0
I'm busy setting up my Plex server on a separate Wi-Fi router and I'm testing the transfer speeds between two MacBook Pros. It's just got me baffled because I've come across something weird. Not being familiar with the terminology, I don't know what words to type into Google so I've come here for help.
I've been using TotuSoft LAN Speed Test Lite to try and see where the bottlenecks are. The test involves uploading and downloading a 20 MB file to and from each computer (I run the program on computer 1 and then again on computer 2).
I expected that the upload speed when said program is run on Computer 1 will be the same as the download speed on Computer 2 and vice versa.
Note that the test doesn't measure that directly. It measures upload and download speed on one computer using the other as the target. However, if I repeat the test on both computers, the upload speed on one ought to match the download speed on the other since the same file transfer is used to test both. The results from test 1 are actually:
Readings on Computer 1: 5 Mbps upload, 85 Mbps download
Readings on Computer 2: 9 Mbps upload, 90 Mbps download
Here's a screenshot of similar results from test 2:
Note: I used my USB3 HDD as a target location on my mac as it has a SSD and would make the results less comparable if the network transfer speed wasn't a bottleneck.
Why do the tests appear to show an order-of-magnitude difference between sending and receiving, regardless of which direction the test is performed? I totally get why the upload and download speeds are different over the Internet, but not on the same network.
Christian Johnson-Richards
Posted 2015-04-12T04:06:55.903
Reputation: 11
You set up the discussion by saying you expect the upload speed on one computer to match the download speed on the other. Then you show results, but what you've shown is ambiguous. For example, you show upload and download speeds under Computer 1. Does that mean when you upload on computer 1, that's the download speed on computer 2, or is that the download speed on computer 1 when you upload on computer 2? – fixer1234 – 2015-04-12T06:03:55.510
Okay sorry about the ambiguity, its difficult to get across because in itself theres a lot of overlapping language/jargon.
Right so I connected both computers to a wireless access point that is not connected to the internet and made sure file sharing is enabled on both mac's. I then used this program here: http://www.totusoft.com/lanspeed1.html which i first run on computer 1 and it spits out the two readings of upload and download as per my original post underneath computer 1, I then run the program on the second computer, is that better?
I just want to know why Im getting those results? because in themselves they dont make sense, surely on computer one the upload speed you get when your run the program should be exactly the same as the download speed you get when your run the program on computer 2 or there about – Christian Johnson-Richards – 2015-04-12T06:16:06.743
I mean at least of the same sort, not a factor of 10 out – Christian Johnson-Richards – 2015-04-12T06:16:45.243
I'm going to do it again and take screenshots bc this is bugging me beyond belief – Christian Johnson-Richards – 2015-04-12T06:26:39.917
Sorry I should reword my reply a bit to you're comment; Its neither of the things you suggested, basically how I see it is you can give or take files from computer 1 to computer 2 OR you can give or take files from computer 2 to computer 1 however what I am asking is why is the speed at which you can do this different depending on whether you are giving files away (from the results it seems like this is slower) rather than taking files from another computer on the same network? – Christian Johnson-Richards – 2015-04-12T06:51:44.847
It should be the same thing omg what is life!?!?!?! – Christian Johnson-Richards – 2015-04-12T06:58:20.493
Maybe try to use iperf instead of your tool so you can see both sides simultaniously. https://iperf.fr/download/iperf_2.0.5/iperf-2.0.5-macosx-i686.tgz
– Nikolai – 2015-04-12T07:30:17.623Ill try it, in the mean time any theory's? Its not un normal, i mean everything works fine plex streaming a 4GB movie works fine, idk maybe i should just let it be :( – Christian Johnson-Richards – 2015-04-12T07:44:35.410
I would try using iperf. The application here relies on shared folders which in turn run using some extra protocol. That coupled with the fact that wireless in inherently half-duplex introduces a lot of variables. We also don't know how the tool itself is working as we have no access to the source. Using a tool like iperf which does the testing on both sides simultaneously can remove some of those variables. – prateek61 – 2015-04-12T12:42:59.960
I can't seem to figure out how to work iPerf through trial and error (never used console commands really) so if I want to figure this out I'm gonna have to do a bit more research so please don't expect a reply for a few days ish – Christian Johnson-Richards – 2015-04-12T14:13:32.163