Adsl Speedtests Which One Is Right?

[Supernova]

Supernova, for whatever it's worth, just now at about 8:30 p.m. I downloaded a file - test16 at 16,384 Kb - from the ftp://ftp1.optonline.net/ site in North America you listed, and the download speed based on my browser's reading started at about 200 KBytes per second and then gradually declined to finish out at about 140 Kbytes per second, or, in other terms, started at about 1600 Kbits and finished at about 1120 Kbits... That was with my Firefox browser and nothing else downloading... I'm not sure I have a standalone FTP client installed on my PC anymore.. It's been years since I FTP'd anything anywhere.

I did the same with the 512 MB file, and it started much faster, around 300 Kbs, then gradually fell down to under 100 Kbs, before bouncing back to fluctuate mostly between 130 and 180 Kbs per second...

That is a pretty common pattern I've noticed in other downloads via True's Internet. Things seem to start out with a relatively fast burst, but then the speed falls off pretty quickly and never gets back to the original rate, but instead bumps along at a slower rate until finish.

My results from optonline using Firefox (earlier this evening):

"test16" = file transfer starts off around 140 KB/s and gradually increases to the maximum 197 KB/s within a few seconds. Full speed throughout transfer.

"test512" = slightly faster burst speed and gradually climbs to 197 KB/s (like the first test). However, the download never completes; connection terminates after transferring 60mb or so. Retried twice, same result... Wierd.

Line Rate (down/up Kbps): 1888/512

[jfchandler]

Did you mention what kind of ISP plan you're using to get those results???

Separately, I watched Burn Notice on the H service tonight using a V service... measured the stream with Bitmeter about 11 pm to midnight...

The numbers bounced all over the place, but generally were in the 1500 to 2500 Kbits per sec range...Only thing going into my network.

That's 1.5 to 2.5 Mbit international to the U.S. on a True 8 Mbit premium plan...

In general tonight I didn't see as much downturn in speeds during the evening hours as normally seems to occur...

.

[Supernova]

Did you mention what kind of ISP plan you're using to get those results???

JI-NET Max Lite 2Mbps.

Not the fastest in the world, but it's reliable. Been with them for three years.

[jfchandler]

That;s a pretty good result, considering you're getting 1.5 or so Mbits tops from the U.S. on a rated 2 Mbit plan...

I wish I was getting that kind of performance ratio with True's 8 Mbit plan, but sadly, not even close...

Just finished watching a couple episodes of 24 via N and using a V....Again, the stream bouncing around a lot...

General range, midnight to 1 am...was all over between 1000 and 2000 Kbits...with occasional spikes down to 800 or so and up toward 2500.

So in other words, topping out between 2 and 2.5 Mbit per sec...

DSL Reports Java speed test to LA with V: 1167 Kbps...

DSL Reports Java speed test to LA without V: 1519 Kbps.

DSL Reports Flash speed test to LA without V: 1800 Kbps.

[hml367]

Yet another test for Microsoft Online connections..... http://speedtest.microsoftonline.com/

MSPain

[Pib]

Did you mention what kind of ISP plan you're using to get those results???

JI-NET Max Lite 2Mbps.

Not the fastest in the world, but it's reliable. Been with them for three years.

Couldn't agree more about the reliability of JI-NET. I was with them for about 15 months on their 2Mb BizConnext plan...that was their only plan available to me on my TOT line. It got 1.7Mb download speed to many international sites all the time...and their uptime reliability was like 99.99%....it went don't one time for abot 3 hours one morning, I called JI-NET, they jumped all over the problem/fixed it. JI-NET said the problem turned out to be a TOT hardware issue as JI-NET was providing their ISP service over a TOT line in my case. I only left them them because their plans are a little pricey (the BizConnext plan is 1890 Baht/mo) but it did come with a static IP and 3 email addresses and supposedly a 10 to 1 user/contention ratio.

