Hello from a newbie and question re streaming while testing

[bretherton]

Hello to all out there from Virginia.

 

My first question is, does video streaming while testing affect the testing results?

 

I am having problems streaming video from the internet (not porn).  All seems to work just great during the day but in the evening multiple buffering pauses make the activity impossible.

 

OOKLA reports that Verizon FIOS is giving me in excess of 75 Mbps at all times, however I suspect a bandwidth issue, one that varies (greatly) with time of day, so am going to gather results from TMN.

 

Thanks

[CA3LE]

Welcome to TestMy.net bretherton!

So if I understand correctly you're having a hard time streaming all that porn. Streaming porn while testing definitely affects your results.

Just messing with you dude.

But yeah, any transfer during the test will affect results. Your result will be your speed minus the transfer you having going on. If you add both together, stop the transfer and retest you'll get a result that matches. That's actually one way to test the accuracy of speed tests. TMN lines up in that test.

It happens to me all the time. I run a test and it's slower than normal. Sometimes down to 1 Mbps and sometimes only a few Mbps off my max. But it makes me think something's wrong, even if it's only a little because my connection normally runs on a flat curve (not much fluctuation). If it's severely affected it's usually because I have usenet going. Other times I've found transfers minimized that I slipped my mind. Various reasons but every time it makes sense and I go, "Oh yeah." ---

Funniest time it happened, kinda unrelated... I'm on my iMac which unless I'm purposely limiting my connection (for development) it always performs over 100 Mbps. Sometimes as high as 125 Mbps. Anyways. I run a test and I'm getting < 1Mbps. I run a multithread and my results are slightly higher. Then, I look at my bandwidth in Activity Monitor to make sure nothing else is going on... zilch. like 84 bytes in and out. I'm really confused. I run Mercury speed tests to Google [w/larger test images], Wordpress [w/extra large test images], Amazon (large images), Cloudfront, Tumblr [w/larger test images], National Geographic... a bunch of locations that have been testing really well for me... same thing. I'm running super slow... maximum I saw was about 2 Mbps. As I lean back in my chair to think I see my iPhone on the desk, with the effen cord plugged in. I was testing with it weeks earlier and I set the networking priority to use the iPhone's connection first. I had multiple streaming connections going while I plugged my phone in and nothing was disturbed so it never occurred to me that I was running off my Verizon connection. Which runs like crap in my office. I literally smacked myself in the forehead, I look closer at my TMN results, "Verizon Wireless" - damn it! I just wasn't paying attention. I unplug it, retest on Mercury to googleusercontent -- 121 Mbps. Mystery solved.

First thing you need to do if you want to improve your speed is stop wasting your time with all the other tests. TMN has far more capability that any other speed test online. It's a true test and each different type offered has purpose and value to understanding your bandwidth. Most importantly the results here are not adjusted. TMN is a neutral unbiased opinion and now with Mercury you can literally test MY net because it's can be customized to test from the sites you normally use. Every test result is processed exactly the same without bias.

I don't know how the other guys do it. I pretty much just focus on building and don't pay attention to the other tests. Sure, I look at them.. they're my competition but every time I update myself on what's out there it makes me laugh. Are they serious?! Results so far from reality it's a joke. Sometimes they'll jive but NOT when you need them to. Not when there's an actual issue. And then providers will force clients to use these inaccurate tests. Ookla themselves admits in their own wiki, "The fastest 10% and slowest 30% of your results are discarded" [related post] -- so when you feel those issues and but test at full speed, that could be part of the reason. Why anyone would use a test that removes that data is beyond comprehension. ...why > 100X more people would use it over TMN ... I don't know. I guess people don't like when someone holds a mirror up to their face and shows them they've been walking around all day with a booger hanging out of their nose.

With any test, if you remove part of the data it nullifies the result.

"Hello Mr. Bretherton. I have your EKG. You see these two lines? Okay, anything that was above this line and below that line... we're just going to ignore. Your heart may have stopped somewhere in there, not sure because we discarded that data. ... ... aaannnyways ... ...Everything checks out, have a good day!"

I don't know about you but I'd find a new Doctor.

Support the little guy and please spread the word if you like TestMy.net, this is 100% word-of-mouth.

[bretherton]

Thanks for the response.  I suspected that streaming would affect the results but wanted official confirmation.

 

I think TMN is great and exactly what I was looking for.  I originally found it by 'googling' for something that would automate the gathering of results over a period of time.

 

I only have two minor issues:

1.  I switched on multi-threading but now cannot persuade TMN to go back to single threading.

2. A couple of times I was preparing a 'post' but then TMN cut in to conduct a test, consequently my 'unposted input' was lost and I had to start again.  Now I try to remember to switch off testing before preparing a post

 

TMN is a great tool, thanks for developing and releasing it.

[CA3LE]

Thanks for the response.  I suspected that streaming would affect the results but wanted official confirmation.

 

I think TMN is great and exactly what I was looking for.  I originally found it by 'googling' for something that would automate the gathering of results over a period of time.

 

I only have two minor issues:

1.  I switched on multi-threading but now cannot persuade TMN to go back to single threading.

2. A couple of times I was preparing a 'post' but then TMN cut in to conduct a test, consequently my 'unposted input' was lost and I had to start again.  Now I try to remember to switch off testing before preparing a post

 

No problem, I'm glad that it's what you're looking for.

 

There was a bug with the multithread enable and disable links that was recently corrected.  Sorry about that. If you're still stuck on multithread clear your cookies and that will resolve the issue.

 

With the auto test, although you can surf in that window I don't really recommend it.  When I run an auto test I open it in a different browser window or tab and ignore it until I get the 'complete!' email.  It doesn't care that you're typing in the window... you told it to test every x minutes, so it runs that cycle.

 

It doesn't automatically open in a new window because I don't like being forced to open in a new window myself (in general).  If you want to open in a new window hold the shift key while you press the start button (new tab is command/control+click).  Then pop over to the other window/tab and let the test run in the background.  If you happen to notice activity on the test pause what you're doing and wait till it's done before you resume for best results.

 

TMN is a great tool, thanks for developing and releasing it.

 

I appreciate the kind words.  Keep visiting, I have much more in store for you.

[Pgoodwin1]

Haha. I had to go look in the mirror