Throttling, Level3, United States Connect

[Teera]

I have technical dilemma. Can you please tell me if any other the following still hold true, or are true in the first place as far as using ISPs in Thailand. I suspect that I may have been throttled. Though I thought that this could only happen if you download a lot of data over several billing cycles. The details are:

1. 8/1 MB line

2. was able to get 1.5MB download consistently for a few weeks

3. After Thursday afternoon the link from the United States slowed to only 40k

4. Google and Youtube DNS Servers are normal speed

5. Speed test in Bangkok is normal

6. speed test in Singapore and Hong Kong are Normal

7. level3 undersea cable from the TOT International Gateway to San Francisco shows a slow rate of only 50k

Does TOT throttle users and has anyone had the similar problems with their connect speed to the US in the past few days?

Is it better to use a VPN, proxy or another local ISP? I need a constant 1.5 download in the Thonlglor area.

Please note that Google and Youtube don't count as there is a regional hub in Singapore and Hong Kong.

Sorry that this pic below copied and pasted from the tot (Tel org of Thailand) is small. I coped it from the tot website at:

http://www.totiig.net/totiig/

[joncl]

I thought the CAT 10gb fiber we use goes via San Jose (if-13-27.icore1.SQN-SanJose.as6453.net) and it is OK here at the moment but slower than normal as it is the start of peak times here in BKK;

Edit: added the 1st USA hop via San Jose to level3 above

[Pib]

I'm on the TOT 4Mb/512Kb plan and I normally get around 0.6-1.0Mb to San Francisco depending on the time of day. Here my Speedtest results to SF as of 6:45pm

[wjhall]

Tried looking at your Speedtest results but just got the actual Speedtest application

Not tried this one before so gave it a whirl

San Francisco

Ping 250 ms

d/l 2.61 m/bs

u/l 1.55 m/bs

Then tried

Bangkok

Ping 48 ms

d/l 4.97 m/bs

u/l 3.27 m/bs

[joncl]

I'm on the TOT 4Mb/512Kb plan and I normally get around 0.6-1.0Mb to San Francisco depending on the time of day. Here my Speedtest results to SF as of 6:45pm

Indulge me please .. what is the test to San Jose like? Allot faster?

[Pib]

I'm on the TOT 4Mb/512Kb plan and I normally get around 0.6-1.0Mb to San Francisco depending on the time of day. Here my Speedtest results to SF as of 6:45pm

Indulge me please .. what is the test to San Jose like? Allot faster?

No, not a lot faster. I did do a test to San Jose a few minutes after doing the San Francisco test and the San Jose metrics were pretty much identical. I don't remember the exact upload number (but it was very close to the SF numbers), but the download number was 1.01Mb. I don't matter if a zillion gigabyte undersea fiber optic cable runs to SF or SJ, what ever bandwidth your ESP allots to you is what you will get...a person's ISP is the choke point, not the high speed undersea lines. On my 4Mb TOT plan on the "average" my international bandwidth is approx 1Mb here in western/suburb Bangkok...sometimes I run around 0.6Mb and on a few occassions I've got 3Mb+ numbers. My physical TOT phone/DSL line is fine and has good attenuation and SNR numbers in rain or shine. I expect TOT throttles/shares its bandwidth among customers at around 20-25% of your package's max "IP" speed; not max sync speed...and a person's max international speed seems to vary depending on where you connect to the TOT backbone.

Below are my San Jose and San Francisco results as of Monday morning at approx 7:35am.

[t.s]

this is what i was seeing today.

true 4mb package (upgrade to 7)

3:53 pm from sukhumvit 11.

[Pib]

I just love True speedtest results, especially the bogus ping number. For example round trip to San Jose is 8000 miles times 2 = 16,000 miles. Divide 16,000 by the speed of light which is 186,000 miles/second and you get 0.086 or 86ms. And actually the speed of propagation through the fiber/wire and added delays from electronics/servers along the way would really result in an even higher (slower) number than 86ms like reported above. Something in the 200 to 300 ms range is more realistic. Since the ping number is inaccurate, I wonder how truthful / accurate the download/upload speeds are, especially the download speed. Don't know how the True network skews the results...probably proxies.

[jfchandler]

The True Internet speed numbers derived from Speedtest, at least in regards to connections to the U.S., are bogus... They haven't don't bear any resemblance to the real speeds that are measured. I did a whole separate TV post on the subject...

In my experience, a more realistic measure of U.S. connections speeds can be had using dslreports.com

[Pib]

Why do you feel he dslreports.com speedtest is more realistic. I gave it a try and it gave me upload/download/ping numbers like most any speed test site. Just interested as to why you feel it's more realistic. Thanks.

[jfchandler]

A couple of reasons...

1. I did a fair number of tests some time back (different days and times), and compared the results that came from SpeedTest with a variety of other testing sites, including DSL Reports. The SpeedTest numbers tended to come up significantly higher than the other sites, though it does vary from time to time...

2. I read a lot of comment here on TV particularly on what to make of SpeedTest's results (there are other technical-oriented threads on the subject), and after concluding they were considered unreliable at least as far as True customers are concerned, went out reading and looking on the Internet for what was considered a more reliable test measuring site. DSL Reports got good recommendations, from what I read.

3. I don't know that there's anything wrong with SpeedTest per se... Rather, it's apparently something True does in the way they handle their system or connections that tends to give artificially high speed numbers when using that site.... In other words, they're apparently gaming the system, somehow. Again, at least as far as international/U.S. connections are concerned.

Why do you feel he dslreports.com speedtest is more realistic. I gave it a try and it gave me upload/download/ping numbers like most any speed test site. Just interested as to why you feel it's more realistic. Thanks.

[Pib]

Your rationale works for me...thanks.