Sunday, June 20th, 2021 • Welcome to the no-nonsense zone!

Internet or Wi-Fi Speed Test: How You Can Figure Out the Correct Numbers

At one point or another, you might have wondered how to do a proper Wi-Fi speed test or the right way to figure out how fast your Internet actually is.

In this post, you’ll find the answer to that question, namely the absolute speeds of your connections. Take absolute with a big grain of salt, though, since these speeds vary a great deal.

As a bonus, I’ll also reveal how I test Wi-Fi speed — including the latest Wi-Fi 6/E — for my reviews on this website. Spoiler: It’s easier, though more time-consuming than you think.

Before we continue, make sure you understand how Wi-Fi and the Internet are two different things.

Dong’s note: I first published this post on December 19, 2018, and last updated it on June 7, 2021, to include additional relevant information.

Internet or Wi-Fi Speed Test: Netgear CM600 Cable Modem
Internet or Wi-Fi Speed Test: Figuring out those numbers can be a challenge.

Why you should care about your Internet speed

It’s generally helpful to know how fast your Internet is. Faster is always better, but most importantly, you need certain speed grades to do specific tasks. Take video streaming — one of the most bandwidth-taxing online activities — for example; you’ll generally need a minimum download speed of:

  • 3 Megabits per second for DVD quality.
  • 5Mbps for HD quality.
  • 25Mbps for Blu-ray (4K) quality.
  • 80Mbps for 8K video.

Keep in mind that this is the speed required by a single stream. If you have more than one person streaming simultaneously, you generally need to multiply those numbers with the concurrent clients to figure out the necessary real-time bandwidth.

Also, there are a lot more online applications than streaming. Some of these applications — such as automatic updates — even occur within a connected device without you getting involved.

So yes, again, the faster Internet is always better. But there’s more than the download speed when it comes to the Internet.

Internet connection explained

When testing an Internet connection, you’ll get two main numbers, download and upload — sometimes they are called downlink and uplink. And you might also see a few other values, including ping, jitter, and package loss.

Following is the breakdown of what they mean.

Internet speed: Upload vs. download

Download speed represents how fast you can pull things from outside your local network — specifically from the remote server that hosts the test data — onto your device. That said, streaming a movie, surfing a website, downloading a file, getting an email, etc., use the download pipe.

Upload represents the speed of the opposite direction. Things like sending an email, posting a photo or comment to Facebook, saving a file to Google Drive, etc., use the upload pipe.

(Note: Cloud-recording security cameras like those from Google or Arlo can put a massive strain on the upload.)

In most traditional broadband connections, such as one via a cable modem, the download speed tends to be much faster than the upload. But you might get the same rate for both directions with a more modern connection, like a fiberoptic service.

While we use more of the download pipe, data transmission — delivered in small portions called packets — needs both directions to work. That’s because your computer uses the upload pipe to send a confirmation to the remote server that it has received a packet and is ready for the next one.

That said, if the upload pipe is all clogged up, you can’t download anything at all.

Packet loss

As the name suggests, packet loss happens when a packet of data is either not received or partially received.

Most of the time, this is the consequence of a bad physical connection or incompatible MTU settings. In this case, the packet will be resent. Packet loss shouldn’t be higher than 1 percent.

See also  Going Gig+ Broadband? Keep Tab of Your MTU and Jumbo Frame

Lag (or latency): Ping vs. jitter

Both ping and jitter relate to the delay in a connection, but they are slightly different.

Ping is the fixed latency or lag at a given time — in a speed test, it’s generally measured at the beginning. It’s the amount of time, shown in milliseconds, needed for a data packet from one party to reach another or vice versa.

The shorter your ping is, the better your Internet connection is for applications that require real-time interaction, such as Wi-Fi calling or online gaming. Generally, a ping of 15ms or shorter is considered excellent.

Jitter, often referred to as Packet Delay Variation (PDV) or ping variations, measures the changes in ping values over time — the entire test.

In other words, if every packet takes the exact amount of time — no matter how long — to arrive at the destination, then there’s no jitter.

The higher the jitter value, the more likely a packet loss will occur. Jitter value should be below 30ms and is generally below 10ms.

See also  QoS Explained and How to Get Better Voice and Video Calls over Wi-Fi

How to do a real Internet speed test

There are many speed test websites, such as Speedof.me, Fast.com, or Speedtest.net.

Don’t get too obsessed about which to use. They are all the same in the test methodology. The results will vary, though, due to the server’s location (and its Internet speed). So pick one that’s best for your location, and in that sense, Speedtest.net is great.