Now the JI-NET Max Lite program at around 930 Baht is a better deal price wise I think for home use but you have to be on a TOT line "with ADC circuits." Always seems to come down to who is providing your phone line, who sometimes teams up with a another telecommunications company, to provide the line to your residence. JI-NET does provide plans up to 4Mb speed now (eg., Bizconnext at 2800 Baht)...these plans don't show on their web site...you have to call their sales office. http://www.jasnet.co.th/home/index.php Maybe in the future they will lower their prices and I'll switch back to them, but I expect since they have to pay the phone company a cut for the concessionaire agreement/provide ISP service over a phone company's line, their price will always be higher along with better/faster service....but being about 3 times higher compared to most TOT/True/etc plans is a little to much for me. The MaxLite plan seems to be one of their plans that is not priced way too high...wish I could have gotten that plan considering I'm a TOT line captive audience.

JI-NET definitely provides reliabe service....I'm sure they could get a lot more customers with more competitive prices on some of their plans ...you listening JI-NET?

[jfchandler]

Thanks for this post... Lots of good info there on a person's Net connection...much more than the typical speed test site....

Yet another test for Microsoft Online connections..... http://speedtest.microsoftonline.com/

MSPain

[Pib]

Yet another test for Microsoft Online connections..... http://speedtest.microsoftonline.com/

MSPain

I've always got much lower results using this speedtest, even when I was in the States. What about you? Cheers.

[jfchandler]

By comparison, True's supposed 8 Mbits plan is 1199 for regular and 1399 for premium, but True Move and/or UBC customers get the premium plan at the regular price.

Their customer service here in BKK has been very good, dedicated English language call center, and very good reliability for their Internet, much better in fact than their cable TV service...

Unfortunately, as you can see from the various streaming and speed test results from the U.S. that I posted over the night, True's international connection speeds -- at least for me -- seem to be pretty comparable to Supernova's 2 Mbit plan with JiNet. Though based on what you guys are saying, even for comparable actual performance, their pricing is much higher than True's...

In the end, it doesn't really matter how any company rates or calls a particular plan. What matters is what kind of performance, speed and reliability you get for a particular price.

[hml367]

What it seems to me is different people have different isp's which are "the best". Different people have different ways of guaging their internet connection speed which are "the best".

Then most people change isp's at least a few times while they are in Thailand.

Personally, besides and dial up I have had, I have had 5 different broadband isp's since I have been using the internet in Thailand. They were all good in the beginning, then for one reason or another their service changed..... and I changed isp.

People that are claiming very fast download speeds are probably in a fortuitous physical location in relation to their isp's physical equipment.

Another thing I have noticed, in line with some peoples' claims that isp's manipulate speed tests, is that different browsers give different results. I have recieved rediculous reports of speed when I have used Google Chrome. I am under no illusion that I ever got those speeds.

Then it comes down to the actual computer and its hardware and maybe more importantly, the network parameters being used by that computer. If those parameters are not "in sync" with what the isp is providing, you will not be getting optimum results. Beginning with Windows Vista and much more prevalent in Windows 7 the network settings are adjusted by the OS itself dynamically. You can get programs that will change some of the settings for you, however, so far I have not found a way to make the settings stick.

With all the different "this is best" advice on forums, I am even more confused than normal.

By the way, currently I am using AIS Super 3G in Chiang Mai and I am getting international speeds equal and better to anything I have had in the past. Unfortunately, most of Chiang Mai (and anywhere else outside of Bangkok) does not have available all that is available in Bangkok.

MSPain

[Pib]

Supernova, for whatever it's worth, just now at about 8:30 p.m. I downloaded a file - test16 at 16,384 Kb - from the ftp://ftp1.optonline.net/ site in North America you listed, and the download speed based on my browser's reading started at about 200 KBytes per second and then gradually declined to finish out at about 140 Kbytes per second, or, in other terms, started at about 1600 Kbits and finished at about 1120 Kbits... That was with my Firefox browser and nothing else downloading... I'm not sure I have a standalone FTP client installed on my PC anymore.. It's been years since I FTP'd anything anywhere.

