TRUE Internet 50Mb/20Mb

[ace1]

Hi,

I have just had TRUE Internet 50Mb/20Mb installed, 799 baht a month. When using www.speedtest.net it is regularly showing speeds of 53Mb/23Mb.

All good....BUT....Using Facebook the internet seems to drop out and all contacts go offline. I am unable to scroll quickly through newsfeed as it locks waiting to load. Various other websites are affected too, if not everything.

If I turn on my VPN (free via Opera) connecting to Singapore the the troubles go away.

So it would appear to be connection within Thailand, or problem with router ? Not really sure.

Has anyone got any ideas  ?

Thanks for your help :)

[muratremix]

It is because of true's transparent proxy.

if you keep vpn open, surfing won't be a problem.

[ace1]

4 minutes ago, muratremix said:

It is because of true's transparent proxy.

if you keep vpn open, surfing won't be a problem.

Thanks....so I am OK from my computer but anyone using my WiFi on phones etc will have problems.

Anyways of changing/disabling this ?

Thanks

[ace1]

Running a PING test from the router I am inconsistently getting a 25% packet loss and when VPN is turned on I don't appear to be getting any packet loss.

 

[muratremix]

just one packet loss does not mean there is a major problem.

transparent proxy can't be disabled. True engineers are idiots.

 

I used to have a second, more powerful (AC1200) router and I used it to keep VPN always open in my whole home network so I always get fast speeds.

Current provided huawei fiber ont probably does not support vpn. You'll need a second wireless router one that supports openvpn or pptp/l2tp if you want always on vpn in your home.

[brewsterbudgen]

Use  http://speedof.me/ for a more accurate idea of what speeds you are getting. 

[ace1]

Not got to the bottom of it yet but things have gotten a bit better.

Phones are working OK, missus was over exaggerating ! I connected my laptop directly to the router and it is working fine. Interestingly I have no problems with the WiFi card as I have 2 places where I stay and it hasn't got any problems with a non TRUE WiFi in the other place. 

Hey ho can live with it connected by wire for now.

Cheers,

[ignis]

19 hours ago, brewsterbudgen said:

Use  http://speedof.me/ for a more accurate idea of what speeds you are getting. 

 

Maybe doing something wrong ?

 

True Speed test

104.9 down

34.54  up

 

But on this speed test 'speedofme'

5.96  down

7.84 up

 

The one I normally use 'testmy.net'

104.64 down

9.17 up

 

Why is 'speedofme' so different ?  tested 3 times this is the highest readings

[brewsterbudgen]

1 minute ago, ignis said:

 

Maybe doing something wrong ?

 

True Speed test

104.9 down

34.54  up

 

But on this speed test 'speedofme'

5.96  down

7.84 up

 

The one I normally use 'testmy.net'

104.64 down

9.17 up

 

Why is 'speedofme' so different ?  tested 3 times this is the highest readings

According to them:

 

How It Works

SpeedOf.Me tests your internet connection by directly downloading/uploading sample files from your browser. It reflects your actual browsing and download experience.

Download Test

It begins with downloading the smallest sample size (128 KB). While it shows the real-time bandwidth graph, it also watches the download duration. If it takes less than eight seconds, next sample would be tried otherwise it continues with upload test. Among all downloaded samples only the last one which took more than eight seconds will be accepted and your download speed is actually based on the last sample file.

Upload Test

When the download test is completed, similar incremental mechanism will be used to send back data to SpeedOf.Me test server. The upload test starts with a smaller sample file and will gradually increase until it takes more than 8 seconds to upload a sample file.

Note:
If you are behind a proxy server or firewall, you may experience problems with upload test. It may not work well or it may give you inaccurate results. In these cases you should use another browser as different browsers react differently regarding this issue.

Smart Test Method

In order to guarantee that your Internet connection is tested thoroughly, SpeedOf.Me tests your bandwidth in several passes. Sample file sizes increase gradually until it takes longer than eight seconds to download the sample file. As a result, SpeedOf.Me unlike other speed tests, is able to measure connection speeds in a wide range. From a slow 10 Kbps Mobile GPRS to a very fast 100 Mbps cable user (or even higher) automatically.
Sample files are in these sizes: 128 KB, 256 KB, 512 KB, 1, 2, 4, 8, 16, 32, 64 and 128 MB.