Since you’re on this page, do a quick test right now.



(Note: This test will use up your data. Also, Ookla may collect certain information from your connection.)

So how fast is it? Totally fast, and you’re happy with it? Good for you! You can move on now. But if for some reason it’s not what you expected, keep in mind that chances are it’s not your correct Internet speed anyway.

(Note: If you were using a phone’s cellular connection during the test, then that was indeed the real speed of your Internet. However, mobile Internet speed always varies a great deal depending on where you are.)

Let me explain. When you did the test, there might have been other devices in the network, also using part of the connection’s bandwidth.

If you have an ultra-high-speed broadband plan, the local Wi-Fi or your computer’s wired connection might not be fast enough to deliver the Internet speed in full.

That said, to find out the real speed of your broadband connection, you need to do a bit of preparation before the testing.

Netgear CM600 Cable Modem 3
Your Internet speed is generally only correct at the modem or any similar terminal device at your home’s broadband drop.

What you need to do a real Internet speed test

  1. A test computer with a network port. A Gigabit port is fine most of the time, but if you want to test a full Gigabit, Gig+, or faster broadband connection, a Multi-Gig-capable (2.5Gbps, 5Gbps, 10Gbps) computer is a must.
  2. Connect the test machine directly to the Internet source — such as a modem, a gateway, or a Fiber ONT — using a network cable. The objective here is to remove all middle devices, like a slow switch or router, that can be the bottleneck.
  3. Check to make sure the computer is the only device using the broadband connection during the test. For example, if the Internet source is a gateway, make sure you disconnect all other devices from it (unplug all other network cables and turn off its Wi-Fi.)

And that’s it! Now on the test computer, do a few speed tests as you did earlier, the number you get is your real Internet speed.

By the way, you’ll likely still get a different result each time you do a test. That’s just how the Internet is. Feel free to use the highest number as your broadband score.

Now, if it’s still significantly lower than what you pay for, it’s time to call the provider to complain.

Again this complicated real test is to make sure you get what you pay for. You can always do a random test on any device to find out the connection speed at hand. And if that’s fast enough for your need, don’t bother to do anything else.

How to test a router’s Wi-Fi speed test

There are many apps designed to test a router’s Wi-Fi speed. None of these are accurate for a couple of reasons.

First, to conserve battery life, mobile devices almost always optimize their Wi-Fi adapter for battery life and not speed. That’s because the former is way more important in real-world usage.

Second, none of these apps replicate what users do in real life. They are all synthetic. So, avoid using Wi-Fi test apps if you want to know how fast your Wi-Fi really is.

For the same token, you shouldn’t use Internet speed test apps to test Wi-Fi, either. A router’s Wi-Fi speed is likely much faster than the speed of a broadband connection.

And even when you have an ultra-fast Internet, there are many variations in the world wide web that can adversely affect broadband speeds. As a result, it’s rarely accurate, if at all, to use the Internet to test your Wi-Fi’s throughputs.

Generally, it’s OK to use a phone to do a speed test. Just keep in mind that it’s not 100% indicative of your Wi-Fi or broadband speeds — rather, it’s the speed of the phone’s Wi-Fi capability. Basically, if either your Internet or the Wi-Fi connection is faster than a certain speed grade, say 500Mbps, you need to test it via a real computer with the fastest network/Wi-Fi adapter to make sure.

The best way to find out a router’s Wi-Fi speed is to copy data from one computer (a server) to another (a Wi-Fi client) within the local network, using a single wired-to-wireless connection.

If you use multiple Wi-Fi clients for the testing, the router’s Wi-Fi bandwidth is shared, and therefore, you can’t find out how fast its Wi-Fi can truly be. That’s not to mention Wi-Fi software drivers tend to be optimized for battery life and therefore favor download over upload.

And that means, again, you need to do some preparation.

What you need to do a real Wi-Fi speed test

In my experience, the real-world speed of Wi-Fi 5 (802.11ac), even at its best, has never surpassed the sustained rate of a wired Gigabit connection. But Wi-Fi 6, and especially Wi-Fi 6E, can be significantly faster.

See also  Wi-Fi 6E Explained: Better Wireless Connections at the Expense of Range

In any case, keep this in mind: The connection speed between a pair of network devices is at the mercy of the lowest party involved. You can read more about that in this post about network basics, but the gist is the rate you see is that of the bottleneck device.

