Weird Google Compute Engine routing within US

Question

We are seeing some very weird routing/latency issue for GCE within the US:

$ traceroute fixdemo.fxcorporate.com
traceroute to fixdemo.fxcorporate.com (204.8.241.23), 30 hops max, 60 byte packets
 1  216.239.43.216 (216.239.43.216)  0.824 ms 216.239.46.192 (216.239.46.192)  0.822 ms 72.14.233.134 (72.14.233.134)  1.296 ms
 2  216.239.43.216 (216.239.43.216)  1.274 ms 216.239.46.192 (216.239.46.192)  0.976 ms 64.233.174.94 (64.233.174.94)  1.272 ms
 3  216.239.49.32 (216.239.49.32)  1.271 ms 64.233.174.94 (64.233.174.94)  1.261 ms  1.260 ms
 4  216.239.49.24 (216.239.49.24)  1.244 ms 216.239.46.190 (216.239.46.190)  2.066 ms 216.239.43.216 (216.239.43.216)  1.243 ms
 5  72.14.233.134 (72.14.233.134)  1.236 ms  1.235 ms 216.239.46.192 (216.239.46.192)  1.216 ms
 6  216.239.46.192 (216.239.46.192)  1.212 ms 216.239.43.218 (216.239.43.218)  0.788 ms 64.233.174.94 (64.233.174.94)  0.824 ms
 7  216.239.49.24 (216.239.49.24)  0.829 ms 216.239.43.216 (216.239.43.216)  0.781 ms 216.239.49.32 (216.239.49.32)  0.838 ms
 8  216.239.43.216 (216.239.43.216)  0.837 ms 64.233.174.94 (64.233.174.94)  1.045 ms 216.239.43.216 (216.239.43.216)  0.848 ms
 9  72.14.238.107 (72.14.238.107)  15.354 ms  15.389 ms 209.85.241.23 (209.85.241.23)  13.589 ms
10  72.14.237.131 (72.14.237.131)  13.596 ms  13.599 ms  13.587 ms
11  209.85.254.121 (209.85.254.121)  13.594 ms  11.643 ms  13.533 ms
12  * * *
13  ae4.cr1.ord2.us.above.net (64.125.28.49)  14.256 ms  13.891 ms  11.877 ms
14  ae0.cr2.ord2.us.above.net (64.125.20.246)  13.883 ms  13.975 ms  12.056 ms
15  ae6.cr2.lga5.us.above.net (64.125.24.29)  38.438 ms  40.585 ms  40.586 ms
16  ae3.er4.lga5.us.above.net (64.125.31.246)  42.630 ms  42.603 ms  41.923 ms
17  208.185.12.62.IPYX-065351-ZYO.above.net (208.185.12.62)  215.528 ms  216.819 ms  216.719 ms
18  fixdemo.fxcorporate.com (204.8.241.23)  216.402 ms  216.298 ms  216.474 ms

Source is located in us-central1-a, destination in New York. Still, we see a latency via traceroute and mtr of > 200 ms, which starts within the above.net network. The same problem exists for a source in the europe-west1-a zone. I guess this is the result of some heavily asymmetric routing starting at hop #17.

Any idea how to solve this?

Answer 1

you could be right, as you know it is not simple to troubleshoot routing problems, and not much can be done on the internet user side.

I would like to share this well done document about troubleshooting with Traceroute:

https://www.nanog.org/meetings/nanog45/presentations/Sunday/RAS_traceroute_N45.pdf

Regards Paolo