 43.446 ms 43.726 ms 40.642 ms

Now Juniper, it would be really nice if we also could see those
intermediate MPLS nodes at hop 3, 4 and 5, but I know those are AS2119
anyway

Bjørn

Leo_Bicknell1 · June 6, 2011, 2:49pm

In a message written on Mon, Jun 06, 2011 at 04:38:26PM +0200, Bjørn Mork wrote:

You can convince your traceroute to do that for you:

  -A --as-path-lookups Perform AS path lookups in routing registries and
                              print results directly after the corresponding
                              addresses

I have not had good luck with that feature.

Here's a FreeBSD traceroute, using the same host I referenced before:

% traceroute -a efes.iucc.ac.il
traceroute to efes.iucc.ac.il (128.139.202.17), 64 hops max, 52 byte packets
1 [AS1280] exit.blue.sql1.isc.org (149.20.48.1) 4.658 ms 2.718 ms 1.778 ms
2 [AS1280] int-0-4-0-0.r1.pao1.isc.org (149.20.65.9) 3.656 ms 2.363 ms 0.944 ms
3 [AS1221] ge-9-15-1G.ar1.PAO2.gblx.net (64.215.195.21) 50.539 ms 50.508 ms 59.709 ms
4 [AS3549] DANTE.TenGigabitEthernet7-3.ar1.FRA4.gblx.net (207.138.144.46) 166.476 ms 166.240 ms 166.243 ms
5 [AS20965] iucc-lb1-gw.rt1.fra.de.geant2.net (62.40.125.122) 230.590 ms 230.587 ms 230.545 ms
6 [AS378] gp1-gp0-te.ilan.net.il (128.139.188.1) 230.546 ms 230.579 ms 230.528 ms
7 * * *

Now, I happen to administer hop #2, and know the packets are leaving
on a link to Global Crossing (glbx.net). How AS 1221, which is
Telstra, ends up in there is beyond me.

From a Juniper box with a full table:

bicknell@re0.r7.pao1> traceroute as-number-lookup efes.iucc.ac.il
traceroute to efes.iucc.ac.il (128.139.202.17), 30 hops max, 40 byte packets
1 int-2-0-0.r1.pao1.isc.org (149.20.65.16) 4.448 ms 3.973 ms 3.222 ms
2 ge-9-15-1G.ar1.PAO2.gblx.net (64.215.195.21) 50.724 ms 67.135 ms 50.761 ms
3 DANTE.TenGigabitEthernet7-3.ar1.FRA4.gblx.net (207.138.144.46) [AS 3549] 159.652 ms 159.571 ms 160.003 ms
4 iucc-lb1-gw.rt1.fra.de.geant2.net (62.40.125.122) [AS 20965] 279.761 ms 275.059 ms 230.929 ms
5 gp1-gp0-te.ilan.net.il (128.139.188.1) [AS 378] 231.149 ms 231.222 ms 231.258 ms
6 * * *

At least it's not wrong, but there is no ASN listed because the /30
on the link between 2 and 3 is internal, so it views hop 3 as being
internal to my ASN, when it is not.

But more importantly, not all traceroutes have this feature, and as I
said it's about being polite.

Valdis_Kletnieks · June 6, 2011, 3:08pm

s/-a/-A/ - FTFY.

Leo_Bicknell1 · June 6, 2011, 3:26pm

-a Turn on AS# lookups for each hop encountered.

-A Turn on AS# lookups and use the given server instead of the
default.

Is there a particular server you would like to recommend?

Ryan_Rawdon · June 6, 2011, 4:38pm

In some regions Google has revere DNS for IPv6 already:

Ryan-Rawdons-Work-MacBook-Pro:~ ryan$ traceroute6 ipv6.google.com
traceroute6 to ipv6.l.google.com (2001:4860:800f::67) from [redacted], 64 hops max, 12 byte packets
  1 [redacted]
  2 2001:4870:4000:5::1 7.953 ms 8.540 ms 9.697 ms
  3 2001:506:8:ffff::3 24.895 ms 26.349 ms 24.993 ms
  4 2001:4860::1:0:5db 25.654 ms 32.175 ms 26.444 ms
  5 2001:4860::1:0:9ff 112.931 ms 60.227 ms 56.638 ms
  6 2001:4860:0:1::14f 37.783 ms 24.640 ms 34.077 ms
  7 iad04s01-in-x67.1e100.net 25.574 ms 24.359 ms 24.719 ms
Ryan-Rawdons-Work-MacBook-Pro:~ ryan$

They're doing it at about the same level as their IPv4 reverse DNS, that is, servers seem to have reverse DNS but intermediate hops in their infrastructure do not:

Ryan-Rawdons-Work-MacBook-Pro:~ ryan$ traceroute mail.google.com
traceroute: Warning: mail.google.com has multiple addresses; using 72.14.204.83
traceroute to googlemail.l.google.com (72.14.204.83), 64 hops max, 52 byte packets
  1 [redacted]
  2 [redacted]
  3 64-129-129-49.static.twtelecom.net (64.129.129.49) 7.057 ms 7.956 ms 7.439 ms
  4 dist-02-xe-0-0-0.milw.twtelecom.net (66.192.246.82) 7.082 ms 6.975 ms 8.139 ms
  5 216.239.48.108 (216.239.48.108) 9.187 ms 32.129 ms 107.103 ms
  6 66.249.94.46 (66.249.94.46) 10.923 ms
     66.249.94.54 (66.249.94.54) 8.839 ms 12.205 ms
  7 iad04s01-in-f83.1e100.net (72.14.204.83) 8.592 ms 9.672 ms 9.032 ms
Ryan-Rawdons-Work-MacBook-Pro:~ ryan$

Heinrich_Strauss · June 6, 2011, 7:08pm

From what I understand, they filter intermediate hops so that only Google internal machines see the paths. Destination hops should have reverses, though.