A true yard stick for Internet speed ? Testmy.net

[THAIPHUKET]

Testmy.net claims to different, convincing is you can select the test data quantity. As you can see I used up to 100MB test download

My equipment

100 Mbps 3 BBB line

Windows 10 up to date

Browser Chrome

Asus UX 305

Hosts Singapore Host can be changed but only internationally

Question what does the test tell about 3 BBB service?

[muratremix]

It means you don't know how to test your speed properly.

If you noticed, your speed maxed out at 17.8 Mbits on a single connection.

You need to choose 100-200 or 250Mb manual test file. If you leave it at auto, it will give you low results.

[muratremix]

Wireless connection could be also limiting factor. Try do it with ethernet cable connected and using a bigger test file (250mb)

[KhunBENQ]

I get a value in the mid 20s to Singapore whether I use 50 MB download or use 200 MB download.

I have ToT fibre, 30 / 3 and best conditions (connected via cable, WiFi switched off, papa alone on the net).

To the OP: 7 Mbit/s to Singapore is "so la la".

To get more insight use a real overseas target.

Just did 200 MB to San Jose (please don't ask me why they use those silly (smiley) symbols ).

Edited May 9, 2016 by KhunBENQ

[THAIPHUKET]

Very helpful. That kind of feedback I was hoping for.

[akampa]

I use Testmy.net and find it quit accurate here is a screenshot which I did I with Testmy.net multitest also use 3BB but only have 15mps the second is with speedtest .net but only to the U SA .

Edited May 9, 2016 by akampa

[KhunBENQ]

From the OP:

Hosts Singapore Host can be changed but only internationally

Why "only"?

How much do you care about lightning speed to some Bangkok test server while most of the relevant content comes from out of the country.

Singapore and Hongkong being very important for Google, MS, FB and other big players.

And of course content coming directly from your home country (less than you might think).

"content delivery network" is the term to google.

https://en.wikipedia.org/wiki/Content_delivery_network

Edited May 9, 2016 by KhunBENQ

[THAIPHUKET]

From the OP:

Hosts Singapore Host can be changed but only internationally

Why "only"?

How much do you care about lightning speed to some Bangkok test server while most of the relevant content comes from out of the country.

Singapore and Hongkong being very important for Google, MS, FB and other big players.

And of course content coming directly from your home country (less than you might think).

"content delivery network" is the term to google.

https://en.wikipedia.org/wiki/Content_delivery_network

YOU are very right, I should not care about BKK speed. Except I am trying to figure out what the 3BBB speed in Thailand should be before that speed is throttled down by the CAT ???? international gateway. Which would be beyond 3BBBs control. If I understand so far correctly.

In other words What is the relevance of 3BBB speed chart? Is the 100Mbps line worht the money? I cannot see a difference in loading web sites

Secondly

To quote MURATREMIX = You need to choose 100-200 or 250Mb manual test file. If you leave it at auto, it will give you low results.

What good are OOKLA et altera at all?

[KhunBENQ]

In other words What is the relevance of 3BBB speed chart? Is the 100Mbps line worht the money? I cannot see a difference in loading web sites

I see you are on the path to enlightenment, asking the right questions

[THAIPHUKET]

Ok, I try again. Here my new download with a package of 200MB from Chiang Mai against Singapore multiple sources

What does the BBB chart tell you. Erratic or a stable download? How acceptable ?

Better phrased ? :-)

[KhunBENQ]

Just did a download test from Singapore with 100 MB.

In the graphics I added a red line at 30 MBit/s which is my "paid package" rate (ToT fibre, 30 / 3).

You can guess that the variations are from ToT throttling/capping back to 30.

Kind of a poor control loop.

Your result shows a very stable but lower rate.

Don't ask me "what is better".

17 Mbit/s is still OK for the typical "home use".

In the meantime I understand who really needs these high rates like 50+?

It's the heavy cloud users who have most of their data in the cloud.

BUT: who has (or would like to use) a Thailand based cloud for his data

So anytime also these users depend on international bandwidth.