That said, these are what you need to test a device’s Wi-Fi speed:

  1. A computer that plays the server’s role and hosts the test data. This computer must have a Gigabit network port — or a 10 Gigabit Ethernet port if you intend to test a Wi-Fi 6 router — and use a solid-state drive as its storage. Connect this computer to the fastest LAN port of the Wi-Fi router that you want to test. (If the router doesn’t support Gigabit, you can forget about it. It’s already too slow, anyway). Now on this server computer, share the folder that contains the test data so that it’s accessible to other computers that connect to the same router.
  2. A second computer — be it a laptop or a desktop — to play the Wi-Fi client’s role. This computer needs to have a highest-end (fastest possible) Wi-Fi adapter of at least the same speed grate as that of the router. This computer should also use a solid-state drive as its storage. Connect this computer to the Wi-Fi network of the router. Make sure you separate the bands when possible, be it 2.4GHz, 5GHz, or 6GHz, into different networks to know which brand is being tested.

And that’s it! Now from the second computer (the Wi-Fi client), browse for the shared folder on the server computer and copy the data over. Time how long that process takes, do some simple math with the amount of data involved, and you’ll figure out how fast the connection is.

For example, if the data you copy is 2000 megabytes and the copy process takes 30 seconds, then the speed is 66.7 megabytes per second or 533.6 Mbps.

You can move the Wi-Fi client around to find out how the distances — between the router and the client — affect the Wi-Fi speed.

Similar to Internet speed, the Wi-Fi speed also tends to fluctuate. That said, feel free to pick the highest number of the same location as the speed of your router.

Gigabyte 10Gbps Adapter Card
Internet or Wi-Fi Speed Test: A super fast wired adapter is a must.

Wi-Fi speed test used in Dong Knows Tech’s reviews

It usually takes me about a week to finish evaluating a router. For Wi-Fi throughput speed, I use the same test method above. That means there’s a server that connects to the router in questions via a wired connection.

Important note: My testing method requires the router to a Gigabit (or faster) LAN port(s) to work. For this reason, I generally don’t review routers that use the old Fast Ethernet (10/100Mbps) wired standard, nor should you consider them.

After that, I use various high-end Wi-Fi clients to conduct the throughput tests. The following are the general specs of my equipment, which I, by the way, upgrade quite frequently — I will update this post when the changes are significant enough.

Server specs for Wi-Fi speed test

My server is a custom-built computer with the following specs:

Again, this server hosts the test data, which I use to copy to clients via the test router’s wireless connection to figure out the Wi-Fi speeds. The server connects to the test router via a 1Gbps, 2.5Gbps, 5Gbps, or 10Gbps, whichever highest available, wired connection.

Clients specs for Wi-Fi speed test

Generally, I use three clients for throughput testing. All of them use relatively high-end Intel CPU, 16GB of RAM, and an NVMe SSD as the primary storage.

  • Client #1: This is a desktop computer using an Asus PCE-AC88 4×4 Wi-Fi 5 adapter — the fastest Wi-Fi 5 client on the market. I use this mostly for a close range (10 feet or shorter) throughput test of the Wi-Fi 5 router.
  • Client #2: Apple MacBook Pro 15 mid-2015, top of the line. This laptop has a 3×3 Wi-Fi 5 adapter. I use this to mostly test the range speed (at a 40 feet distance) of Wi-Fi 5 routers. This laptop runs both MacOS and Windows 10.
  • Client #3: A Dell XPS 9550 15-inch laptop. This machine has top-of-the-line specs, and I have upgraded it to Wi-Fi 6E.

I also use a few other laptops, tablets, USB Wi-Fi adapters, and phones of different Wi-Fi standards, including some extra Wi-Fi 6 and Wi-Fi 6E devices, for additional tests.

Wi-Fi router speed test: Data and configurations

The following are the data and how I conduct tests on Wi-Fi broadcaster (routers, access points, Extenders, etc.) to get their official sustained real-world wireless speeds.

Depending on a particular device, there might be more testing. For example, a Wi-Fi 6 router will also be tested as though it were a Wi-Fi 5 router. But all of them have to go through these.

  1. I upgrade the hardware to the latest firmware available.
  2. I use single large files for test data, which generally take less time to copy than multiple small files. Depending on the tests and how fast a router is, I use a 2 GB, 6GB, 10 GB, or 20 GB test file. Generally, I use the smaller test file when a bigger one would make the test take too long.
  3. When possible, I change the router’s settings to favor speed (and not compatibility, which is generally the default). I also test each Wi-Fi channel separately, as well as using the Auto channel setting.
  4. For official performance scores, I test the router with just one Wi-Fi client at a time. I place this client at two specific locations that are (a) less than 10 feet (3 m) and (b) 40 feet (12.2 m) away from the router.
  5. I do multiple tests at different times during a day and different days of a week and use the highest consistent numbers as the final scores.
  6. Besides the performance, I also used the router for an extended amount of time, from a few days to even a few weeks, with many clients of different Wi-Fi standards and tiers to find out how the router functions in daily life.

