CA3LE

The most common question that people ask us is why their results here differ from speedtest.net. I want to explain something that may help you understand why this is.

According to the Ookla Wiki [updated link] the following is true about your tests taken at speedtest.net...
The fastest 10% and slowest 30% of your results are DISCARDED The remaining data is averaged together to determine the final result Throwing away the fastest 10% and slowest 30% of the results in my opinion does not make for an accurate test. Isn't THAT the data that you're here to see?

Also note this isn't only true for speedtest.net. Ookla is huge, most of the speed tests online run their software.

Source: wiki.ookla.com

Multithreading makes your connection look better
Ookla speed tests are also multithreaded. Meaning that they open more than one connection to the host and combine the speeds. This often can mask congestion issues. For instance, imagine that you have a pipe along your route that's limiting you to 10 Mbps. If you open more than one connection through that pipe you'll be able to achieve a faster speed... but testing that way will not clue you in that there is actually a problem along the route. Remember, you shouldn't have to multithread your connection to pull your full speed.  
 
Multithreaded speed tests, like the majority out there, are designed to benefit the Internet provider not the consumer.  They often display your maximum throughput not your throughput over the course of an entire upload or download.  Omitting the worst portion of your test resulting in inflated scores that may make you feel warm and fuzzy but aren't going to help you see and resolve connection issues.
 
Multithreading can show that you're able to max out your connection by combining the speed of concurrent connections but a great connection can max out without having to multithread.


So, if you have to multithread your connection to get your full speed, you should be asking why that is.  With TestMy.net it's instantly apparent if there is a congested route. Your speed here reflects the actual loading time of data within your browser... not the combined efforts of multiple threads being altered (dropping the top 10% and bottom 30% of the results) and added together, all through a plugin that isn't suited for the task.
 
:: EDIT ::
If you'd like to multithread with TMN, I offer that as well.  Try the multithread speed test.  TMN's multithread test gives you the option to select from an array of servers.  Allowing you to test across multiple routes, to more than one server at a time.  TestMy.net is the only Internet speed test with this ability.

So THAT'S why your results differ.   --  TestMy.net is a harder test to ace and a lot less forgiving than other speed tests.  But isn't that what a benchmark should be?

:: EDIT ::
Another thing to consider is if you're running windows, particularly if it's older than Windows 7 you may need to tune your TCP stack to see your full speeds. Windows doesn't always come out of the box optimized for fast connections. Flash based speed tests fail to detect this problem.

To make this change for free I recommend TCP Optimizer. This makes changing those settings very easy and nearly fool proof. Just open it, slide the bar over to your speed that you're supposed to have, check "modify all adapters" ...apply the settings and reboot. You should have faster speeds after your re-test if that was the problem.

Here's an outside source talking about exactly this issue, I'll hunt down more examples. >> Download Speed Test - Something fishy?

A few related topics ::
Accuracy.. Slow upload, drastically different from other speed tests Resuts Vs. Actual Download speeds? TestMy.net shows different results than other speed tests TestMy v Ookla Difference Between Speed Test Sites? Satellite Service Testing Locking at around 14Mbps Questioning speedtest.net resolved with TestMy.net  
Outside Sources ::
Internet providers caught inflating speed test results [myce.com] Use Testmy.net or else!! NO MORE speedtest.net posts will be allowed!!! by the Admin on [xplornetsucks.com] Accurate Speeds Or Speed Test Error? on [wildblueworld.com]

Thank you Obeahman!
 
This is one of the best testimonials we've received in a long time. Feedback like this drives my development.
 
It's really nice to hear stories of TMN working for people. It's pretty crazy how under served so many people are and how most tests don't seem to notice. It's hard to convey to the majority of people why TMN is different so first hand testimonials are really helpful. There are major differences here that in my opinion render other tests useless when most needed. You witnessed it yourself.
 
I'm happy to hear that you found TestMy.net and that it's done what I originally set out to do, help you get what you're paying for. Hopefully in the future more people find us here. 
 
Much appreciated on this end as well.
 
 

Hey all. Just wanted to say a big thank you to the creators and maintainers of testmy.net.
 
I live in Jamaica and we don't have fast or quality internet service here unless you live in the more populated areas. Where I live there are no physical internet connections available so I've had to use fixed wireless options for years. I've been through every single one that's been available and your speed test has helped keep my sanity. Basically all of them have tried to send me to the sites that make them shine and make me look like the crazy unreasonable customer.
 
