2

I have a ubuntu machine (Ubuntu 16.04.1 LTS) connected to my home router via ethernet. I had been noticing poor performance when connecting via ssh, so I ran some ping tests from the ubuntu machine. It starts high, but decreases at a strangely consistent pace. Here are the results:

$ ping -c 30 www.google.com
PING www.google.com (216.58.217.164) 56(84) bytes of data.
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=1 ttl=56 time=99.7 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=2 ttl=56 time=98.1 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=3 ttl=56 time=97.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=4 ttl=56 time=96.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=5 ttl=56 time=95.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=6 ttl=56 time=94.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=7 ttl=56 time=93.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=8 ttl=56 time=92.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=9 ttl=56 time=91.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=10 ttl=56 time=89.9 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=11 ttl=56 time=88.9 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=12 ttl=56 time=87.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=13 ttl=56 time=86.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=14 ttl=56 time=85.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=15 ttl=56 time=84.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=16 ttl=56 time=83.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=17 ttl=56 time=82.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=18 ttl=56 time=81.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=19 ttl=56 time=80.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=20 ttl=56 time=79.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=21 ttl=56 time=78.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=22 ttl=56 time=77.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=23 ttl=56 time=76.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=24 ttl=56 time=75.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=25 ttl=56 time=74.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=26 ttl=56 time=73.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=27 ttl=56 time=72.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=28 ttl=56 time=71.0 ms
64 bytes from iad23s44-in-f164.1e100.net (216.58.217.164): icmp_seq=30 ttl=56 time=61.9 ms

and another website to test

$ ping -c 30 www.apple.com 
PING e6858.dsce9.akamaiedge.net (23.218.213.175) 56(84) bytes of data.
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=1 ttl=59 time=99.8 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=2 ttl=59 time=98.1 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=3 ttl=59 time=97.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=4 ttl=59 time=96.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=5 ttl=59 time=95.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=6 ttl=59 time=94.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=7 ttl=59 time=93.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=8 ttl=59 time=92.2 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=9 ttl=59 time=91.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=10 ttl=59 time=90.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=11 ttl=59 time=88.9 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=12 ttl=59 time=88.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=13 ttl=59 time=87.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=14 ttl=59 time=86.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=15 ttl=59 time=85.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=16 ttl=59 time=84.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=17 ttl=59 time=83.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=18 ttl=59 time=82.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=19 ttl=59 time=81.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=20 ttl=59 time=80.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=21 ttl=59 time=79.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=22 ttl=59 time=78.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=23 ttl=59 time=77.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=24 ttl=59 time=75.9 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=25 ttl=59 time=74.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=26 ttl=59 time=73.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=27 ttl=59 time=72.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=28 ttl=59 time=71.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=29 ttl=59 time=70.0 ms
64 bytes from a23-218-213-175.deploy.static.akamaitechnologies.com (23.218.213.175): icmp_seq=30 ttl=59 time=69.0 ms

Anybody know what might cause this behavior?

Update:

Pinging the router results in the same thing, so the problem must be something on the local network. Heres the report from mtr anyway:

                                               Loss%   Snt   Last   Avg  Best Wrst StDev
1.|-- 192.168.1.1                               0.0%    10   74.6  81.9  74.6 100.0   9.1
2.|-- lo0-100.WASHDC-VFTTP-311.verizon-gni.net  0.0%    10   74.6  72.7   8.9  97.4  23.3
3.|-- B3311.WASHDC-LCR-22.verizon-gni.net       0.0%    10   74.5  73.6  17.9  97.3  20.6
4.|-- ???                                      100.0    10    0.0   0.0   0.0   0.0   0.0
5.|-- ???                                      100.0    10    0.0   0.0   0.0   0.0   0.0
6.|-- 0.et-10-3-0.GW13.IAD8.ALTER.NET           0.0%    10   74.2  76.0  44.8  97.0  12.7
7.|-- 204.148.79.46                            10.0%    10   74.1  76.7  53.8  96.9  11.0
8.|-- 209.85.252.121                            0.0%    10   74.1  77.6  62.9  96.8   8.3
9.|-- 209.85.143.145                           10.0%    10   74.0  78.4  71.9  96.8   7.3
10.|-- iad23s44-in-f164.1e100.net               10.0%    10   73.9  79.2  73.9  96.7   6.9

Not sure what 4 and 5 are.

$tc qdisc
qdisc noqueue 0: dev lo root refcnt 2 
qdisc pfifo_fast 0: dev enp9s0 root refcnt 2 bands 3 priomap  1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
  • Wow. You've found a (insane) problem that's visible in ping, perhaps you could try traceroute. The pattern of latency would be harder to check in a traceroute, but you can use ping on the intermediate routers which traceroute shows you the address of (e.g. on the way to google.com). You might prefer the mtr version of traceroute, or traceroute --icmp. My guess is you will find the problem also occurs within the first few routers. – sourcejedi Nov 17 '16 at 21:25
  • I notice the way that both traces are centred around whole numbers of ms, and decrease by 1ms each time, starting from 100ms, looks very deliberate. If that was implemented inside the network, it would almost certainly be within 0.1ms, i.e. your home router. Also you the router would need to track state. That is, in order to know how much to delay the ICMP reply to make a total delay of 100ms, you would need to remember when the ICMP request passed you originally. – sourcejedi Nov 17 '16 at 21:34
  • 1
    A few updates. I tried starting two pings like you said and they both start at 100. Next I tried pinging the router 192.168.1.1, and got the same thing, 100 ms and counting down by one. So the problem is somewhere between the machine and the router – Matt Nov 18 '16 at 1:58
  • 1
    ISP is Verizon, using the router that is supplied by Verizon. There really isn't anything unusual about the setup. I will run nping and see what else I can find. – Matt Nov 18 '16 at 2:19
  • 1
    The machine hasn't been running long, so it might have been like this all the time. I noticed that right after a reboot it is often working fine, but after it has been running for while, everything slows down. – Matt Nov 18 '16 at 15:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.