Hello there,
I want to bring up a Wifi latency issue I’m experiencing with WhatPulse. I found this out by noticing that my Wifi was flaky on a cyclic behaviour (verified this by inspecting latency between my router and Mac by pinging it).
I then slowly removed running applications until I found the culprit – WhatPulse. I’m guessing WhatPulse is doing some cyclic monitoring or something to the Wifi?
The following are four ping outputs of which identifies that WhatPulse + Wifi is a problem for latency (raw data is here https://gist.github.com/kevinjalbert/862144b2470b6afd94cb):
~❯ tail wifi-with-whatpulse
— 192.168.1.1 ping statistics —
300 packets transmitted, 298 packets received, 0.7% packet loss
round-trip min/avg/max/stddev = 2.366/44.129/3788.314/247.017 ms
~❯ tail wifi-without-whatpulse
— 192.168.1.1 ping statistics —
300 packets transmitted, 299 packets received, 0.3% packet loss
round-trip min/avg/max/stddev = 2.615/9.293/206.989/17.424 ms
~❯ tail eth-with-whatpulse
— 192.168.1.1 ping statistics —
300 packets transmitted, 300 packets received, 0.0% packet loss
round-trip min/avg/max/stddev = 0.398/0.517/0.714/0.049 ms
~❯ tail eth-without-whatpulse
— 192.168.1.1 ping statistics —
300 packets transmitted, 300 packets received, 0.0% packet loss
round-trip min/avg/max/stddev = 0.391/0.509/0.704/0.051 ms
I’m wondering if anyone could confirm that this issue is indeed to do with WhatPulse and if there is a potential solution?
Regards,
Kevin
Hi,
It’s not a WhatPulse thing, it’s an OS X thing. They’ve had issues for years now, but the latest seem to have impacted a lot more people then usual. The solution, or better said work around, is here: https://medium.com/@mariociabarra/wifried-ios-8-wifi-performance-issues-3029a164ce94 (the ifconfig bit)
10.10.1 said to have a fix, but it doesn’t for everyone. I still have issues as well, that work around works for me.
Thank you for the linke, I gave it a read through. It is interesting how airdrop can play such a large factor in wifi stability. I took the following advice and disabled my AWDL interface.
After I disabled the AWDL interface I am still seeing ‘huge’ latency spikes when WhatPulse is running. I understand that there are underlying issues with OS X’s wifi. I’m just concerned that this is something that WhatPulse is directly causing (from my perspective it appears that is the case). I can easily witness the latency spikes by enabling and disabling WhatPulse while I’m on Wifi.
Wifi=ON, AWDL=OFF, WhatPulse=ON
— 192.168.1.1 ping statistics —
300 packets transmitted, 300 packets received, 0.0% packet loss
round-trip min/avg/max/stddev = 2.330/20.901/304.584/40.037 ms
Wifi=ON, AWDL=OFF, WhatPulse=OFF
— 192.168.1.1 ping statistics —
300 packets transmitted, 300 packets received, 0.0% packet loss
round-trip min/avg/max/stddev = 2.265/5.633/103.555/7.471 ms
Can anyone else confirm a similar behaviour (ping -c 300 <gateway/router IP>)?
It’s not the client, really. All the OS X kernel does is send a copy of the network traffic to the client (or wireshark for that matter), asynchronously. It doesn’t wait for the client to receive it, it simply does not care whether the client receives it. As it (and any other OS) does not want to rely on another program for its network performance.