Test Servers

SpeedOf.Me hosts its sample files on CDN. It benefits from 85 (and counting) super servers called PoP. Each PoP may utilize several servers. They are located in various places around the world near primary Internet exchange points. These super servers have a massive computing and caching power and are directly connected to all the major backbone networks.
When the test is started, the quickest and most reliable server, not necessarily the nearest, would be chosen. There are many factors involved in selecting the best test server for you. However, it is done automatically using CDN technology to ensure the best testing quality. Here are SpeedOf.Me PoP locations:

• North America:
• Ashburn(2), Atlanta(3), Boston, Chicago(3), Dallas(3), Denver, Los Angeles(9), Miami(2), New York(2), Newark(2), Philadelphia, San Jose(2), Seattle(2)
• Europe:
• Amsterdam(2), Copenhagen, Frankfurt(4), Helsinki, London(5), Madrid, Milan, Oman, Paris(2), Stockholm(2), Vienna, Warsaw
• Asia:
• Bangalore, Batam, Beijing, Chennai, Hong Kong, Jakarta, Kaohsiung, Mumbai, Noida, Osaka(3), Seoul, Shanghai, Singapore, Tokyo(3)
• Australia:
• Melbourne, Sydney
• South America:
• Barranquilla, Buenos Aires(2), Lima, Medellin, Quito, Rio De Janeiro, Sao Paulo

Accuracy

Unlike other speed tests that require third party applications (plugins), SpeedOf.Me tests your Internet speed directly from your web browser not through Flash or Java Plug-ins with lots of overhead.

SpeedOf.Me also downloads large contiguous sample files. This is what you usually do when downloading a webpage or media. This is while other speed test services download small chunks in parallel and do heavy adjustments to calculate your approximate speed.

SpeedOf.Me utilizes several test servers (PoPs) in various locations which makes the test result more realistic. Other speed test services have a different approach about test server location. Some choose the nearest physical test server, probably inside your ISP. It seems illogical because the speed between you and the selected server is always higher than your real speed [for connecting] to the rest of the Internet.
On the other hand, some speed test services use a single test server to test users' connection speed in all geographical locations. For example, imagine a user is in Asia or Australia while the test server is in the US. The result would be inaccurate again because the end user is very far away from the testing server.
All of these make SpeedOf.Me the most accurate and realistic speed test among others.

Which browser is better?

Generally speaking, the more HTML5 features the browser supports the more precise the test result would be.
Our recommended browsers are latest version of Chrome, Firefox or Safari.
If you are using Internet Explorer, please be advised that the upload progress may not be shown in some older versions of IE. So, please be patient while the upload test is in progress.

[KhunBENQ]

3 hours ago, ignis said:

Why is 'speedofme' so different ?  tested 3 times this is the highest readings

You doing nothing wrong.

 

True speed test goes to a True server in Bangkok (or maybe even closer to your location).

It says nothing more than that your subscribed package is set up OK (100 down, 35 up ?).

It has no other practical value when connecting with the real world.

 

Testmynet allows selection of server location. What does it show? Where was the server?

The result is surprising as testmynet does not have servers in Thailand.

 

This "speedofme" is new to me.

I tried it and it seems to connect to a server in Singapore automatically?

Singapore is the location outside of Thailand with the highest data rates for obvious reasons (big land cables, no submarine cables involved). A huge portion of internet traffic from Thailand goes via Singapore and then to rest of world via submarine cables.

The results can vary in a wide range depending on destination country.

 

Quote

5.96  down

7.84 up

Indeed not a "great" value especially the down being lower than up lets me assume it's throttled due to overload.