I did the same with the 512 MB file, and it started much faster, around 300 Kbs, then gradually fell down to under 100 Kbs, before bouncing back to fluctuate mostly between 130 and 180 Kbs per second...

That is a pretty common pattern I've noticed in other downloads via True's Internet. Things seem to start out with a relatively fast burst, but then the speed falls off pretty quickly and never gets back to the original rate, but instead bumps along at a slower rate until finish.

My results from optonline using Firefox (earlier this evening):

"test16" = file transfer starts off around 140 KB/s and gradually increases to the maximum 197 KB/s within a few seconds. Full speed throughout transfer.

"test512" = slightly faster burst speed and gradually climbs to 197 KB/s (like the first test). However, the download never completes; connection terminates after transferring 60mb or so. Retried twice, same result... Wierd.

Line Rate (down/up Kbps): 1888/512

For comparison, here's my results on the TOT 4096/512Kb Basic Fit Plan using IE8 at 11:30am/Friday/30 July in western Bangkok:

"test 16" = file transfer starts off around 150Kb/s and quickly increases to a high of 200Kb/s and they stablizes at around 192Kb/s

"test512" = file transfer starts off aound 150Kb/s, quickly increases to around 200Kb/s and then slowly climbs to a high of 240Kb/s and then stablizes at around 230KB/s at the 110Mb download point...at about the 140Mb download point I abort the test since the download speed is changing very little/very stable.

Cheers.

Edited July 30, 2010 by Pib

[hml367]

I am trying to learn, not argue, Supernova.

The difference in speed is negligible when testing with ftp or http, why is there a File Transfer Protocol? Is it left over from "slower" days?

MSPain

^ The speed difference is negligible.

[Supernova]

^ Chill dude... No one's arguing. You asked a question, I answered.

The difference in speed is negligible when testing with ftp or http, why is there a File Transfer Protocol? Is it left over from "slower" days?

MSPain

This document offers an excellent comparison of the two protocols.

Be sure to read through the entire document, especially the section on Transfer Speed.

[jfchandler]

Since you guys ran the results for this site midday today, and mine originally were last night, I decided to re-run mine today - 3:30 pm - to aim for a bit more comparability in time of day... Measured in Kilobytes by Firefox browser download window

Test 16 file - started at 250 KBytes, then steadily dropped below 100, then back up to 190, then down to 50, before finishing at about 110. Typical True zigzag performance, at least for my connection.... Probably averaging somewhere better than 100 KBs (800 kilobits per sec)...but the browser doesn't give an average.

Test 512 file - took maybe five ups and downs during the period I let the download run, starting at 150 KBytes, falling to 90, hovering around 100 to 110 quite a while, rebounding to 140, then falling to 65, then up to 130 before finally ending about 80 KBs... So again, probably averaging somewhere in the 100 (800 Bbits) to 140 KBs (1120 Kbits) range...

Not very impressive, considering the equal or higher scores others are posting with rated 4 Mbits plans from other ISPs. And probably slower overall than the comparable test I ran last night...

Supernova, for whatever it's worth, just now at about 8:30 p.m. I downloaded a file - test16 at 16,384 Kb - from the ftp://ftp1.optonline.net/ site in North America you listed, and the download speed based on my browser's reading started at about 200 KBytes per second and then gradually declined to finish out at about 140 Kbytes per second, or, in other terms, started at about 1600 Kbits and finished at about 1120 Kbits... That was with my Firefox browser and nothing else downloading... I'm not sure I have a standalone FTP client installed on my PC anymore.. It's been years since I FTP'd anything anywhere.

I did the same with the 512 MB file, and it started much faster, around 300 Kbs, then gradually fell down to under 100 Kbs, before bouncing back to fluctuate mostly between 130 and 180 Kbs per second...

