Thank you for being a valued part of the CNET community. As of December 1, 2020, the forums are in read-only format. In early 2021, CNET Forums will no longer be available. We are grateful for the participation and advice you have provided to one another over the years.

Thanks,

CNET Support

General discussion

Are internet speed connection tests rigged?

Aug 5, 2016 5:50PM PDT

I have a question regarding internet connection speed. Are the so-called "speed checks" rigged? I ask this because recently I received an email from CNET that contained a solicitation inviting you to check your connection speed. Well, I did and my result averaged around 6Mbps. My ISP is AT&T, which boasts its U-Verse is superior. I'm subscribed to their 12Mbps plan. So after I learned that the solicitation link clocked me at 6, I tested my connection on U-Verse's speed check which shows me clocked at around 12Mbps download and around 3 upload, which is fine but makes me wonder which is right? I've always kind of wondered what the actual speed is when at times I find myself waiting patiently staring at a screen as the page loads. Any thoughts or previous discussions on this subject? I'm wondering if speed is just a marketing scheme to entice one believing their connection is better than their competitor. And is there a trustworthy method of finding out my actual connection speed without bias? Thanks.

--Submitted by Alan L. of Columbus, Ohio

Post was last edited on August 5, 2016 5:52 PM PDT

Discussion is locked

- Collapse -
ISP speed is only from your terminal to their router
Aug 12, 2016 8:28PM PDT

The toughest challenge in Internet connectivity today is that "Last Mile" from the ISP's Point-Of-Presence to the subscriber location.

They can run multi-gigabit fiber optic lines into your neighbourhood, but the last bit from their utility office or underground vault to your office/residence is the most difficult and that's where the cable modems and DSL modems come in.

This is, of course, notwithstanding those few lucky neighbourhoods where they have "fibre-to-the-home" service but that's rare.

The stated bandwidths are only between your ISP supplied terminal/modem to their router (a.k.a. "default gateway").

Beyond that there's all sorts of issues that network engineers have to deal with like congestion, packet loss, jitter, etc. across multiple router hops on the internet.

And this is notwithstanding traffic loads and bandwidth at the server at the other end.

Usually ISPs who run speed tester servers will have high bandwidth within their own network so if you subscribe to 25 mbps download you should contractually get that speed to the ISP's network but beyond that all bets are off, as they say.

- Collapse -
Yes, But
Aug 13, 2016 10:01AM PDT

I think I know what you are saying, but some of the terms are unclear and a bit confusing. I have a cable modem and a separate router/wifi in my house. NEITHER belongs to the cable company (ISP). They are both mine. My PC default gateway is pointing at my router. Here is what I'd expect to see:

1. LAN - if I had a server on my home network running the server component of SPEEDTEST.
2. ISP Test - This would run on a SERVER on the cable company's network (not really a router).
3. Internet test - A server somewhere out on the Internet.

1. LAN test - I would expect to see close to 1 GB/sec download and upload (wired) or about 2.3 GB/sec WiFi
2. IP TEST - The maximum should be whatever the ISP sold you BUT that would depend on traffic on their network, such as your neighbors streaming.
3. Internet - Will vary but the maximum speed could be as high as the ISP sold you.

You are correct that this will show three varied results, and those can change depending on network traffic. It is hard to come across an "ideal" test environment.

- Collapse -
Different Dell Laptops
Aug 12, 2016 8:41PM PDT

The speeds on my laptop are very similar, no matter the test I use. That being said, the wireless speed on my laptop is 14-18Mbps. Either my daughter or my wife's Dell laptop is consistently 10-15Mbps faster. I assume I just have a bad wireless device (yes, the driver is up to date). Any suggestions? BTW, my wired speed is 45Mbps. Thanks in advance.