My current provider just upgraded my package after almost 15 years and wanted me to do daily speed tests to make sure everything was fine. They told me to use openspeedtest.com and I told them I would use testmy.net instead (been using you for years and loving it, THANK YOU again) but they said they wanted to measure jitter as well so I relented. Well, after a while I  noticed the connection was acting up and while testmy.net gave me the real world, openspeedtest.com was giving results saying I had full speed and nothing was wrong.
 
To cut a long story short, after 2 months of telling them something was wrong and trying to convince them to use testmy.net instead, they have finally relented and admitted there's a problem  on their network. So thank you again to each and every one of you who has and anything to do with this site. Please continue to do what you do and fight for the little guy.
 
Much appreciated.

btw, the reason this website exists was to test my Cox cable internet back when it was first released in 1996 (I was 15) ... to see what throughput I was actually getting from Cox Communications.   (TMN got its name later)
 
Really though, I think the automatic test you've been running is already telling us a lot.
 
?t=u&d=01-16-2022+%2F+01-17-2022&y=u&l=25&q=quid's Speed Test Results
 
A couple things I would check,
 
Router placement: How far away is the router and what obstructions are present? Are there any mechanical devices near the router that may be interfering with the signal. Is the router confined in an enclosed space? Other devices: What are the other devices doing? Can you run a test for a period with all other devices disabled? (1 pm to 1am your time should be a good period given what I see above)  
Is your arlo power wired or battery? If it's a battery only unit I can give you instructions for how to give it hardwired D/C power, this not only will save you money on those expensive a123 batteries but it will make it more reliable in cold temperatures. You might even have a suitable power adapter that you can repurpose for free.
 
If you can, take a laptop outside near the arlo camera and run some tests. Go to https://testmy.net/mysettings and change the first (e.g. laptop, 1, 2, 3) so that we can tell the results apart. It could be that double pane windows, metallic window tint and/or materials your house is made out of are making it difficult to communicate. Your laptop may have more powerful wifi than the arlo but it may still provide some clues. Test first inside the door (maybe use identifier 1), then outside (identifier 2), then back inside (back to identifier 1). 
 
So first, I'd make sure the placement of the router is optimal. Then I'd make sure there aren't computers or devices mucking it up. Could be a torrent in the other room you forgot about (maybe a family member or roommate), could be your phone uploading a bunch of stuff to the cloud when you get home after a long day taking photos. There are many scenarios that can consume upload and make it an issue for everything network-wide. If your upload is it all taken up, you're internet's going to suffer across the board. If you disable all devices and the issue clears up but you don't know where to start, then maybe reconnect them one by one until you see the problem. Pinpoint and target the machine that seems to be slowing things down.
 
No matter what, I'm going to make sure you get early access to my new tool because it seems like it's right up your alley. It tests your home connection every second of every day while also respecting your network bandwidth. I've been running it myself since 2020 and have never noticed it running, unless I have connection issues. Then it makes its presence known. It's designed to run indefinitely and use minimal resources, all around. Like everything at TestMy.net it only requires a modern web browser.
 
This is actually different from the code in development I was previously talking about. And then in thinking about what you were saying,"test my upload throughput for an extended period" got me thinking since yesterday about a totally different upload test method, a hack to the current method that would provide deeper insight. Seriously, thank you! I don't know if it will work but it will be fun to experiment with. If it does work I can see a huge benefit.
 

I'd like to test my upload throughput for an extended period such as one or two hours.  I realize that this will entail uploading Giga bytes of data and tie up a server -- I'd be willing to pay for such a test.  Are there any tools for doing this?

Thanks!
   btw, the reason for this is that I'd like to see what throughput I am actually getting from Cox Communications.  I recently installed some Arlo Pro 4 cameras and found that they dropped offline a lot.  In fact I have two cameras that cycle offline and online every couple minutes (even with nothing on the network but one camera, a smartphone (running the Arlo app) and a Netgear WAC120 access point, and router the camera goes offline).  So, I tried using my cell phone as a mobile hotspot (which delivered about 80 Mbps down and 50 Mbps up -- and that worked.   Arlo then, (unsurprisingly) said that my 150 Mbps down and 10 Mbps up plan from Cox was not enough.  Honestly, a 10 Mbps upload should be fine.
   So, I upgraded to 1 Gbps down and 35 Mbps up Cox plan.  The Arlo camera performance is much better, but, even running one camera, it drops offline occasionally.  And the network speed seems to drop a bit too sometimes.
   The speed tests you offer, as well as the automatic ones, are really pretty nice.  Still, since it is not continuous, it would miss occasional communications slow downs (which would derail my Arlo camera).  And cable plans don't guarantee a minimum throughput unfortunately. 
 