That is a pretty common pattern I've noticed in other downloads via True's Internet. Things seem to start out with a relatively fast burst, but then the speed falls off pretty quickly and never gets back to the original rate, but instead bumps along at a slower rate until finish.

[hml367]

Thank you for the information.

And nothing to chill about for me. I was merely stating up front that I was not arguing because so many people take that stance... as you seem to have done.

MSPain

^ Chill dude... No one's arguing. You asked a question, I answered.

The difference in speed is negligible when testing with ftp or http, why is there a File Transfer Protocol? Is it left over from "slower" days?

MSPain

This document offers an excellent comparison of the two protocols.

Be sure to read through the entire document, especially the section on Transfer Speed.

[msg362]

OOPS I'm beginning to wish I never asked the question!!

But as a total idiot, let me just say I have ( at about7.00 pm ) downloaded 950 Mb from an apple server ( I know no where!!) to update my software. It took about 25 minutes! the fastest I've ever had!

does that help the discussion?

[Pib]

Just as a clarification, the TOT 4Mb Internet Plan FTP download speed I gave in Post #71 above was using IE8 only and not a separate/integrated Download Manager like IDM, Orbit, etc. When using the IDM download manager I get around a 350KB download speed. Cheers.

[jfchandler]

Finally, some degree of truth in advertising (boasting).... at long last....

In HK on 30mb connection i cannot get that maximise on single thread download.

in KR on 100mb connection i cannot get that to maximise on single thread download.

in TH on 16mb connection i cannot get that to maximise on single thread download.

The problem is when you get into these Ultra high bandwidth products you have to multi thread ie Torrents, Multiple simultaneous downloads, Download managers, Usenet multiple connections to server.

You simply cannot analyse a high bandwidth product and say you are not getting what you pay for internationally based on a single thread download. It just does not work like that at those speeds. 7 years ago yes.

if you an maximise your bandwidth whether it's 16mb 20mb,50mb or 100mb somehow it means that international bandwidth is available to you whether u know how to or care to utilise it is up to you. You cannot expect single (old school) downloads to maximise your high bandwidth. It is unrealistic.

Your Speedtest.net ping is not correct agreed. However mine is 253ms-305ms to US west coast which is normal and confirmed through Ping. So is my speedtest.net then valid?

If you use a high bandwidth speed test for example that used by BELL for their Fibre customers it shows realistic results for thailand as well. Enter any number for your number. Bell Fibre Speed Test

I would take the above as really one of the more accurate Speedtest available today if you do not trust speedtest.net this one is well thought out it's high bandwidth, large file size.

Do you trust that one JF? It gives accurate results in Hong Kong, Korea and as far as i could tell accurate results in Thailand.

This post has been edited by negreanu: 36 minutes ago

[negreanu]

Correct neither can i get a single threaded download to Maximise my 100mb in Korea or 30mb in Hong Kong. Or indeed my 20mb connection at home in the UK.

So you expect the impossible from a high bandwidth product. The bandwidth i purchase from all my ISP's is available if you care to utilise it correctly. If you are expecting to max out your bandwidth on new high speed products such as we use today with a single thread download you are living in a dream world. It just does not work like that anymore.

How do you feel about the BELL high bandwidth speedtest? or is that being manipulated by True as well?

If you look at your posts. You deny everything that gives a result of fully using your bandwidth from Thailand and praise and say the ones that are correct are the ones that give results that don't.

Please answer this if we take Bell's speedtest and indeed speedtest.net and run it in Hong Kong, Korea and Thailand i get fairly accurate results for the allocated bandwidth. However if we run all your given speedtests in Korea, Hong Kong and Thailand the results are no where near accurate.

So you cannot possibly be saying that my Korean ISP, Hong Kong ISP and Thai ISP are all lying are you???? Your into the realms of conspiracy theory and tin foil hats if you are.

Edited August 25, 2010 by negreanu

[jfchandler]

Negreanu, not sure why you've recommended the Bell Canada speed test above... But I tried it this morning, about 11 am, and it appears to give very unrealistic results...