- Collapse -
Variables
Aug 13, 2016 10:18AM PDT

You do have to take into account a lot of factors. I'm going to guess that, if your "wired" speed is 45 mbps, that your ISP sold you about 50 mbps. WiFi could be affected by other 2.4 Ghz or 5 Ghz devices on your network, such as wireless phones. Also, some users have more than one WiFi router/repeater. You could be correct that it could be differences in wireless devices and, in addition to the drivers, I'd also update firmware in the router and, if possible, that of the devices. Lastly, there are other issues. What is running in the background? Virus software, indexing and most background software will change your results so you may not be comparing apples to apples. Then there is the TCP/IP stack. Are both machines on the same version of their OS and patch level? Are you running battery or wall power? Some power saving schemes reduce power to the WiFi transmitter (is why I asked).
These speed tests are good if you are making changes to a computer and want to see the effect on download speeds but the number of things affecting the results can be large. In WiFi, even moving to a different room can affect the results (number of retries), so it is important to try to keep other things the same. If the two laptops are different, that can also have an affect. It may be that the test, itself, is running faster on one machine. Good luck. Oh, something you might try... get a wireless USB card and use that on both machines for the test. And decide, in advance if you are testing 2.4 GHZ or 5.0 GHZ (or both) and check that both machines are on the same WiFi channel, especially if you have a lot of neighbors on the same channel.

- Collapse -
many factors involved
Aug 12, 2016 9:28PM PDT

Others could give you more technical details about what can affect your actual internet speed, but it often comes down to routing. Unless you are very close to the server of the site/service you are accessing, your computer is going through many intermediate sites as it tries to connect to the site you have requested, and if there is a bottleneck due to high traffic, software/hardware issues, weather or something else at one or more of those points, you are either going to have a slow-loading site or it may not load at all.

One good way to know if you are actually getting the speed your ISP promises is to download one or more largish files and calculate your total download/upload speed. Keep in mind if you are uploading a lot, you likely won't be able to get maximum download speed until you reduce your upload. I am able to download at 1.1 MB/s, which corresponds to my promised bandwidth. I can generally send/receive email and browse the internet decently at the same time, though streaming high def video may be a little choppy if my file transfers are close to max.

If many websites are loading slowly for you at various times of the day, there may be a problem with either your system, your modem (if you use one), or your ISP. A modem is fairly easy to test, especially if your ISP has decent customer service. A bad cable can be trickier to pin down.

Good luck tracking this down. Hopefully you don't have a dishonest/poor ISP.

- Collapse -
It's complicated
Aug 12, 2016 10:32PM PDT

I live in a rural area in California but have Xfinity/Comcast and have those 80+ down speeds a lot of the time regardless of whose test I use. But when the internet is likely to be busiest, I test the speeds using servers in Boston and Florida and lo and behold at times the speed drops substantially, no matter what the local tests show. (I'm retired and sometimes bored.)

For me, the issue is do I get the speed I need 95% of the time. I do. So what about the other 5%? Since I'm accessing some web site's server system, if I don't get the speed is it my service or their problem?

When I check that out by testing speeds from test servers in the locale of the web site's operation, I frequently learn the web is busy somewhere between them and me. Also pings to the specific servers sometimes indicate they've got problems with their operation.

Speed tests rigged? I haven't seen that unless they are all rigged the same. But when things are not good and I run an Xfinity speed test, sometimes the third time I run it speeds do mysteriously improve.

- Collapse -
Rigged Speed Tests?
Aug 13, 2016 5:07AM PDT

My suggestion is to go to http://speedtest.net to check your internet connection speed. Cnet's test gave me the same results as yours; about 6.5 mbps, but speedtest.net showed it to be 30 mbps which I know to be closer because I often download large files from usenet and my newsreader reports the same download speeds ( in the range of 25 to 30 mbps).

- Collapse -
Internet Speed by Cox in Phoenix, Arizona
Aug 13, 2016 7:28AM PDT

I just did a test a few minutes ago and used a 3rd party test to perform it. With Cox in Phoenix, My Download speed with my new HP Desktop was 58 Mbps and upload speed with 32 Mbps and the ping was 41. This is not the actual internet speed since it will measure differently on my Android tablet devices, but it is fast enough to give me a speed that allows me to watch UHD on my TV while watching You Tube and Netflix on my TV which comes in via my Router. It costs about $80 a month, but is well worth the cost.