Edited May 10, 2016 by KhunBENQ

[muratremix]

Ok, I try again. Here my new download with a package of 200MB from Chiang Mai against Singapore multiple sources

What does the BBB chart tell you. Erratic or a stable download? How acceptable ?

Better phrased ? :-)

Dude. If you are hard-capped at this speed, either 3BB is capping you, which is not unexpected because Chiang Mai - Bangkok domestic fiber connections are limited and oversaturated.

Or, you are limited by wifi speed and wifi signal noise. Did you test this speedtest with ethernet connection and disabling wifi adapter?

[KhunBENQ]

Worth repeating that I do my tests under ideal conditions.

All devices except PC off the net.

WiFi off (cutting the soap opera viewers :>).

Connected via LAN cable (always).

Edited May 10, 2016 by KhunBENQ

[wpcoe]

I like testmy.net speed test, but wish they would do burst and multi-threaded tests like ThinkBroadband does.

e.g.

[KhunBENQ]

Can you comment on this type of test?

Too lazy to research.

Mine:

[Pib]

I like testmy.net speed test, but wish they would do burst and multi-threaded tests like ThinkBroadband does.

e.g.

You can set testmy.net to Multthreaded or Linear (one thread) testing. Setting is in upper right hand corner of page.

[Pib]

DSLReports is another speedtester a person might want to try. You can select whether you are using Fiber Optics, Cable/DOCSIS, DSL, or other type plans. Below are my results at 7:10pm Bangkok time with a VPN connection to San Francisco.

I'm on a True 15Mb down/1.5Mb down Cable/DOCSIS plan (which has bursting capability) here in Bangkok using StrongVPN service (using a L2TP connection).

So many speedtesters all testing with their proprietary program/algorithm and different servers...a person can get all kinds of results when testing among them. One might give you slow results and the other one fast results.

[wpcoe]

Thanks, Pib, for those two pointers. Using burst testing on testmy.net for my 30/3 True DOCSIS line on a Wednesday mid-day:

The first four (good) results were to Singapore and the last (not-so-good) result was to San Francisco.

I need to spend some time understanding the DSLReports test. For instance: what destination do they use? The cith with lowest latency? What is "Buffer Bloat"?

Not using VPN I got 32.4Mbps / 3.23Mbps, using Strong VPN in SFO I got 12.71Mbps / 2.84Mbps and for both tests got an "F" for Buffer Bloat.

[Pib]

Yea, it's understandable you got a better result/faster speed to nearby Singapore due to a more direct an connection. The destinations and latency to each destination is shown while the test is running, like this snapshot I did below of a test in progress. Bufferbloat is a function of latency/distance to a test server....shows the amount of buffering/bunching-up of data along the way...usually the farther away a server is the higher the buffering/jitter/bufferbloat. Where you got a score of F, if you have been testing to the Moon your score would have probably be a Z. Clicking the "gear" icon you can change/select the destination(s) tested to...and then Reset to Default after to just allow the test to automatically select various destinations.

[wpcoe]

I knew it showed latency to five servers, but how do you know to which server the speed testing is conducted by default, if none is specified by clicking on the gear icon? Or, is it being tested to all five locations and then somehow compiled to a single result?

Boy, talk about assumptions. I assumed the F was a Failing score, as in my schoolboy days of grades A, B, C, D, F! Never crossed my mind that it could have even meant Fair, much less a long scale of A-to-Z. 555+

[Pib]

It's a multistream/multithread tester and unless you pick a specific server(s) to test to by going into the Settings (the Gear icon) it will automatically test to a variety of locations using multiple streams/threads.

In the download and upload windows you will see a number next to a cloud....that number represents the number streams/threads being used during the test. After the test is complete you can these click the Results + Share icon to view the test Run Log which shows the locations/IP addresses the multistreams reached out to in order to run the test.

In the Settings you can also set the number of streams from 1 to 32. And remember, just press the Reset to Default icon in the Settings area to go back to default/auto settings.

