How can speed test take longer/shorter if speed is same?

[THAIPHUKET]

Logic would say if the measured speed is 10 MB the time to take the measurement should be same.

In fact, sometimes the measurement seems to take forever until is spit out the lottery number.

Nix verstehen

[OldChinaHam]

The distance between two points, your computer and the server which you are connecting to for your speed test, is never the same.

That is one reason.

Fluctuations in line quality from one moment to the next.

Number of users on the net.

Etc, etc, etc.

I am no network engineer.

[Tywais]

One way the tests are run is to send a few small groups of data for initial calibration to determine the optimum size for the real test. That will show as an additional real time as opposed to the measured time. Then one or more large groups are sent, usually via multiple simultaneous connections, to the computer and the time measured to arrive. There can be a turn around time between each group and preparation for the next group. Some are incrementing block sizes, for example 100k, 500k, 1M, etc. and overhead in preparing the next group plus calculating than averaging them will add to the overall real time.

That is just one of many possible reasons for it.

//edit - also there is the possibility of packet losses in the transmission of the data and has to be resent adding even more time to the real time (stop watch) measurement.

[OldChinaHam]

Which site, or ISP, are you using to conduct your speed test?

Is it located in Thailand? Outside Thailand?

Most ISP's will offer "dummy" files which you can download and measure the time it requires for a specific size.

Try a 100 MByte size download and measure the time.

I can't really do this with any reliability because I am on a shared network. Sometimes I will get 15 Mbps, and sometimes 800 Kbps.

If you really want to know what your effective speed is, why not download a 100 GB size file over the weekend and measure how long it takes to download.

Probably you cannot find one. So then you might download a 4.4GB size file from some site like Opensuse which has very fast servers, usually. And then repeat 5 or 6 times during any 24 hour period. Then you will know if you have a good line or not.

If you have your own server in the US or Europe, then you can use FTP to download a 10 GB file using Filezila or something.

I often do this to see what the line I am using is really worth.

Usually in CM, I learn it is pretty worthless, though.

And, when I ask the people who operate the equipment that my computer is connected to, WHY IS THE SPEED DIFFERENT each day?

They never provide me with the same answer twice.

[Tywais]

Believe you are missing his question OCH. He is wondering why the reported speedtest info doesn't match up with the real (physical wait time). That is holding up a stop watch and start/stop it they don't match up.

[OldChinaHam]

When you do do your speed test.

Please make sure you have nothing else running, no apps, no surfing, no nothing.

Only your Filezilla, or your browser connected to the download site, such as Opensuse for the 4.4 GB OS test download.

Sometimes a computer will be overtaxed with apps and slow, or sometimes you might have a virus that is uploading or downloading in the background.

(If you should wish to try Filezila to download a huge file over a 24 hour period, then you can get the opensource software here: https://filezilla-project.org/ But I am sorry, i cannot list my server here for you to try. Try Opensuse instead.)

[OldChinaHam]

Believe you are missing his question OCH. He is wondering why the reported speedtest info doesn't match up with the real (physical wait time). That is holding up a stop watch and start/stop it they don't match up.

Thank you, I see. But I must have tested my line speed a million times, in a hundred ways during the past 13 years, and I have never thought using the speed "dial" or numbers on the website of the ISP or speedtest service to be reliable, due to lag time of the browser, or not knowing when the test actually begins.

So this is why I always use a method where I can begin the download at will, and then I can use download software to watch the file downloading, and watch the speed fluctuate by the minute or second along the x-axis, and finally I know exactly when it has stopped downloading.

Normally, an ISP will allow you to pick a files size, such as 1MB or 10MB for downloading. Then just take out your "stopwatch" and time it.

If the file size is too small, then you will not get any accurate speed. (Maybe this is why the "meter" method is always wacko, they use a filesize which is too small to be valid.)

But I always prefer using bigger files which can tell me what the line speed is over a longer period of time. For this, a 10GB file is good because this usually requires an hour or two to download if you have a decent internet connect.

Hope I am not still missing the point.

BUt this IS a very interesting question, one which we have "tested for" as consumers so many times, fearing the ISPs are giving us the "run-around".

[Pib]

Also keep in mind that different speed test programs/sites use different algorithms and some results are based on downloading/uploading files of different sizes vs just one standard file size. Downloading a BIG file is probably a good test to see how your internet plan works with torrents/download managers, but maybe not a good test if you just browse/surf the internet a lot where each webpage usually consists of many smaller files/links to other servers.

[THAIPHUKET]

You read me correctly Tywais, the stopwatch example says it perfectly. Background is I have wild swings in my TOT connection from 2,3 to 11,5 , eg. TOT is my permanent house guest. I use and I can only use a yardstick that TOT understands, difficult enough as it is, which is http://speedtest.adslthailand.com/ . Test always against BKK for sake of consistency and to cut off any excuses : But that´s abroad.