Ran it 3 times consecutively, and each time it showed a download speed of 10,000+ Kbits per second...that's almost 10 Mbits per sec...

Do you really want me to believe that True's 8 Mbit plan INSIDE Thailand is producing a real 10 Mbit connection to Canada??? I think not..

So, I did some other comparisons, which produced more realistic real results.... I'll let the readers decide which ones they think are real....

Bell Canada - DL 10,000+ kbits

Stanford Univ. - Bay Area, Calif.

Download 1.25 Mbits (1,280 kbits)

DSL Extreme - Los Angeles ISP

Download 1,830 kbits

I also ran the Microsoft data center diagnostic that another member had recommended above, and its results were even slower....embarrassingly slower... Not sure why.....

Before you say it.. I know I know... Stanford Univ. doesn't really know how to accurately measure Internet bandwidth speeds... They're not up with the modern world.. new technologies...fast Internet speeds... I'm sure they'll be disappointed to hear that....

[lomatopo]

Turn everything off, FTP files up and/or down; I use FileZilla. Instantaneous speed is displayed, as are the final results: file size, total time (get the calculator out, remember the difference between bits and Bytes, and kilo- and mega-). Using FTP, within Thailand, I seem to get nearly full wire speed after accounting for overhead. Of course this assume the other end has a bigger pipe than me.

[negreanu]

You just make a fool of yourself every time.

Like i said above you ridicule everything that gives a good speed test result and praise and believe everything that does not.

You can disregard DSL Extreme as the file size is too small. See how your connection to their server is still accelerating and then it ends.

Stamford is way outdated for today's high bandwidth speedtests. That particular Stamford speedtest is 10 years old.

Bell's Speedtester is made for todays High bandwidth services as is speedtest.net. (It was developed for their Fibre customers). And if you look at the forums for it most agree that it's a pretty accurate speedtest for high bandwidth customers on Bell and elsewhere.

Look Bell is accurate 100% i can download at the value it reports every time. So Bell is confirmed with actual physical downloads.

Bell is confirmed in HK, Bell is confirmed in KR and backed up with exactly the download rates i get on files.

[jfchandler]

Negreanu, you're selling snake oil and beachfront property in Las Vegas too, right???

Do you really want me to believe that True's 8 Mbit plan INSIDE Thailand is producing a real 10 Mbit connection to Canada??? I think not..

Edited August 25, 2010 by jfchandler

[negreanu]

Yes I do. I know people on 4mb Premium True who can achieve download speeds equiv of 5-6mb.

What is your current Sync rate on your router JF?

Your on Premium right which is True's international product. Not within thailand product.

I don't even know why i waste my time with your uneducated ramblings there is no technical or sound reasoning behind any of your points.

You just ramble and ignore all my relevant points i make and continue on your i am being short changed by True campaign and that Thailand has a problem with international bandwidth still.

You just dont know how to utilise your allocated bandwidth or wish not to do so. No matter what speed your connection in what country by any provider. You would not be happy simply because i think you do not understand how high bandwidth products work.

There are 2 ways to look at this. You can b1tch and complain for the next 20 years that you are not getting the bandwidth you pay for from True on your obscure speedtests and websites. Or you can adapt and say ok how can i ensure i maximise my bandwidth that is available to me.

Enjoy it!. Looking forward to 50mb next week. But if the year 2000 Stamford speedtest says i aint getting 50mb i wont lose any sleep . However I will post a review of it and screenshots just to wind you up even more.

Final question what is it you do with your international bandwidth that u cannot maximise it doing so?

Edited August 25, 2010 by negreanu

[negreanu]

Probably the most demanding thing to do is stream video right.

Well if you take 16mb True you can stream for example these 2 easliy with Istats indicating download rates of around 1.7MB/s at 1080p. No stuttering pausing jumping etc. Very smooth.

Make sure you click 1080p on the control bar.

Edited August 25, 2010 by negreanu

[jfchandler]

Back to the personal insults, huh???