It's really a versatile speed tester since you can adjust locations "and how my streams" to test.

They have another neat tester for "video streaming" at this webpage since so many people complain about pausing/poor video streaming.

http://www.dslreports.com/tools/streamtest

[Pib]

I knew it showed latency to five servers, but how do you know to which server the speed testing is conducted by default, if none is specified by clicking on the gear icon? Or, is it being tested to all five locations and then somehow compiled to a single result?

Boy, talk about assumptions. I assumed the F was a Failing score, as in my schoolboy days of grades A, B, C, D, F! Never crossed my mind that it could have even meant Fair, much less a long scale of A-to-Z. 555+

For the test to the Moon I'm sure they would have expanded the grading scale all the way to Z.

You know, I've had VOIP service between the Thailand and U.S. since retiring here around 8 years ago and I've had 3 different U.S. VOIP providers...they all provided good service...I only switched around due to pricing (I'm currently with NetTalk) and when talking to friends/relatives in the U.S. like on the U.S. East Coast which is a half a world/12 times zones away my VOIC calls are clear as a bell...sounds like they are just across the soi...and I sound the same to them on the other end. However, every speed test I run which also gives a VOIP score my VOIP score was like a D or F....but like I said my calls are just fine...just like I'm calling across the soi.

I take all speedtests with a big grain of salt.

[THAIPHUKET]

In short, no way to objectively say if you are getting 100% from 3BB for your money assuming oily domestic data transfer. Correct?

[Pib]

If you are getting full advertised "domestic" speed which can be easily & quickly confirmed by testing to a local server like using the 3BB Speedtester or variety of other speedtesters using in Thailand servers, then you are getting what you paid for. Your international speed will be "somewhere between zero and your advertised domestic speed....and just because a person say upgrades his package to twice the advertised speed that does not mean his international speed will increase two fold...in fact, it may not increase at all.

[THAIPHUKET]

So it is all bad-mouthing that these tester produce overrated results, kept in cache . In other words the test are real-time?

[Pib]

So it is all bad-mouthing that these tester produce overrated results, kept in cache . In other words the test are real-time?

No in not all bad-mouthing as some testers do produce overrated results because they get fooled into pulling test files from "local cache servers."

[THAIPHUKET]

at 12 am , testmy.net Chiang Mai against Singapore, 200MB pachage , astonishing download speed of 58Mbps !

[mahtin]

Another convention sometimes used is that, since 12 noon is by definition neither ante meridiem (before noon) nor post meridiem (after noon), then 12 a.m. refers to midnight at the start of the specified day (00:00) and 12 p.m. to midnight at the end of that day (24:00).

Not that this thread is in any way confusing...

[Jdiddy]

So i just did a speed test on my capped mobile, ran out of data yesterday should be maxing out at 384kbit so dunno quite whats going on lol

http://www.speedtest.net/my-result/a/1942601717

[Pib]

Just as crossfeed I very recently bought a PureVPN two year subscription for $2.08/month. Their speed is excellent, but their connection app sucks/is buggy at least on my computer so I have stopped using it. I have just setup Windows manual VPN connections which is easy to do and don't use the app.

Anyway, plus or minus a few minutes of 4pm today Thailand time I did a "Combined" test (i.e, auto download & upload test) and below are the results for my True DOCSIS 15Mb/1.5Mb. Yes, my download speed exceeds my plan's basic download speed of 15Mb but that is due to DOCSIS "bursting" which allows for higher short term download speeds which usually last longer than the speedtest...bursting usually does not increase upload speed. Bursting capability is great for browsing since you are just spending a brief time to load each webpage, but bursting does not help with live video streaming. My connection to Singapore is smoking fast (30.3Mb) when I ran it...did the test twice with basically the same speed...and the connections speed to LA and SF were 19Mb and 16Mb, respectively. See test results snapshot below.

VPN PPTP Connection to Singapore and using to Testmy Singapore server

VPN PPTP Connection to LA and using to Testmy LA server

VPN PPTP Connection to SF and using to Testmy SF server