I report Wi-Fi performance in megabits per second (Mbps).

See also  Home Wi-Fi Explained: Hertz, Frequency Bands, Streams, Channels, Range, Speeds, and a Lot More

Wi-Fi mesh speed test: Hardware placement

A multi-hardware unit setup. I always test them in the wireless setup. That’s because, in the wired backhaul configuration, the performance of the extender unit (satellite) will be similar to that of the router unit.

In this case, I place the satellite 40 feet away from the router unit. After that, the test client is placed at 10 feet and 40 feet away from the satellite.

Also, I use the star topology for the testing, meaning the satellite units (if more than one) are placed around the router unit. In other words, a 2-pack mesh will deliver the same test results as a 3-pack one.

See also  Mesh Wi-Fi System Explained: How to Best Use Multiple Broadcasters

Important note on speed test

Again, I measure the official test score by using just a single high-speed Wi-Fi client at a time. That’s the only way to figure out all broadcasters’ speeds consistently.

If I use multiple clients, due to different speed tiers, standards, and the fact Wi-Fi bandwidth is shared, it’s impossible to come up with relevant throughput numbers to say if this router is faster than the other, etc.

So, the scores reported in my reviews are likely those of the best-case scenario. It’s the total real-world bandwidth of the router’s Wi-Fi band in question.

Additionally, I also test all routers anecdotally with a dozen or so concurrent wireless and wired clients in various scenarios — over at least a few days to even a few weeks — to have a real sense of how good (or bad) it is compared with others.

Wi Fi Router USB NAS
A Wi-Fi router USB port can turn an external storage device into that of a network-attached storage server.

Router network-attached storage (NAS) speed test

If a router has a USB or eSATA port that can host a storage device, I test the performance of its NAS feature, too.

In this case, for consistence’s sake, the following is the standard way I go about it:

  • I use at least two random portable SSDs, out of this list, for the test and pick the one with higher consistent scores as the official numbers. So far, any portable SSD has proved to have way higher speeds than the router’s USB port. In other words, no matter what drive I used, the performance was almost the same for each router.
  • The drive is formatted in NTFS. If the router doesn’t support this file system, I’ll use the applicable one and note that in the review. So far, all routers I’ve reviewed support NTFS.
  • I perform the test using Windows 10’s File Explorer (a.ka. Windows Explorer). It’s a simple drag and drops copy test via the common Server Message Block (SMB, a.k.a. Samba) protocol with a 20GB single file as the data.
  • The test computer has the old and insecure SMBv1 disabled by default. This means the router must support SMBv2 for the test to work. (If SMBv1 must be enabled, I’ll note that in the review.)
  • I do the test using a wired Gigabit connection. If the router has a multi-gig port, I’ll use that port, too.
  • I perform each test (write and read) at least three times and pick the highest consistent numbers (within 5 percent) as the router’s scores.
  • I report NAS performance in megabytes per second (MB/s). (You can find the scores of all routers I’ve tested in this post of the best routers for with NAS feature.)

Note: The test for the router’s NAS feature is just for reference in terms of raw speeds. I myself use a real NAS server (and I’d always recommend that you do, too.) That said, I only use the router’s USB port for a short time and do not try all available applications or functions, nor do I check for conflicts or issues.

Wi-Fi and speed test: You can’t really put your finger on it.

After years of working with hundreds of routers, I have to admit that Wi-Fi testing can get repetitive and tedious. It’s also not 100% accurate. Also, no one can try every scenario, considering a router can have a lot of features and settings.

It’s impossible to say for sure how fast a router’s Wi-Fi is because there are so many factors and elements that can affect a test’s outcome. That’s not to mention the fact a router can change dramatically via firmware updates.

For this reason, I try to keep my testing as consistent as possible. In the end, my goal is to show how a particular Wi-Fi device does, compared with others.

In other words, my testing doesn’t mean to represent your experience of a router in terms of throughputs. Instead, it describes how better, or worse, a choice it is, among others, at the time of the review.

In a way, Wi-Fi is like red wine. The experience changes depending on when, how, and with whom you open a bottle. It’s tough to pinpoint what is what in the complex outcome. But over time, after so many bottles, chances are you’ll be able to tell how fine the wine is just by the nose.