  So, I would like run a continuous test to document the actual throughput that a real-time devices (like security cameras) have to work with.  My application is not critical I suppose.  I need the camera to see when our cats show up at our back door to let them in (the occasional snowfalls are hard on them) -- so this is important to us.  It is surprising though how many folks complain about WiFi security cameras going offline.  It may be just that the ISP is not providing enough upload bandwidth.  And I can't get DSL around here which would provide a guaranteed bandwidth.

The UK server seems to have been removed, and the German one is not responding, any idea if more European servers will be added?

Still working on this, will update in a few days.

I have a new program I'd like to start beta testing very soon.
 
I've been running it myself nonstop for well over a year now. It's reliably been keeping track of my home connection's uptime / downtime. It's also helped me make TMN's services more reliable and redundant in the process because the server-side needs to be highly available for the program to make sense. Can't be sending false alarms because of server-side hiccups.
 
I really just want to open it up to people in its current state, there are minor bugs I need to address but none of them have to do with the functionality, usefulness or stability of the program. More aesthetics. I feel like it's a new concept so it will need to find it's design along the way as I get feedback. The core functionality is pretty sweet already. I'm just unsure if people will get what it's all about at first glance.
 
In your case, it does exactly what you're looking for. It passively monitors your connection every second then notifies you when your connection goes down and when it's back online. It has an interesting, unique and reliable method of operation. Designed specifically to be able to run indefinitely on unstable connections and always recover itself back into a ready state, using minimal resources.
 
Give me a few days to button this new program up, I'll send you a message and update this thread.

I have a new program I'd like to start beta testing very soon.
 
I've been running it myself nonstop for well over a year now. It's reliably been keeping track of my home connection's uptime / downtime. It's also helped me make TMN's services more reliable and redundant in the process because the server-side needs to be highly available for the program to make sense. Can't be sending false alarms because of server-side hiccups.
 
I really just want to open it up to people in its current state, there are minor bugs I need to address but none of them have to do with the functionality, usefulness or stability of the program. More aesthetics. I feel like it's a new concept so it will need to find it's design along the way as I get feedback. The core functionality is pretty sweet already. I'm just unsure if people will get what it's all about at first glance.
 
In your case, it does exactly what you're looking for. It passively monitors your connection every second then notifies you when your connection goes down and when it's back online. It has an interesting, unique and reliable method of operation. Designed specifically to be able to run indefinitely on unstable connections and always recover itself back into a ready state, using minimal resources.
 
Give me a few days to button this new program up, I'll send you a message and update this thread.

When our Internet goes down, this program stops & doesn't restart?  We need to see how often our internet is down.

You should want to keep your old results to compare against. Some of the averaging tools here like My Average work best with a history.  (e.g. CA3LE's My Average)
 
You can narrow your My Results search to a certain date to get the same effect as deleting results.
 
If you'd really like to delete your old results you can also do this under My Results. In the results details under the graph, check the boxes next to the results you'd like to delete then hit the trashcan at the top of the table. You can toggle all with the checkbox next to the trashcan. If you have a huge number of results, alter the results/page setting to display more results per page to make the process quicker.
 
Originally when I started database logging there was no client-side delete functionality. The reason it was added was to delete possible erroneous results. Wasn't ever intended on being used to totally reset results --but it can be.
 
side note: In an upcoming release you'll have the ability to reset all of your logs without actually deleting anything.

I started an Auto Test this morning at 11:44. The test was set to repeat at 15 minute intervals, 5 times.
 
At 12:22 I noticed 'Retesting at 01:28:20pm' and at 12:29 it showed 'Retesting at 01:43:21'.
 
There is an issue when calculating the time of the next test, perhaps caused by spanning 12-noon?

Thanks CA3LE
 
Totally understand Maintenance Downtime, gotta keep gear updated.
 
