Bandwidth!

[Rimmer]

Hi Guys

Possibly someone already did this excercise? But if not maybe other folks can post their figures as some of them seem very low.

All tests from ADSL Thailand site on the local service.

Test figures over a one week period

Dial up data from one TT&T line in Jomtien

The first column is the bandwidth.

The second column is the D/L speed

The third is the provider

39.4----4.93----JI NET-DIAL UP

37.7----4.71----JI NET-DIAL UP

35.4----4.43----LOXINFO-DIAL UP

34.0----4.25----LOXINFO-DIAL UP

29.4----3.68----JI NET-DIAL UP

27.6----3.45----LOXINFO-ADSL

26.7----3.40----LOXINFO-ADSL

20.1----2.51----LOXINFO-ADSL

216.8----27.10----IP STAR

225.5----28.19----IP STAR

212.4----55.00----IP STAR

Please note these are bandwidth and download speed figures NOT connection speeds

[tuky]

What test did you use Rimmer?

[chownah]

Now I'm confused. I've read alot of posts about maxnet since I'm considering. Lots of posters say how bad it is and that they get speeds as slow as a dialup...but...they say that they have download speeds of ball park 20kbps. Now rimmer shows that dialup is getting ballpark 3.4kbps.

Seems like the people complaining that their speeds are as slow as a dialup are wrong....seems like their speeds are alot faster than a dial up...maybe 6 times faster. If I could get 6 times faster download by switching from my TT&T Enrich dial up connection to the cheapest TT&T maxnet (256/128 I think) I'd jump at the chance. For me they would cost the same since my dialup charges are four to five hunderd baht per month. Have these guys just forgotten how really slow a dial up is?

[chownah]

I did some calcs on Rimmer's data...it appears that the first column divided by 8 gives the second column....except for the last one (maybe its a mistake). I think what this means is that the first column (whatever it is measuring) is measured in kbps (kilobits per second) and the second column is just the same measurement measured in Kbps (kilobytes per second) because one byte is equal to 8 bits. So....what does this all mean....

Also I just went to a speed test site and they said that my dialup modem was downloading at 80kbps....strange since my modem only goes up to 56kbps and my connection speed is only 40kbps....seems like the Twilight Zone is the only answer.

[monty]

Chownah,

Lots of people mix up the speeds used on internet connections!

Two numbers are used, the first being connection speed indicated in kilobits per second. (kbps)

For example adsl speeds are indicated 256/128, meaning 256kbps download and 128 kbps upload. A dial up modem's theoretical maximum is 56kbps, but will normally connect at between 40 and 50kbps.

The second indication is the download speed. These are indicated in kiloBytes per second (kBps)

1 kiloByte equals 8 kilobits...

So a 40 kbps connection speed can download at 5 kBps (4 kBps X 8 = 40 kbps)

Most test sites give you your connection speed (in kbps). So if somebody tests and gets 20kbps, he is getting roughly half of what a dial-up modem (40-50kbps) would get him...

What Rimmer posted means just that the test shows him that he is for example getting the equivalent of a 27.6 kbps connection speed (what he calls bandwidth), which allows him to download at 3.45 kBps (27.6/8=3.45)

What the Maxnet users are upset about is when they test and the test indicates 20 kbps (which allows them to download at 2.5 kBps). Dial up connects at over 40kbps, allowing for downloads of 5kBps

[monty]

Chownah, what did the download test download?

A textfile gets highly compressed by the dial-up system, giving you seemingly high speeds.

A .jpg image or .mp3 file would normally come in at roughly your connection speed since these are higly compressed filetypes already...

[Buff-horns]

Hi Guys

Possibly someone already did this excercise? But if not maybe other folks can post their figures as some of them seem very low.

All tests from ADSL Thailand site on the local service.

Test figures over a one week period

Dial up data from one TT&T line in Jomtien

Please note these are bandwidth and download speed figures NOT connection speeds

<{POST_SNAPBACK}>

you need to itemise what times you did the tests .

and how far from the exchange you are.