No, I don’t smell any router — not purposely. But I can sense how good (or bad) a router is rather quickly. Still, I always take time to use it to make sure the assessment is fair and correct, within reason. (And there have been more than once that my initial hunch turned out incorrect.)

That said, the speed test is just one of many things I do to evaluate networking devices. As a rule, again, I always spend the time to have some real-life experience with them before publishing my reviews. You can count on that.

62 thoughts on “Internet or Wi-Fi Speed Test: How You Can Figure Out the Correct Numbers”

  1. Hi Dong, Wow! This post was awesome in helping me understand (as much as I’m able) the workings of the invisible internet! But I have a question that may or may not belong right here. Forgive me ahead of time if it doesn’t.
    If I have a VERY unstable connection for a long time, say over a year of constant loss of service, dropping repeatedly 25-30 times a day, could that damage a PC? Or laptop or tablet or even a phone that is or has been connected to the internet through my gateway? I ask because -Awful Connection and all of my machines and my phone are acting super hinky. Throwing security alerts at me when I try to go online, telling me my system(s) are compromised or are being attacked. No malware of any kind to be found. I had one computer worked over by two separate tech people from local computer shops with high ratings and also at Best Buy and none of them could find a thing. Also, the computer, phone or whatever I’m using works like a champ AWAY from home, but as soon as I come home and connect one to my gateway, it’s a mess from there. They all started to mess up within a week of each other, and I don’t have them connected to each other. I am at a loss and my ISP is a word I won’t say here.
    Thanks again for this post!

    Reply
  2. Hi Dong – I have 4 different wifi networks running (copper, 4g and 2 on Starlink (the Starlink + a wifi extender)) and wonder if there is an application that can test all 3 simultaneously or nearly simultaneously without requiring me to connect to each one, one by one in succession? It’s time consuming to do the latter and not accurate as in particular the 4g & Starlink vary a lot by time. Thanks ,Adrian

    Reply
      • Thanks Dong. What I have are 4 networks my computer can connect to:
        Starlink
        4g from Vodafone
        Copper line from BT Openreach
        And then an extender which extends whichever I choose

        I am struggling to test the speed of each effectively to see which I should ultimately select to focus on as I can’t really get enough data and am after a piece of software that can measure the down and upload speeds of all these networks systematically.

        Does that make better sense?

        Adrian

        Reply
        • That still doesn’t make sense, Adrain. I can’t read minds. You need to use the same langue as I do, so to speak. 🙂 Check out the posts on networking basics on the site and be specific.

          Reply
        • Why can’t you use speed test.net while connected to each one individually? If you want to test the quality of each one, you could use something like the Cisco WebEx test. Lastly, you could do a free demo for speedify which allows you to leverage all four connections at once.

          I’m really guessing at what you’re going for at this point though.

          Reply
          • Thanks, Marcus. I think I got what Adrian was trying to say now. So here go, A:

            1. It seems you have three Internet connections one of which you use via a Wi-Fi network. So you don’t have four “networks”. Check out this post to know how Wi-Fi and Internet are different.
            2. You can use one Internet connection at a time. So to test them all simultaneously, you need to use three devices. Or you can test them one at a time.
            3. It’s generally not possible to combine multiple Internet connections into one to increase bandwidth.

            Hope this answers your questions.

  3. Thank you for tagging this fantastic writeup. Very in-depth and thorough. I’ve been through a couple WiFi 6 routers now including the RT-AX86U and for the life of me, I cannot obtain over 600Mbps WiFi speeds but hard wired on all of them have been 950/950 so I’ve deduced that there must be external factors at play as both a WiFi 5 AC 160mhz laptop and S21 Ultra can’t break 600Mbps down.

    Reply
    • It’s likely the Jumbo Frame settings on the router and/MTU setting of your client, J. You can try turning on (or off) the Jumbo Frame on the router and see if that makes a difference. It’s in the LAN – Switch Control section within the web interface. Don’t mess around too much with these settings…

      Reply
  4. Hi Dong,

    I have ATT 1000 Mb fiber service. When I connect directly to the ATT Gateway via ethernet from a desktop, I regularly get speeds of 750-850 Mb for both upload and download. My normal setup is an RT-AC88U connected the gateway. When I test from the same desktop (same cable) through the router, I am getting 300-350 or so. I’ve tested several times of day and have tried to make sure nothing heavy is stealing bandwidth (streaming etc.). I have swapped the cable between the router and gateway a couple times (5E).

    It’s been about a year since I set all this up, and at the time I think I tested both ways but I am not 100% certain.