Kudos on your ‘Failover’ logic for Test Server selection. When I noted DALLAS was used, I of course checked MIRRORS & no Toronto!?!!
As I wasn’t sure if You were aware of Toronto status so I spoke up.
 
BTW, is DALLAS the Failover Test Site for North America/Continent or Planet  ?
( maybe a Feature Tweak there   )
 
I offer a FEATURE Tweak of MIRRORS :
- specifically for Toronto which currently  lacks duplicate nodes, update the Site Descriptor with ‘Down for Maintenance’ and make Site ‘UN-pickable’ .
 
Thanks again … your site is still the Best Kick-Ass Speed Testing beast !

It was down for maintenance and is back online now.
 
The server list is at https://testmy.net/mirror. Most locations have multiple servers to fall back on but Toronto has much lower traffic so it currently only has one. 
 
Sorry for the inconvenience.  
 

I'm lurkin' too. Reconfiguring my server cluster for high availability. Some pretty killer backend upgrades. Pretty sweet, all coming together nicely.
 
You may be able to notice when the site has migrated to the new setup. Some really heavy database queries in testing are returning at least twice as fast as they currently do. In the latest configuration I'm running directly off 100% Optane memory. I swear, it's alien technology.
 
It's not about how fast it is, there's faster (on paper, if only scratching the surface -- totally different story in the real-world) -- but it's at what queue depth that matters. And there are NAND drives out there that perform amazing. But what none of them have compared to Optane...
 
Optane can write directly to memory without the need for DRAM cache, it's THAT fast Optane doesn't need to first erase the cell before writing Optane has nearly unlimited write endurance Optane loses no performance as the drive fills up Optane does not require trim (goes on and on really) -- it's in between RAM and NAND  
For my workload it feels the same as going from HDD to solid state. Or from solid state to raid-10 solid state. It's that big of a difference, mind blowing really. Over just a few days of testing I thrashed it with about 60 TB of data and it never broke a sweat -- literally, the drive didn't fluctuate in temp.
 
SMART/Health Information (NVMe Log 0x02) Critical Warning: 0x00 Temperature: 40 Celsius Available Spare: 100% Available Spare Threshold: 0% Percentage Used: 0% Data Units Read: 48,816,551 [24.9 TB] Data Units Written: 105,693,840 [54.1 TB] Host Read Commands: 937,508,694 Host Write Commands: 1,177,333,140  
It just sits there at 40°C like, "Is that really all you've got? Do you realize who I am? HIT ME! I DARE YOU! This is my idle temp FOO!"
 
 -- in other words, it's bad ass. It's a different kind of memory entirely.
 
Old but relevant video describing 3D XPoint (cross point)
 
 

Definitely need to work on SEO.
 
Best thing people can do is spread the word. 
 
And you're right, to the casual user -- they have no idea what makes TMN unique. Unfortunately the majority just follows where the rest of the crowd goes. A vicious cycle, feeding into itself.      -- implications far deeper than just speed tests.  lol
 
One of these days I hope to get my message across better, where more people want to engage. Until then, all I can do is keep developing. At the end of the day, I'm happy just serving those who already understand the difference.
 
Forums everywhere in general, died. The Internet changes and you have to change with it. I may eventually push the forums off into an archive in favor of a more modern method of communication.

I had forgotten all about the sleeping tabs. I went into Edge Settings > System and performance adding [*.]testmy.net to the "Never put these sites to sleep" and it has successfully made it farther than before.

I wonder if there are new power management settings and Edge may be putting the tab to sleep or something.  I'll put it on the bench and see if I can reproduce and get back with you.
 
For now, try scheduling the same test using Chrome or Firefox and see if this resolves the issue.  
 
If there are new settings I may be able to work around them.
 
Sorry for the extremely delayed response.  I got your message right away and 10 minutes after you sent it I started scheduling some tests and investigating but wasn't able to reproduce the issue (but wasn't using Edge).  I had it scheduled out for a while and then by the time it finished forgot to update this thread.  My bad.  -- Again, I'll give it a try using Edge, may provide more info.  Pretty certain it's not an issue effecting large numbers of users.

I'm not sure either because I haven't been able to ever get it to do that.
 
My guess is something on the provider end is messing with TestMy.net's output. It may be trying to accelerate previously loaded content.
 
