Otdfbt

I am experiencing an annoying condition on my network. I first noticed it when my video calls started freezing for few seconds every minute or so.

After digging in, I noticed that there is a pattern. The network drops exactly every minute.

Here's a graph showing latencies of pings:

(it's pinging Google's servers and the graph is generated using a nice program PingPlotter)

As can be seen, the latencies increase exactly every minute. Sometimes the pings timeout.

More info about the setup:

• the client computer is a Mac OS laptop, connected to my home 2.4ghz wifi network


• the network is ADSL with modem/AP being close to the laptop. The modem is quite old Comtrend model. It doesn't have support for 5ghz wifi.

Things I tried:

• the problems do not happen on other clients connected to the same network (tried a Windows laptop and an Android phone)


• I tried watching the behaviour on different wireless networks, but couldn't get conclusive results. The problems do not happen when I connect the Mac a phone hotspot (screenshot). However, they do happen on a wifi in a cafe (screenshot).


• I don't have an ethernet port on the laptop or the adapter with me here, so can't connect the laptop to the problematic network via cable.


• restarting the modem to factory defaults didn't fix it


• reproduced using native ping command:

$ ping 8.8.8.8 -i 0.2
PING 8.8.8.8 (8.8.8.8): 56 data bytes
...
64 bytes from 8.8.8.8: icmp_seq=243 ttl=120 time=46.194 ms
64 bytes from 8.8.8.8: icmp_seq=244 ttl=120 time=44.044 ms
64 bytes from 8.8.8.8: icmp_seq=245 ttl=120 time=43.062 ms
64 bytes from 8.8.8.8: icmp_seq=246 ttl=120 time=43.593 ms
Request timeout for icmp_seq 247
Request timeout for icmp_seq 248
64 bytes from 8.8.8.8: icmp_seq=247 ttl=120 time=580.329 ms
64 bytes from 8.8.8.8: icmp_seq=248 ttl=120 time=415.218 ms
64 bytes from 8.8.8.8: icmp_seq=249 ttl=120 time=213.723 ms
64 bytes from 8.8.8.8: icmp_seq=250 ttl=120 time=50.411 ms
Request timeout for icmp_seq 253
Request timeout for icmp_seq 254
64 bytes from 8.8.8.8: icmp_seq=251 ttl=120 time=839.641 ms
64 bytes from 8.8.8.8: icmp_seq=252 ttl=120 time=634.475 ms
Request timeout for icmp_seq 257
64 bytes from 8.8.8.8: icmp_seq=253 ttl=120 time=1135.625 ms
64 bytes from 8.8.8.8: icmp_seq=254 ttl=120 time=930.481 ms
64 bytes from 8.8.8.8: icmp_seq=255 ttl=120 time=726.878 ms
64 bytes from 8.8.8.8: icmp_seq=256 ttl=120 time=1202.733 ms
64 bytes from 8.8.8.8: icmp_seq=257 ttl=120 time=999.965 ms
64 bytes from 8.8.8.8: icmp_seq=258 ttl=120 time=794.788 ms
Request timeout for icmp_seq 264
64 bytes from 8.8.8.8: icmp_seq=259 ttl=120 time=1256.415 ms
64 bytes from 8.8.8.8: icmp_seq=260 ttl=120 time=1055.366 ms
64 bytes from 8.8.8.8: icmp_seq=261 ttl=120 time=854.074 ms
64 bytes from 8.8.8.8: icmp_seq=262 ttl=120 time=691.366 ms
64 bytes from 8.8.8.8: icmp_seq=263 ttl=120 time=493.404 ms
64 bytes from 8.8.8.8: icmp_seq=264 ttl=120 time=290.603 ms
64 bytes from 8.8.8.8: icmp_seq=265 ttl=120 time=86.724 ms
64 bytes from 8.8.8.8: icmp_seq=266 ttl=120 time=114.285 ms
64 bytes from 8.8.8.8: icmp_seq=267 ttl=120 time=42.717 ms
64 bytes from 8.8.8.8: icmp_seq=268 ttl=120 time=52.216 ms
64 bytes from 8.8.8.8: icmp_seq=269 ttl=120 time=59.150 ms
64 bytes from 8.8.8.8: icmp_seq=270 ttl=120 time=49.539 ms

• to see if it's the latency or bandwidth dropping, I run iperf and seen bandwidth dropping to zero. Cropped output (see seconds 60 to 65):

...
[ 10] 57.00-58.00 sec 248 KBytes 2.03 Mbits/sec
[ 10] 58.00-59.00 sec 156 KBytes 1.27 Mbits/sec
[ 10] 59.00-60.00 sec 137 KBytes 1.12 Mbits/sec
[ 10] 60.00-61.01 sec 0.00 Bytes 0.00 bits/sec
[ 10] 61.01-62.00 sec 0.00 Bytes 0.00 bits/sec
[ 10] 62.00-63.00 sec 0.00 Bytes 0.00 bits/sec
[ 10] 63.00-64.00 sec 0.00 Bytes 0.00 bits/sec
[ 10] 64.00-65.00 sec 0.00 Bytes 0.00 bits/sec
[ 10] 65.00-66.00 sec 90.3 KBytes 743 Kbits/sec
[ 10] 66.00-67.00 sec 312 KBytes 2.55 Mbits/sec
[ 10] 67.00-68.00 sec 282 KBytes 2.32 Mbits/sec

• there are no new entries added to /private/var/log/system.log when the drop occurs


• tried closing potential apps causing this (chrome, onedrive, ...) and the problem persisted. Didn't expect this to work anyway, as it's not happening on different network than my home wifi.


• run nettop and tcpdump to see if there is any traffic generated during the drops, but didn't see anything


• as the last thing, I actually tried restarting my laptop. And to my surprise the drops got much less significant:

(the differences in the baseline can be attributed to VPN - I was connected to the VPN before the restart and then connected again at 12:54)

Q:

What can be causing this? What to diagnose next?