    Of course as you have pointed out in your other posts, I don’t really need that faster speed, but they advertised 300 Mb and 1000Mb for the same price when they rolled out the service so I thought why not. I am mostly curious if I have messed up a setting in the router.

    Thanks,

    Scott

    Reply
    • I would start by making sure that when you connect the router, that both sides so the port at 1000 full.

      Assuming your firmware is up to date, I would do a hard reset (NVRAM), reprovision it, and then do another hard reset and reprovision it. I know it sounds crazy, but there are plenty of reports out there about configurations not fully wiping out and I have experienced it myself.

      I’m sure Dong will have other suggestions but that’s where I would start.

      Reply
    • Likely you have the Jumbo Frame or MTU settings of the router and your Network card set in different values, Scott. Change the former in the router and see if that fixes it. The JF is in the LAN area of the web interface.

      Reply
  5. Thank you Dong for a great article and much detail. One question though. I upgraded my old router from 2.4 Ghz to a dual band 2.4 and 5Ghz. I see that my wired and wifi laptop are still getting the same upload/download speeds of about 30/4Mbps respectively before the upgrade. I also see on the router web interface that these laptops are set at 5Ghz.
    Why don’t I see speed increases? Should I see speed increases? The gateway is connected on DSL (with filter).

    Reply
  6. So I have two routers I am testing to make sure I have my new Asus AX89X setup right before I sell the Netgear RAX200. I have a gigabit internet connection. I read the posts about network setup and made sure both have 160Mhz bands enabled. I have two computers sending 1 video file thats over 2GB on a shared photo on laptop 2 with ethernet cable. Laptop 2 is downloading the file with a intel ax 200 wifi card and I got the exact same 113MB per second for both routers (They should be around 1400-1435Mbps right?). I’m pretty sure I’m bottlenecked somewhere because my internet is getting slightly speeds at 900-945Mbps with speedtest.net. Laptop one has a intel AX200 wifi 6 card. Laptop 2 has a gigabit ethernet port and cable. Both laptops have adequate M.2 NVME drives with at least 530MB read 230MB write. WAN shouldn’t be faster than LAN. What do you suggest I try next?

    Reply
    • Your setup yields 1Gbps at best, Nathan. The number you got was correct. Keep reading, you’ll find out why. My only advice is: Don’t make assumptions! 🙂

      Reply
      • Thanks for the advice. I definetely will keep reading and I thank you for taking the time. I just wanted replicate your results and 1435Mbps is quite alot different than 904Mbps considering its not using internet its all over the network. The only reason I care is because Im thinking about setuping a NAS in the future. Back to the forums. I’m gonna read everything.

        Reply
  7. Dong, i’m a big fan and have learned quite a bit from your articles.
    I have a question regarding 2 different ports in my home. I’m getting about ~90mbps and ~290mbps using SpeedTest with another port 10 feet away. I have fiber and a home run with a switch panel, and placing the router at the wall through the panel. I can’t figure it out, and I am wondering if this is simply a wall wiring issue that would reduce my signal so much.
    Appreciated in advance.

    Reply
      • Understood, but they are both in the same condo unit, and that would appear strange as both are next to each other in the switch panel and then 10 feet away across the living room floor layout. Could this be a wiring problem? Perhaps one of the eight wires affected? Thoughts?
        Thanks in advance.

        Reply
        • It could very well be. For example, I have a 30 foot extension that will only run at 10m. One of the pairs is bad. If you go to network and settings, then adapter properties, click on status. It will tell you what speed the port is running at. If it says 100mbps on the same computer and connected patchcord, it’s most likely a bad cable.

          Reply
  8. Great info here, Dong. This post is long and gets technical, so feel free to remove or edit. I’m posting it in case anyone needs to do a deeper dive into testing.

    For the more curious or advanced users or those not able to get results from nearby servers, there are some factors to consider. Most of these have been mentioned, but expanding a bit:

    Additional details regarding inconsistent or poor test results:

    1. Test hosting site server capacity. Stick to the major players like Speedtest, Google, etc. Most major ISP’s also have dedicated servers. (ATT, Comcast, Level 3, etc). Just Google them.

    2. Peering – Also related to the above, but different. This is what connects ISP to one another and others, such as Google, Facebook, Netflix, etc. When you traverse several peers to get to a server, it adds latency and unpredictability since your ISP is only responsible for its own peering links. If a server is on an ISP network with poorly managed or undersized peering links, results will be poor.