 

My results to Singapore:

testmynet 28.7, 14

speedofme 13.3, 10.7

(ToT fibre)

[Pib]

Speedof.me is a "Single thread" HTML5 tester (see below) vs Speedtest.net and testmy.net which use multi-threads (a.k.a., multi-streams).  I think there may be some way to setup testmy.net to do single-threaded tests but I'm not sure anymore.  And my  "threaded" I mean number of "streams, pipes, etc."  So if your ISP throttles individual threads/streams to X-amount, to get more speed to a location you need to use multi-threading which is what the great, great majority of speed testers do.  Normally they will use 4, 6, or 8 threads to ensure the full speed capability of your internet plan is tested.  Normally they will not go beyond 8 because that requires the server you are testing to to be resources with the capability to do more than 8 threads....maybe think of it as connecting 8 pipes between your water pump(s) and water storage tank and the tank can not handle more than 8 pipes connected to it.

 

On my AIS Fibre 100Mb plan I can only pull 21Mb speed to Singapore with speedof.me because AIS apparently throttles single-threads to approx 21Mb.   Now test to Singapore on Speedtest.net (a multithread tester) which will use 4 to 8 streams (threads) and I get 100Mb speeds pretty much24/7...use testmy.net which defaults to the Multithread test and I get 75Mb to 100Mb depending on the time of day.

 

If I use another HTM5 speedtester called DSLReports Speedtester and test to Singapore I can pull 100Mb speed since it defaults to using up to 6 streams (threads).  But if I go into its settings and change the number of streams allowed to 1 stream my test results drops to around 18Mb...pretty close to Speedof.me results which apparently only uses a single stream and you can not adjust the number of streams used in speedof.me  If I change DSLReports tester to 2 streams my speed goes up to almost 36Mb...change it to 3 streams and it goes up to around 60mb, change to 4 steams I'm up to around 80Mb, to 5 streams I over 90Mb by a little, change to 6 streams I'm up to 100Mb.  Now although you can enter up to 32 streams in DSLReports it appears it will not actually use over 6 even if you enter more than 6 streams...you will see an error during the test than any stream number entered over 6 apparently resulted in a error and only 6 streams were actually used.   The default setting in DSLReports is 32 simply because you can also select multiple sites to test to if desired, unlike many testers where you are only testing to "one" location.  As mentioned, giving this DSLReports example since you can adjust the number of streams in it's settings for your testing.  

 

If using DSLReports, first select "one" server to test to like Singapore and then change the Number of Download and Upload Streams to 6 each...Save...then run the test again.  Now with that result go back in and change the number of streams to 1...run the test again.  If your results are greatly different then you ISP is most likely limiting each stream (thread) to X-amount.

  

So if your ISP throttle "single-threaded" operations and/or the site your a using like maybe some streaming video site that only allows a single stream,  you speed could be way down from the results you get from multi-thread OOKLA-based testers like speedteet.net, ISP testers such as True's tester, and even HTML5 multithreadtesters.

 

Using both a single stream tester like Speedof.me or one you can set to single thread if desired like DSLReports Speedtester combined with multi-thread tester give your system the best overall test because in your day to day interneting you could actually be using both single-threading and multi-threading operations.

 

https://speedofme.zendesk.com/hc/en-us/articles/223659007-Why-are-SpeedOf-Me-test-results-different-than-other-speed-tests-

[Pib]

Below are DSLReports Speedtester results to Singapore when using the default number of Download and Upload Streams...the Default for Download is 32 streams but really only uses 6 per test location and tells you so with an error notice in the final result...the test still run it just telling you in a round about way on 26 of 32 streams were used to Singapore.  The Default for Upload is 6 and you get no error notice since you are not using more than 6 streams.  Guess each of their servers is resourced for a max of 6 streams (threads).

 

Result with Default Setting. Notice for my AIS Fibre 100 Download / 10 Upload plan I'm getting full speed to Singapore with a multi-thread/stream test.  Notice the areas I've circled like the areas that show the number of streams in the cloud (i.e., number of streams the test was set for) .  Basically the test was for a 6 streams Download and 6 streams Upload to Singapore.   Although the Download Cloud shows 32 when you look at the bottom of the result it says 26 of 32 download streams didn't work/start which means only 6 streams were really used.

 

 

Now the results for a 1 Stream Download and 1 Stream Upload.  My download speed is now down to around 18Mb download but my Upload is still around 10Mb even with only 1 stream.  With AIS apparently limiting each stream to aroun 20Mb that still more than enough to get full Upload speed on my 10Mb Upload plan.  But since my plan is 100Mb speed for Download I would need around 5 or 6 streams to get the full 100Mb speed.  As mentionedd in my earlier post, multi-thread testers like OOKLA-based tester such as Speedtest.net and even testmy.net unless set just right will default to multi-threads giving you higher/full possible speed results.