I do a lot on my end to try to bust host caching and clear your cache programatically. But you may get an occasional outlier. Seems to really only affect the multithread download test.
 
You can delete that outlying result by checking the box next to that result and click the trash can at the top of the result details table.
 

 
When I'm not programming or building server infrastructure on the backend I often scan the results for outliers. After so many years of doing this and adjusting the program they're pretty few and far between but I would like to develop a solution for any remaining.
 
When I've rotated back into programming mode I'll make sure to make the necessary adjustments to prevent this. I think before the result is logged your average will be taken into account, it's already being queried at that stage. If the result is much higher than your average and the result is over 1 Gbps a re-test occurs with further cache busting measures. After re-test, if the second result is within a range of the first result it will allow logging. And if not, the test will die with an error so we don't create a loop.
 
When I find these in the results I usually see the client re-test soon after, the re-test always returns to the median.
 
 
It would really help if I could reliably make the issue happen but it's an intermittent issue I've never personally seen. Unless I witness it I can only try to fix it blind. But I think we have a roundabout solution, just need to program it in there.
 
Thank you for the constructive feedback, it's always very helpful. You are building TestMy.net.

So, yeah, just gonna leave this here. Have not been able to get this again, and not sure why it happened. Any ideas?
 

That's awesome
 

Definitely need to work on SEO.
 
Best thing people can do is spread the word. 
 
And you're right, to the casual user -- they have no idea what makes TMN unique. Unfortunately the majority just follows where the rest of the crowd goes. A vicious cycle, feeding into itself.      -- implications far deeper than just speed tests.  lol
 
One of these days I hope to get my message across better, where more people want to engage. Until then, all I can do is keep developing. At the end of the day, I'm happy just serving those who already understand the difference.
 
Forums everywhere in general, died. The Internet changes and you have to change with it. I may eventually push the forums off into an archive in favor of a more modern method of communication.

That would be lovely.
 
Is it difficult to get the word out about testmy.net? My hunch is that 99% of internet users don't perceive and won't understand the problem testmy.net addresses in addition to many who just assume testmy.net is wrong. But then again people often surprise me and one won't get anywhere assuming 99% of people wont like a product. Or maybe some of it may be SEO related? On google, testmy.net comes up in the middle of the second page. On duckduckgo, its the first result revealed after one clicks "see more results". I myself found testmy randomly through an LTE equipment forum post. I'm super grateful I did, and look forward to what new things come out. To the casual internet user one test tool may seem as good as any other, I would think. But there I go getting myself in trouble thinking again...

I'm lurkin' too. Reconfiguring my server cluster for high availability. Some pretty killer backend upgrades. Pretty sweet, all coming together nicely.
 
You may be able to notice when the site has migrated to the new setup. Some really heavy database queries in testing are returning at least twice as fast as they currently do. In the latest configuration I'm running directly off 100% Optane memory. I swear, it's alien technology.
 
It's not about how fast it is, there's faster (on paper, if only scratching the surface -- totally different story in the real-world) -- but it's at what queue depth that matters. And there are NAND drives out there that perform amazing. But what none of them have compared to Optane...
 
Optane can write directly to memory without the need for DRAM cache, it's THAT fast Optane doesn't need to first erase the cell before writing Optane has nearly unlimited write endurance Optane loses no performance as the drive fills up Optane does not require trim (goes on and on really) -- it's in between RAM and NAND  
For my workload it feels the same as going from HDD to solid state. Or from solid state to raid-10 solid state. It's that big of a difference, mind blowing really. Over just a few days of testing I thrashed it with about 60 TB of data and it never broke a sweat -- literally, the drive didn't fluctuate in temp.
 
SMART/Health Information (NVMe Log 0x02) Critical Warning: 0x00 Temperature: 40 Celsius Available Spare: 100% Available Spare Threshold: 0% Percentage Used: 0% Data Units Read: 48,816,551 [24.9 TB] Data Units Written: 105,693,840 [54.1 TB] Host Read Commands: 937,508,694 Host Write Commands: 1,177,333,140  
It just sits there at 40°C like, "Is that really all you've got? Do you realize who I am? HIT ME! I DARE YOU! This is my idle temp FOO!"
 
 -- in other words, it's bad ass. It's a different kind of memory entirely.
 
Old but relevant video describing 3D XPoint (cross point)