[GoodHeart]

I have yet to get comparable results from these four meters whilst using them back-to-back, from the same computer. Still, it makes for an interesting exercise.

Speakeasy

http://www.speakeasy.net/speedtest/

McAffe’s Internet Connection Speedometer

http://us.mcafee.com/root/speedometer/test_0150.asp

Bandwidth Speed Test

http://www.wugnet.com/myspeed/speedtest.asp

DSL Reports

http://www.dslreports.com/stest

[Khutan]

Well I can only say how wonderful True is this morning.

I got a test at 37 Mbps..... I couldn't believe it so I ran the test again.

Amazing as it is 8 times the rated speed of the modem - Truly amazing...

Here is a snapshot of the test

I did a proper test with wugnet.com and found the speed to be actually reasonable.... as the second picture will show. I am on a 2mbps package, so to get outside the country at that speed, isn't too bad

[chownah]

Chownah, what did the download test download?

A textfile gets highly compressed by the dial-up system, giving you seemingly high speeds.

A .jpg image or .mp3 file would normally come in at roughly your connection speed since these are higly compressed filetypes already...

<{POST_SNAPBACK}>

When I went to a web site and had them do a test of my speed the results came back 80kbps when my connection was at 40kbps. I have no idea what file they downloaded...I didn't have to pick the file....everything was automatic....I didn't do anything.. and after I waited awhile the result came back. I was a passive observer to a process that claimed to measure my download speed. The reported result was twice as fast as my connection speed. This is not right. Something is bogus here. They obviously made a big mistake. It appears that mattnick has run into the same thing in his latest speed test results.

Edit: I just went out and did it again....got the same results. This is the link where I went: http://us.mcafee.com/root/speedometer/test_0150.asp

Edited October 25, 2005 by chownah

[Rimmer]

Chownah,

Lots of people mix up the speeds used on internet connections!

Two numbers are used, the first being connection speed indicated in kilobits per second. (kbps)

For example adsl speeds are indicated 256/128, meaning 256kbps download and 128 kbps upload. A dial up modem's theoretical maximum is 56kbps, but will normally connect at between 40 and 50kbps.

The second indication is the download speed. These are indicated in kiloBytes per second (kBps)

1 kiloByte equals 8 kilobits...

So a 40 kbps connection speed can download at 5 kBps (4 kBps X 8 = 40 kbps)

Most test sites give you your connection speed (in kbps). So if somebody tests and gets 20kbps, he is getting roughly half of what a dial-up modem (40-50kbps) would get him...

What Rimmer posted means just that the test shows him that he is for example getting the equivalent of a 27.6 kbps connection speed (what he calls bandwidth), which allows him to download at 3.45 kBps (27.6/8=3.45)

What the Maxnet users are upset about is when they test and the test indicates 20 kbps (which allows them to download at 2.5 kBps). Dial up connects at over 40kbps, allowing for downloads of 5kBps 

Right, thanks, I put the original post in the Pattaya section but the moderator moved it. so my original intent to find out if other folks get similar figures in Pattaya is not now happening, never mind.

To clarify I did a test just now and got the following results off ADSL Thailand 'local net' bandwidth test.

Quote

Your current bandwidth reading is: 245.00kbps

which means you can download at 30.63 KB/sec. from our servers.

Test Time: October 25, 2005, 7:28 pm

Test site is ADSL Thailand tested on the local net server. Does that help at all?

[chownah]

...........

To clarify I did a test just now and got the following results off ADSL Thailand 'local net' bandwidth test.

Quote

Your current bandwidth reading is: 245.00kbps

which means you can download at 30.63 KB/sec. from our servers.

Test Time: October 25, 2005, 7:28 pm

Test site is ADSL Thailand tested on the local net server. Does that help at all?

