Strange routing over last couple of days

[alexatkin]

Not so great.

Code:

1  Router (192.168.1.254)  0.244 ms  0.305 ms  0.383 ms
2  145.255.242.254 (145.255.242.254)  4.266 ms  4.477 ms  4.541 ms
3  185.24.14.1 (185.24.14.1)  7.744 ms  7.848 ms  7.921 ms
4  * * *
5  185.103.229.237 (185.103.229.237)  16.989 ms  17.160 ms  17.238 ms
6  ge-0-0-1.sth.swe.call27.net (185.103.229.251)  43.431 ms  43.179 ms  44.005 ms
7  192.121.80.47 (192.121.80.47)  47.250 ms  44.131 ms  44.245 ms
8  216.239.54.181 (216.239.54.181)  47.851 ms  48.106 ms  48.590 ms
9  209.85.245.63 (209.85.245.63)  45.094 ms 72.14.234.83 (72.14.234.83)  53.314 ms  53.389 ms
10  google-public-dns-a.google.com (8.8.8.8)  46.910 ms  46.821 ms  47.294 ms

I'm chasing this up with them as a change from 14ms to 47ms is not something you expect and it makes a noticeable difference to DNS resolution speed.

If that domain name and reading Call27.nets twitter feed, it looks like its routing via their Sweden POP to get to London, utterly bonkers.

Its even worse for my IPv6 tunnel which seems to be routing via Sweden to Amsterdam to get to London.

Code:

traceroute to 216.66.88.98 (216.66.88.98), 30 hops max, 60 byte packets
1  Router (192.168.1.254)  0.179 ms  0.202 ms  0.242 ms
2  145.255.242.254 (145.255.242.254)  4.261 ms  4.293 ms  4.450 ms
3  185.24.14.1 (185.24.14.1)  6.173 ms  6.401 ms  6.439 ms
4  * * *
5  185.103.229.237 (185.103.229.237)  14.419 ms  14.672 ms  14.675 ms
6  ge-0-0-1.sth.swe.call27.net (185.103.229.251)  41.670 ms  41.465 ms  41.445 ms
7  * * *
8  10ge12-6.core1.ams1.he.net (72.52.92.45)  43.589 ms  43.541 ms  43.787 ms
9  100ge9-1.core1.lon2.he.net (72.52.92.213)  41.321 ms  70.150 ms  70.395 ms
10  tserv1.lon2.he.net (216.66.88.98)  46.666 ms  41.609 ms  47.781 ms

I have checked a few different routes, things seem to be fine for Microsoft and my VPS.
The irony there is both of those will default to IPv6 so in real-world use are still being hit due my IPv6 tunnel provider being an affected route. DOH!

[alexatkin]

Wow. I contacted Call27 themselves about the issue via Facebook and this is what I got back:

Code:

Call27 Hi Alex,

Dean here from network. I've made some alterations and dropped HE from direct peering in Stockholm, I've also built a route-map for London to depreference AS15169 (Google Inc) from Stockholm.

cr-1.hex67.uk#traceroute 216.66.88.98

Type escape sequence to abort.
Tracing the route to tserv1.lon2.he.net (216.66.88.98)

1 xe-0-0-1-101.luigi.as34270.net (85.91.226.145) [AS 34270] 0 msec 0 msec 0 msec
2 xe-0-0-1-0.mario.as34270.net (85.91.224.9) [AS 34270] 0 msec 0 msec 0 msec
3 40ge1-3.core1.lon2.he.net (195.66.224.21) 0 msec 0 msec 0 msec
4 tserv1.lon2.he.net (216.66.88.98) [AS 6939] 4 msec 0 msec 0 msec

cr-1.hex67.uk#traceroute 8.8.8.8

Type escape sequence to abort.
Tracing the route to google-public-dns-a.google.com (8.8.8.8)

1 xe-0-0-1-101.luigi.as34270.net (85.91.226.145) [AS 34270] 0 msec 0 msec 0 msec
2 xe-0-0-1-0.mario.as34270.net (85.91.224.9) [AS 34270] 0 msec 4 msec 0 msec
3 195.66.224.125 0 msec 0 msec 0 msec
4 72.14.232.187 [AS 15169] 0 msec 4 msec
72.14.232.215 [AS 15169] 0 msec
5 216.239.58.77 [AS 15169] 0 msec
216.239.57.67 [AS 15169] 4 msec
216.239.58.89 [AS 15169] 0 msec
6 google-public-dns-a.google.com (8.8.8.8) [AS 15169] 4 msec 0 msec 4 msec

I'm seeing 1ms to Google DNS and 1ms to the tunnel server for HE now.

Origin could really learn a thing or two about communication from these guys. Its telling that I was able to resolve the problem quicker than they could. :?

I mean sure the problem should not have occured in the first place and who knows if other peering are also routed wrong, but the quick and detailed resolution is worlds apart from the communication from Origin.

I got a call from Origin this morning to tell me my offer is going to run out in January, so I'm finally looking to move to Zen again (native IPv6 FTW).

I just can't bring myself to stay with an ISP that has such poor communication skills and have openly admitted that their network MTU is still lower than it should it be, a legacy of Digital Region, and no immediate plans for IPv6.

As it is, I'm not holding my breath that I will get an activation before January (I have zero faith in Openreach) so they may end up leaving me with no connection again. Pray to Openreach that my cabinet is not at capacity.

[alexatkin]

Zen connection enabled and now I'm getting a stable 74Mbit down, something Origin has never done since the day they connected me to the BT network.

[alexatkin]

Also 1-3ms lower latency, it varies based on the PPP session as presumably Zen have different virtual circuits from Openreach.

Quote:Tracing route to google-public-dns-a.google.com [8.8.8.8]
over a maximum of 30 hops:

1 13 ms 12 ms 12 ms losubs.subs.dsl6.th-lon.zen.net.uk [62.3.84.27]
2 14 ms 12 ms 12 ms ae1-133.cr2.th-lon.zen.net.uk [62.3.86.133]
3 13 ms 12 ms 13 ms ge-2-0-0-0.cr1.th-lon.zen.net.uk [62.3.80.41]
4 13 ms 12 ms 13 ms 72.14.217.190
5 13 ms 12 ms 12 ms 64.233.174.87
6 130 ms 14 ms 14 ms 216.239.57.51
7 13 ms 14 ms 13 ms google-public-dns-a.google.com [8.8.8.8]

Quote:Tracing route to google-public-dns-a.google.com [8.8.8.8]
over a maximum of 30 hops:

1 5 ms 4 ms 4 ms 145.255.242.254
2 6 ms 6 ms 5 ms 185.24.14.1
3 * * * Request timed out.
4 15 ms 14 ms 14 ms xe-0-3-4.cr-1.hex67.uk.call27.net [185.103.229.237]
5 14 ms 14 ms 14 ms xe-0-0-1-101.luigi.as34270.net [85.91.226.145]
6 14 ms 14 ms 14 ms xe-0-0-1-0.mario.as34270.net [85.91.224.9]
7 14 ms 14 ms 15 ms 195.66.224.125
8 15 ms 15 ms 14 ms 72.14.232.187
9 16 ms 15 ms 15 ms 216.239.57.53
10 15 ms 15 ms 15 ms google-public-dns-a.google.com [8.8.8.8]