    3. Bandwidth does not equal throughput. Bandwidth is the size of the pipe. This is what you are really paying for. Throughput is the speed at which data travels through that pipe. For typical internet traffic, the further away the server is, the lesser the throughput will be. However, there are some tricks!

    4. While a single test on your connection may yield lower results, say 10m on a 100m connection, you should theoretically be able to fill the pipe with additional tests. So, I would expect to be able to do 7-9 additional tests and add them up to be in the 70-90mbps range.

    5. Your device. Older PC’s or phones with older network hardware, weak course, low memory or outdated drivers can be the culprit. At the very least, make sure all firmware, drivers, etc are up to date and test with the most modern PC you have. Borrow a neighbor’s, if needed!!!

    Test sites:

    1. Speedtest often has servers sitting on the ISP’s network. In this case, it will be the most accurate depiction of how well the ISP manages capacity, congestion, etc. Of course, the load on the test server can be a factor, but it is less of an issue today than in years past.

    2. Google’s test should yield nearly identical results. The reason for this is that Google has direct and robust peering with most major ISP’s, so there is typically only an insignificant amount of additional latency.

    3. When you hit a server outside of the ISP network, but is a know reliable test site, there is still the chance of traversing bad peering links.

    Regardless, a couple of things can be done if you’re confident you are testing to far away servers that are reputable.

    1. Perform multiple simultaneous tests. For example, Google, speedtest, Comcast, att, or whatever major isp in your area. It takes a bit of quick timing but is possible. You can also use more than one pc, since cpu utilization can be a factor. Add them up and you should see that you’re able to fill the pipe with multiple streams of data.

    2. Good old-fashioned download. Ubuntu is available for download at servers around the world. Just Google “Ubuntu mirror launchpad”. Many colleges and government entities host the files. The mirror/download site list also tells you what internet bandwidth the host has. Go for the largest ones. Then, do multiple simultaneous downloads from one site and do the math. You can also do this with multiple sites, since the server might throttle multiple requests from the same user (you). There are calculators on the web that can perform the conversion.

    I probably missed some things, but I think I’ve said enough! Happy testing!

    Marcus

    Reply
      • My pleasure. I currently work for an ISP and have worked for several major ISP’s enterprise divisions over the past 20 years. I’ve held roles as a technician, network engineer, data center engineer, sales engineer and now sales leadership. Always happy to contribute!

        Reply
  9. Is there a reason why fast.com seems to be “slower” in test results than speedtest.net?
    Also, I find the Google speed test (Mlab) seems closer to speedtest than fast.
    https://www.google.com/search?q=internet+speed+test
    What I like about speedtest is the plotted results over the duration of the test which can show burst, “shaping” and other performance or throttling intervention.

    Reply
      • Testmy.net results show abysmal speed when speedtest.net shows more than contracted speed. Fast speed little lower within contracted speed

        Reply
        • That’s because Testmy.net doesn’t have a server close to where you are Aiyer. Use Speedtest.net. It has the best selection of servers. Or just use the test of Google.

          Reply
    • Think of all the website speed test this way, use them all, which ever the fastest is the most accurate. Where ever you are the speed test server is not your ISP and your ISP only responsible between you and them and whatever ISP or the tier network they are directly connected to.

      If you are driving from your house to a grocery store. Your ISP own the street that your house is on and if there is a construction on one of the street between your street and your favorite grocery store street are on then there is nothing you or you ISP can do about it. But if you go to a different grocery store about the same distance or even a little farther then your speed would improve. But in term of speed, generally sparking the best grocery store would be the one closest to you, all the speed test websites follow the same logic assuming there is nothing else wrong in between you and them.

      Reply
  10. Dear Ngo,

    I live in Singapore.

    When the speed test using ISP server states the D/L and U/L speeds to be
    70 & 50 Mbps, the Wi-Fi speed per Windows 10 shows only, D/L and U/L
    speeds like 0.00X and 0.00Y Mbps. Why such a vast variation?

    It appears that the speed test data is of no relevance in actual upload and downloads. Is it the case?

    I called the ISP tech support numerous times, but they have no answer.

    Kindly share your knowledge.

    Thanks for your contributions.

    I wish you the best of everything in life.

    Reply
    • Wi-Fi has nothing to do with an operating system, Jay. It seems like there’s something wrong with your Windows 10 machine you used for the testing if I understood you correctly. Try picking a test server near you?