<{POST_SNAPBACK}>

Once again the "current bandwidth reading" of 245.00kbps is exactly equal to the "you can download" of 30.63KB/sec. These two numbers are the same. I would like to know if "current bandwidth reading" means the same thing as your actual current download rate? or is the "current bandwiddth reading" the theoretic maximum download rate but the actual download rate could be lower? Here's why I ask. I have a dialup and when I go on line quite often my "connection speed" is 40 Kbps....but....when I download something a window opens and tells me what rate the data is being transfered and it is highly variable so it is obvious to me that quite often I am not downloading at the 40kbps rate. Can someone explain the proper terminology for these things and how they interrelate? I'm starting to believe that "bandwidth" and "download speed" will always be equal and that they are just two different ways to measure the same thing....is this right?

Edited October 25, 2005 by chownah

[monty]

Chownah,

The tests Rimmer is doing gives you your actual download speed from a server located in Thailand.

adslthailand.com/bandwidthmeter

The highly variable speed you are seeing when downloading is a result of the contention ratio.

Every ISP has to share available bandwidth with a lot of users to keep it economically viable.

This means that for example for 30 dial-up users, only 56kbps is available. The result is that everytime somebody of those 30 users is downloading something (webpage, file,...) your speed will slow down. If nobody else is downloading, you'll get the maximum speed of your connection.

Now what is happening in the broadband world in Thailand, is that the ISP's are selling high speed packages at such low prices, with the result that they have to share the bandwidth with 50(or more) customers.

The effect is that they will get highly variable and low download speeds, their maximum speed might only obtainable in the middle of the night when everybody else is asleep!

The more expensive your package (for the same speed), the higher your actual download speed will be, and the more stable it will stay.

A sharing ratio (contention ratio) of 10 or less will give you pretty good and stable speeds, but in Thailand expect to pay at least 2500 Baht/month for a 512/256kbps link. (or 1500 for 256/128 or 4000 for 1024/256)

Anything cheaper will will almost guaranteed give you shitty speeds!

One just has to realize that every ISP has to pay around 25000 Baht/month for a 1024/512 link to the outside world, collected by our monopolistic, governmental institution The Communications Authority of Thailand (CAT)

Calculate the ISP's profit margin, and you can make a rough guess of between how many people your 1024/512 link will have to beshared when you only pay 790 Baht/month!!!

Edited October 26, 2005 by monty

[chownah]

I understand what you are saying about how sharing bandwidth works. What I would like to know is the definitions of the terms used to measure performance. For instance, to me the "actual download speed" means the actual rate at which information is being transferred to a computer. If a file had 8kb of data then if you had an "actual download speed" of 8kbps (which is the same as 1kBps) then it would take 1 second to download the data. With this meaning the contention ratio will effect the download speed. Is this the way the term "actual download speed" is usually defined? It seems to me that Monty's post uses the term "actual download speed" to mean the maximum download speed that would be attained if the contention ratio was 1.00. For me the concept is not difficult...it is the vocabulary that is problematic...I'd like to clarify the terms and I think it would be best to take one term at a time.....can we start with "actual download speed"?

[monty]

Actual download speed is the speed you actually can download something.

If you can download a 100kB file in 10 seconds, your actual download speed will be 10kBps, or 80kbps.

This the most important speed for the customers, since it will influence the quality of your internet experience.....

This actual speed you can test either by downloading a file yourself, or by going to one of those test sites.

The connection speed is the speed with which you are connected to your ISP.

With a modem it is somewhere between 30 and 50kbps depending on the phoneline quality.

With ADSL it will be exactly the speed for which you subscribed (e.g. 256/128 kbps)

People start complaining when the actual download speed is much lower then their connection speed.

The more expensive your package, the lower the contention ratio, the closer connection speed and actual speed will be.

With a contention ratio of 1:0, your actual speed will be something like 90% of your connection speed. The last 10% gets lost in overhead generated by the TCP/IP communication protocoll (data checking/correction, etc.)

[chownah]

Thanks for your patience. I think I'm getting it. In Rimmer's first post the bandwidth and download speeds ARE identical....they are just two terms used to describe the same thing.....I'm a bit surprised by this....I would think that "bandwidth minus overhead equals download speed" would be the proper way to define it. Can you advise me which is correct, "bandwidth = download speed" or "bandwidth - overhead = download speed"?

