I don't think i'm following your question of "which param can be used", can you elaborate.
FYI this is open source software so if you or your colleagues wish to contribute to help bring up the accuracy please visit the project page: github.com/.../speedtest
rdk, the variability in latency makes me think the speed test server that is being tested against isn't stable (see Jriddle's above where that server stopped responding). Can grab the debug version ( media.thehelpfulhacker.net/index.php ) and try the following. Run "speedtest -l" and find a server that is geographically close to you. Run "speedtest -d -s SERVERNUMBER" a few times to run a debug test against the same server. You can send results to firstname.lastname@example.org and i'll look them over.
One issue though... doesn't speedtest.net exist based on their advertising revenue -- something you're kind of bypassing with command-line tools?*edit* - I see Brian brought this up as well.
email@example.com Please remember that zpeters has uploaded this as a goodwill gesture to all of us on here. As much as I'm sure he wants to know about issues there's a fine line between advising of issues and complaining that the scripts aren't working properly. Speedtest.net in my experience can return very random results.
@zpeters - You link to the testmy.net page and I'm wondering if there's any way you can integrate with their setup at all as they seem to think their testing is more accurate.
Looking at the tests that I've run my upload speeds also seem a little off and when I run comparisons on speedtest and testmy.net they don't agree either so getting an accurate result seems a difficult task.
Thank you for your remarks but I don't think this is the purpose of those forums, trying to blame the other for having asked.
1) I was not aware of such a project, however already using speedtest for years for our clients
2) zpeters asked me to change the -v into -d so seems logic to me that I want to know what difference this would, might give.
last but not least, perhaps your english is better than mine, but I never told black on white that the scripts aren't working properly. The only thing I noticed is that there was a big difference between running the speedtest manually directly from the website and running the scripts.
So before you start shooting......
this is the first and also last time I'm going to reply on such kind of blames, insults of whatsoever....
( I'm not interested in this kind of nonsense, any way you are doing the same thing !! )
You want to make KASEYA strong or break it !!
For a little clarity here, speedtest.net is a frontend (essentially) to volunteer servers. ISP's (and other organizations) around the world volunteer their bandwidth and some processing power to have a local "foot print" wherever you happen to be connecting from. The speed readings you get are not only influenced by your ISP but all of the hops along the way and the "quality" of that server you end up connecting to. The decision of which server to test against is based on IP to GPS coordinates and a brief latency ranking.
Speedtest.net isn't super forth-coming with their algorithm for measuring speed but i do know from developing my frontend that it isn't as cut and dry as it would appear. Various methods exist for testing speed and it all depends on what we are truly trying to capture. If we think back to how TCP/IP works (scaling and "back off" ) and the fact that ISPs or networks along the way may "burst" traffic, it is easy to see how regardless of method the measure will vary every time the test is run. My attempt at finding a happy medium has been to upload and download various sizes of files (since larger files could reach a higher peak transfer rate) and average out the tests. This is not exactly how speedtest.net the website does it's measure, but again i think it's a happy medium.
I have tossed around the idea of adding more command line options to control how and what you are testing and to provide other testing algorithms. It's all in the works (maybe over the holiday break ).
I'm really please this software has raised some discussion that has certainly provided me with some insight into what results people expect from a testing software. I'll continue to develop this and share as I go.
Suggestions, bug reports ect. are best directed to github.com/.../speedtest for tracking.
Happy to help out and look forward to talking to you all on these forums in the future :-)
In your script you only speedtest-32.exe
No need for the 64 bit exe?
This is a real time saver. Thanks for posting.
I've redone this in my own way and I'd like to share. How do I attach my XML?
Use the link below - Use rich formatting - then go to the options tab and you can attach it there.
This is a great script and utility - thanks for sharing this.
Here is my take on this script - I am a bit more flamboyant (you'll see from my frequent "Write Procedure Log Entries" - and I like to be able to watch the script run and it helps hugely for debugging. This is the first one that I've shared here.
I do a few checks and I download the 32- or 64-bit executable depending on the OS, and I run it three times. Then I clean up after myself.
You'll see that I call what I call "Script Outcomes" scripts depending on the results of my IF statements. Depending on the outcome, they will send an e-mail, log a ticket, or run another script. I can provide more info on these if anyone is interested.
My expected result is an e-mail to the person that ran the script, or an update to a ConnectWise ticket. Here's an example of the output:
Speedtest was run successfully on machine sj-pc-180.
===== SPEEDTEST RESULTS =====
Pass 1 of 3...
2220 | Netcetera (London, Great Britain)
Ping: 8.34 ms | Download: 79.49 Mbps | Upload: 72.98 Mbps
Pass 2 of 3...
3047 | FidoNet (London, Great Britain)
Ping: 3.00 ms | Download: 83.37 Mbps | Upload: 77.33 Mbps
Pass 3 of 3...
Ping: 3.33 ms | Download: 84.46 Mbps | Upload: 78.63 Mbps
===== SPEEDTEST COMPLETED =====
Bertie, the BTA Robot
When i tested the script the numbers were way off from the speedtest.net website. Have you compared your results?
@Elliot, please read the previous discussion and the links provided.
oh dammed, just when I wanted to show this tool to my boss, ( the wet of Murphy ) it's failing
Finding fastest server..
2013/12/21 22:55:22 Cannot test latency of 'speedtest01.mobistar.be/speed
test/latency.txt' - 'Cannot contact server'
panic: runtime error: invalid memory address or nil pointer dereference
[signal 0xc0000005 code=0x0 addr=0x20 pc=0x4504e4]
Here is my take on this script - I am a bit more flamboyant (you'll see from my frequent "Write Procedure Log Entries" - I like to be able to watch the script run and it helps hugely for debugging. This is the first one that I've shared here
Bertie, the Robot