      Reply
  11. Have a google mesh setup currently and 1000mb fiber connection. Looking to upgrade to wifi 6 debating a mesh or the asus rtax86u. I’m not even sure mesh is necessary for me in 1600sq ft home but I really just want to get closer to the speeds I pay for wirelessly. Running some wifi6 and wifi5 devices about 20 devices total. What mesh would you recommend? Or would you say the asus is a better bet. Thanks

    Reply
  12. Hello Dong,
    I hope you can point me in the right direction once again. In preparation for installing a new AX86u I ran some benchmark tests with my old AC86u.
    I checked the speed with Ookla and timed the download of a file from my NAS. I did the same tests with my old samsung tablet with N wifi and also with my new S6 Lite tablet that has AC wifi. The old tablet was twice as fast in all tests. When I look at the Network properties in my setting on each device the new tab shows a network speed of 72Mbps and the old tab shows 150Mbps.
    Nothing seems right to me. Is there anything you can point me to?
    Thanks
    Darryl

    Reply
  13. Hi Dong,

    Your site is very informative especially to non-tech guys like me. Maybe you can help me with my issue. I just moved to a new gigabit provider. I also had a gigabit service from the previous one. I have a Nighthawk RA200 AX11000 router. Instead of a cable modem (Netgear CM1000), i now have a Fiber ONT that I connect my router. What I am having issue with is that I only get half the download speed on wifi now as compared to my previous provider. I understand that previously, i get an asymmetric service as compared to symmetric right now. Prior to moving to the new provider, i clocked my download speed at 645mbps, up 9.8mbps. After the install of the new provider, download 277mbps, upload 488mbps. Is there something I need to configure on my router to at least bring up the download speed?
    Thanks!

    Reply
    • Assuming your ISP delivers, Bim, (you can check to make sure by doing a proper test directly from the ONT), I’d recommended a couple of things:

      1. Update your router to the latest firmware.
      2. Reset it to default factory settings and set it up from scratch. There might have been some settings applicable only to the previous internet plan.
      3. Use the router’s 2.5 Gbps port as it WAN port.

      Reply
  14. Hi Dong,
    I think I was testing my speed incorrectly before when I was getting some wonky real low numbers. I was always selecting the server from my ISP rather than letting Ookla select automatically the server with a low ping.

    Do I understand correctly that I should let Ookla select automatically?

    Thanks
    Darryl

    Reply
  15. What exactly is a gigabit network port? I want to test my wifi speed using my MacBook Pro (13-inch, 2017, Two Thunderbolt 3 ports)…I have an adapter that will let me plug an Ethernet cable into my MacBook. Is that sufficient enough for to test (somewhat accurately) my wifi speed?

    Reply
  16. Howdy! Do you know if they make any plugins to help with Search Engine Optimization? I’m trying to get my blog to rank for some targeted keywords but
    I’m not seeing very good results. If you know of any please share.
    Thanks!

    Reply
  17. In your view which 5 GHz channel do you find to offer the best performance? I find that Channel 52 and 100 offer superior performance to the others…

    Reply
    • This depends on the WiFi standard, the time of day and especially where you are. The two channels you mentioned are both DFS and might not be available in certain places.

      Reply
  18. Hi Dong,
    A Vietnamese news website copies your post without introducing the source. Just FYI, in case you’re not aware.

    Reply
  19. Hi Dong, I have an Aimesh with an AC86U as my main router. How do I make the 5GHz band work only with 802.11ac clients and the 2.4GHz band just for 802.11n clients? Thanks!

    Reply
    • You need to turn off the SmartConnect mode (It’s in Wireless section under Advanced Settings) then change the Wireless Mode for each band accordingly. You can choose to name the two bands differently or the same (in this case make sure the password — WPA Pre-Shared Key — is the same too.)

      Reply
  20. I would like to know what 5GHz channel that you are using? Asus (exp. AC86U) is well known as working well in channel 149~161 and Netgear (exp XR500) works well in channel 36~48. Also, what region did you set to those router in Wi-Fi setup (AC86U great in Australia region and XR500 great in US region )? This make major changes as many reviewers use these channels and region trick to “promote” some sponsored routers by putting it to the optimized position to make it looks great (Australia AC86U in channel 149~161 wins XR500 and US XR500 in channel 36~48 wins AC86U) , too tricky in this industry, I hope I can finally find the trustworthy site which disclose it all for the testing setup, thanks

    Reply
    • Hi there,

      Thanks for your input. I actually test routers using ALL channels, one after another, as well as using the Auto setting, which is generally the default. I report the highest scores.

      Hope this helps, 🙂

      -Dong.

      Reply

Leave a Comment

Appreciate the content? ☕ Let Dong know via a Ko-Fi ❣️