 

[wpcoe]

I'm confused by discussions comparing single-thread vs multi-thread testing.  What does it really matter?  Isn't the bottom line:  "How fast can my ISP download/upload data?"  Why should I care whether it's single- or multi-threaded?

[Pib]

Because how your ISP has his network setup determines how single and multiple thread data flow can occur. Since different applications and servers you are connecting to may or may not use multi threading if your ISP throttles each single thread to X-megabits...like limiting the size of each pipe...you can not flow data as fast.

Kinda like how a person might get good speed to a certain location when using multi threading like how a download manager operates or how a multi thread speed tester works. However that person's single threading to the same location like when live video streaming or if using a single thread speed tester like speedof.me may experience too low of speed....like when getting a lot of pausing/buffering.


Best when ISPs don't throttle single stream connections/operations...best having that one big pipe versus needing to use multiple small pipes if the application or site will even use multiple pipes.

[wpcoe]

Thanks.  I hadn't considered apps/servers that can only handle single threads.

[Pib]

Did a series of tests today a little after 11am to Singapore using the DSLReports Speedtester.   I'm on an AIS Fibre 100/10 plan.

 

As talked earlier the DSLReports Speedtester is an HTML5 tester like Speedof.me and Testmy.net, however, with the DSLReports tester you can vary the number of threads/streams used.   I used 6, 5, 4, 3, 2, and 1 thread on download and upload.

 

Below is summary chart showing the download results for different number of threads.   Notice how the speed rolls-off as I stepped down the number of threads used from 6 thru 1.  An below that summary chart is the individual tests...each test using a different number of download/upload threads...1 thru 6 threads.   

 

As you will see as the number of threads increased so did my speeds especially for download.. Didn't reach the full 100Mb download speed of my plan until using 6 threads.  For the upload side only 2 threads was needed to reach the full 10Mb upload speed.

 

This series of tests using the DSLReports Speedtesterwhich can be set to single-thread testing (but it's default is multithread testing) and the Speedof.me tester which is single thread tester (can't change that setting) that AIS Fibre throttles single threads to approx 20Mb since I would get around 21Mb using Speedof.me and around 16Mb using DSLReports test.  

 

So, if you are using  an application and/or connecting to some sever that will only allow single-thread operation then you probably will not achieve the full speed of your plan...it will be limited down to some amount like around 20Mb max on AIS Fibre.  But if using apps/servers that allow multi-threading (more pipes) then you'll get much more speed.  

 

Now I would of course prefer to have a plan where only 1 pipe was used and  allowed to operate at full speed of the plan, but I expect many (most) ISPs may not do that due to bandwidth consideration/costs/real world needs.    So, since applications and servers in our internet world may use single threads or multiple threads, using both single threaded and multi-threaded speed testers to test your internet plan will give you a better idea of how good the plan is.  

 

Would be nice to see some of you post some DSLReports Speedtester results on your plan such as 3BB, True, TOT, etc., while changing the number of threads (streams)...maybe at least do a test with the default setting which automatically uses 6 threads and then run another test where you changed the Settings to only do 1 thread for Download and Upload.

 

Now, on to the charts.

 

 

Summary Download Speed Chart by Number of Threads.   Left side starts at 6 threads and steps on down to 1 thread.

 

 

Individual Number of Threads Used Charts

 

1 Download/Upload Thread Used (notice 1 thread is not enough to reach full "upload" speed...but not nearly
enough to reach full download speed)

 

2 Download/Upload Threads Used (notice 2 threads is enough to reach full "upload" speed and
my download speed has approx doubled)

 

3 Download/Upload Threads Used (notice download speed is higher than with 2 threads)

 

4 Download/Upload Threads Used (notice download speed is higher than with 3 threads)

 

5 Download/Upload Threads Used (notice download speed is higher than with 4 threads)

 

6 Download/Upload Threads Use (notice I have now reached full download speed of my 100Mb plan)