- Collapse -
Does ISP Speed Matter?
Aug 13, 2016 8:08AM PDT

Isn't the price everyone pays for internet service based on how quickly input to and output from the internet appears to you at your computer? That is the response time. The only control over response time is the price you pay the ISP. But your ISP plays only one part of the whole response time picture.
ISP speed is only one of a great number of variables that go towards affecting how long a user waits from when the enter key is pressed until the response is received at the computer. If the requested destination server is overloaded with requests, that will affect when your ISP can use its speed from and to you. Also, the number of networks traversed to get from you to and from a destination server, e.g., Amazon, must be considered. The "hops" between networks mentioned this week will also each add some delay the your response. All of the other variables identified in this week's responses (and probably others) will constitute the response time and not only that, those times will differ for every entry. Your ISP provides one piece of the response time puzzle, charges you for that piece and attempts to show you (accurately, consistently and meaningfully, one hopes) an indication of how their services' prices are justified. This is like auto MPG numbers which are relative to each other, are determined under standard conditions (one hopes) but can never be matched, exactly, by anyone due to other, uncontrolled variables. The same is true with internet response time; in total it depends on controllable, i.e., ISP, and uncontrolled variables! ISP internet connection speeds may be interesting for those technically inclined but basically are neither the whole problem or solution in themselves, the user's response is what users are, or should be looking for.

- Collapse -
Routing............
Aug 13, 2016 11:54AM PDT

Your ISP measures only the speed from your computer to the ISP..........no passing through other routes. On the net, you will be passing through many routes after your data leaves your ISP. The more it is directed, it is likely to pass through a busy or congested router. Your speed is only as fast as the slowest link encountered along the way.

- Collapse -
......and
Aug 13, 2016 11:57AM PDT

...................your ISP ONLY lists speed to their servers from your machine, which is almost always a direct link. Beyond that, anything goes.

- Collapse -
DSLReports
Aug 13, 2016 12:20PM PDT

Also agree re DSLReports.
COMCAST isp. El Cerrito, CA

- Collapse -
Ookla's speed test
Aug 14, 2016 5:21AM PDT

I use http://www.speedtest.net. It's not biased and has different servers you can test from. I've been on U-Verse, Charter, and COX. AT&T will not be a real competitor until they connect a fiber line directly to a house. Charter might be the worst tv provider, but offers faster broadband speeds for less than the others.

- Collapse -
Cable's Marketing Schemes - Part 1
Aug 15, 2016 4:31PM PDT

- Big Island of Hawaii, we are on a TWC/Oceanic/RoadRunner 300/20 Mbps Ultimate cable plan, that was automatically upgraded from 200/20 "Bundled" Package.
Realistically, we do not get 300 Mbps Speeds, period. There are just way too many Skips and Hop's between here and Honolulu.
From 2014 to 2016, we have had their cable tech's so many times, that we have all become good friends, and hang-out.
The Tech's are very skilled with the equipment, cable runs, connectors, and checking speed/bandwidth, BUT, it stops there; as they have absolutely No Control over the "System".
1. TWC does have a dual-system in Hawaii, with Coaxial + Fiber optic.
2. On the BI, only Commercial, Government, and wealthy clients have access to FIBER.
3. There is a Fiber Node, just 1 block away, but we will probably never get access to it for another 10 or more years.
4. Unless GOOGLE Fiber shows up.

- Collapse -
Any chance of you installing a SQUID box?
Aug 15, 2016 4:40PM PDT
- Collapse -
How TWC cable get's artificial higher Mbps ... Part 2
Aug 15, 2016 5:45PM PDT

1. They have a dual-system in place - Coaxial + Fiber Optic.
2. Fiber service is restricted only for Commercial, Education, Government and for the 1%-er's.
3. TWC has set-up their own "speed-test' UI to run thru FIBER. This of course gives the "impression" that you are not crazy and are getting their 300/20 Ultimate speeds. No, not really.
4. Ookla has been the best App to use and up until recently TWC was 1 of the 4 Hosts.
5. TWC has now dropped off the Ookla menu.
6. TWC is still "Marketing" their 300/20 plan.
7. It's called the 300/20 Plan, Because you will PAY $300 + Fee's + Taxes, Per Month.
8. The ONLY WAY for us to get 300/20 cable service is to move to Honolulu and live right next door to their Node, or Magically get Fiber service.

HOW can some of you Help? We would love to go off Cable all-together, so any Real, Feasible and Proven Cable-less Internet SOLUTIONS, would be much appreciated.
And please do not provie any suggestions that will be a WOT on these ISLANDS.
TWC has a Monopoly currently here. GOOGLE could make a huge difference.
As a comparison, Korea & Japan offer Fiber Internet Services, with unbelievable Bandwidth and Speeds.
In Korea, you can get a 1000/500 Mbps package for like 25 bucks!

- Collapse -
Time it Yourself
Aug 17, 2016 5:09AM PDT

The only way to really know is to download a file of a known size and time it for yourself.