Not happy because you finally had to admit that True's 16 mbit package can't/doesn't provide anything close to that for a single connection... Yes...I can understand that...

But I'm open to a challenge, and trying new things... So, this afternoon at about 5 pm, I took two different approaches to multi-threaded downloading you've been promoting, since it's already clear that single streams top out well short of any rated speed package...

Logged onto ITunes, and ordered up downloads of a bunch of concurrent podcasts... Monitored the live incoming streams into my connection using BitMeter... Indeed, with Itunes, it showed a total download stream that bounced around between 6, 8 and 10 Mbits... A pretty good sign that, as previously suggested, Apple-Itunes is using some kind of distributed content system, with that content not likely coming from the U.S.

Then logged onto one of the better known video torrent streaming packages, and opened 9 concurrent video windows. The supposed bitrate of the 9 streams, originating from the U.S., totaled about 4.9 Mbit... The actual download streams coming into my PC never got past 4 Mbits...and averaged in the 3-4 Mbits range -- hardly meeting or exceeding True's 8 Mbits label for my package. And those readings (4 Mbits out of 4.9 Mbits) look about right, since the final two streams never got past buffering (as opposed to actually playing).

Be interesting to see how that performance changes if I repeat the same experiment tonight at 9 or 10 pm....

Edited August 25, 2010 by jfchandler

[siamect]

Probably the most demanding thing to do is stream video right.

.

.

Not necessarily... there are not much point for the stream-server to give you higher speed than the video need to run without interruption...so most likely the stream-server is throttling...

I would think a normal bunch of simultaneous downloads from a handful servers would give you more demanding situation.

Martin

Edited August 25, 2010 by siamect

[negreanu]

ot happy because you finally had to admit that True's 16 mbit package can't/doesn't provide anything close to that for a single connection... Yes...I can understand that...

I did not say only True the same is happens with all High bandwidth providers downloading internationally in every country i have tried.

Which Itunes store are you downloading from? if it is the US store they do not content distribute (there is far too much data to do that in itunes) US content is coming from Apples Data Farm in the US for the US itunes store anyway cannot speak for Thai version etc.

Regarding Torrents depends on the seeds, Never tried streaming video torrents as you have done.

Normally just .Rar etc. file downloads.

Cue up some Torrents, downloads from itunes, microsoft or anywhere else you like open a youtube HD video or such. Look at your bandwidth monitor. I am pretty sure you will see 8mb+ there which is exactly what your paying for.

Don't buy into the apple hype of the distributed content that was posted on here. The reason you get great speeds from Apple/Microsoft etc. is because they invested in the bandwidth necessary to provide their customers.

Edited August 25, 2010 by negreanu

[jfchandler]

It's been an interesting thread, notwithstanding all the personal stuff....

I've learned a lot, including that my download capacity isn't limited to that just which comes thru a single thread, but rather, those threads can be additive on top of each other, at least for certain kinds of applications.

But equally true, it also means that downloading or streaming that needs to rely on a single thread (such as streaming Adobe Flash video or Slingbox streams, etc etc.) can still be constrained by network congestion, even on the faster packages.

I guess the trick is to find an ISP package that at least gives you at least enough speed to cover the single stream things you want to do in adequate manner.

In other words, if what you really want to do is watch Hulu from the U.S. and it streams at about 450 Kbits/second (.43 Mbits), then you want an ISP that can give that can give you 450 Kbits/second reliably for an international connection.

Now, if someone really just needs a reliable .43 Mbits international connection, how much better is a 4 Mbit vs an 8 Mbit vs a 16Mbit DSL package going to be in achieving that.... Interesting question.

[Spoonman]

Cue up some Torrents, downloads from itunes, microsoft or anywhere else you like open a youtube HD video or such. Look at your bandwidth monitor. I am pretty sure you will see 8mb+ there which is exactly what your paying for.

Just hit the go button on my overnight downloads from rapidshare.

Im on True's 8mb plan.