Also, you spoke of a contention ratio of 1:0. Shouldn't this be 1:1?

[monty]

Bandwidth is a bit of a general description, used to describe the available speed on a connection, the size of the pipe if you like.

You are correct that actual download speed equals bandwidth minus overhead, but this is only true on a 1:1 contention ratio. (you're correct on that one also , my mistake)

If on a 10:1 contention ratio, the ISP's bandwidth to the internet will be shared with 10 people, so the actual speed depends on the others users current usage. So the bandwidth available to the end users is constantly varying.

The bandwidth between you and the ISP is not what gives you actual internet download speeds, it's more an indicator of the maximum speed available between you and the ISP.

Edited October 26, 2005 by monty

[chownah]

Bandwidth is a bit of a general description, used to describe the available speed on a connection, the size of the pipe if you like.

You are correct that actual download speed equals bandwidth minus overhead, but this is only true on a 1:1 contention ratio. (you're correct on that one also , my mistake)

If on a 10:1 contention ratio, the ISP's bandwidth to the internet will be shared with 10 people, so the actual speed depends on the others users current usage. So the bandwidth available to the end users is constantly varying.

The bandwidth between you and the ISP is not what gives you actual internet download speeds, it's more an indicator of the maximum speed available between you and the ISP.

<{POST_SNAPBACK}>

In you last paragraph, does the "bandwidth between you and the ISP" mean the same thing as "connection speed"?...so that the sentence could read "The connection speed is not what gives you actual internet download speeds, it's more an indicator of the maximum speed available between you and the ISP."?

Also, in Rimmer's original post all but two pairs of numbers are exactly equal to two decimal places...there is no allowance for overhead...why do you think this is?

Edited October 26, 2005 by chownah

[monty]

The bandwidth between you and the ISP means indeed the connection speed between you and the ISP.

So "The connection speed is not what gives you actual internet download speeds, it's more an indicator of the maximum speed available between you and the ISP" is indeed correct! Just remember that 10% of that bandwidth gets eaten away by overhead.

The numbers Rimmer posted are the results of a testsite. For example http://www.speakeasy.net/speedtest/ gives following result on my connection:

Last Result:

Download Speed: 418 kbps (52.3 KB/sec transfer rate)

Upload Speed: 176 kbps (22 KB/sec transfer rate)

This is my bandwidth available to the internet.

My ADSL router is connected at 512/256 kbps, which is my available bandwidth (connection speed) to my ISP.

From this I can deduce that I have 81% of my bandwidth available for surfing/downloading, which is quite good. If 10% gets lost in overhead, I lose another 9%, probably due to the contention ratio.(I'm on one of the more expensive 10:1 ratio packages)

What Rimmer did not post was his connection speed to the ISP.

If you click on the connection icon of your dial-up connection, you'll see the speed at which the modem is connected to the ISP.

If he would have checked, he would have seen that his connection speed was something like 45kbps, which resulted in an actual download speed of 39.4kbps(4.93kBps). Again, 45 minus 10% overhead and some sharing going on...

Users of the cheap unlimited dial-up packages often see connection speeds of 40+kbps, but actual speeds of maybe 20 kbps. Same as with ADSL this is the result of a high contention ratio going on on the cheap packages!

[chownah]

Monty,

Thanks for the mini tutorial. I'm going to look at some speeds for maxnet in some of the other topics and re-evaluate whether I want to try it or not.

Thanks again,

Chownah

[monty]

I'm on the Maxnet4life package @2515 Baht/monty (512/256)

Pretty happy with it...

[kidtongue]

I'm on the Maxnet4life package @2515 Baht/monty (512/256)

Pretty happy with it...

<{POST_SNAPBACK}>

Hmmm using the wugnet.xom test I get 978 up 473 down. So much for T1. But like anything else, all is constrained by tht Borg collective of 805 MB total.