How Bad Is True Adsl ?

[Khutan]

People probably know my business is IT and also Telecommunications with IT.

I also know the Media read Thaivisa.com so if we want better service from True, then we need some hard facts so the everyone can report fact rather than opinion.

So this is very easy. and if you have an unlimited ADSL account like the millions that true have sold, then this will take 5 minutes.

If you have windows etc do this

Start ==> run

Type cmd

you will get a strange Black box that is a "Dos" window.

If you are very technical type this cd\ then press enter

then CLS and also press enter, otherwise don't worry

Then and this is important type this :

ping proxy.asianet.co.th -t -l 4096

Do nothing for 24 hours

after 24 hours press Ctrl + C and you will get the statistics

I will post my sample in a few minutes.

BTW proxy.asianet.co.th is the recommended proxy from True, so for 24 hours sending a 4 Kilobyte Package it should be perfect...... or is it ???

[Khutan]

Here is a sample from about 2 hours of this command "Pinging" True

Ping statistics for 203.144.144.164

Packets: Sent 2717, Received = 886, Lost = 1831 <67% loss>

I hope you have the idea. This means that I lose 67% of my bandwidth to True. We are NOT talking about outside Thailand, but inside Thailand.

I am very interested in the results of other people and over 24 hours.

PS Minimize the screen, and if it runs for a bit more than 24 hours, then that does not matter

[penguin]

BTW proxy.asianet.co.th is the recommended proxy from True, so for 24 hours sending a 4 Kilobyte Package it should be perfect...... or is it ???

Here is a sample from about 2 hours of this command "Pinging" True

Ping statistics for 203.144.144.164

Packets: Sent 2717, Received = 886, Lost = 1831 <67% loss>

The asianet proxy l use with True is 203.144.160.246

Just running your test now with it and it shows about a 5% packet loss.

Time to change your proxy configuration?

[penguin]

My stats for the past 20 minutes.

Packets: Sent = 1415, Received = 1321, Lost = 94 <6% loss>

Approximate round trip in milli-seconds:

Minimum = 99ms, Maximum = 217ms, Average = 102ms

[JimsKnight]

My Stats for the past 5 minutes.

Packets: Sent =204 , Received =204 , Lost = 0% loss>

Approximate round trip in milli-seconds:

Minimum = 514ms, Maximum = 611ms, Average = 520ms

[MarkBKK]

People probably know my business is IT and also Telecommunications with IT.

I also know the Media read Thaivisa.com so if we want better service from True, then we need some hard facts so the everyone can report fact rather than opinion.

Please buy my new book, How Bad is True ADSL?, Let's Gather Hard Facts and Get it Fixed, the historically accurate account of a plucky band of farang who face up to an evil telecoms empire! Thrill as they complain till they're blue in the face to a Thai who neither understands them, nor cares!! Be seduced by the defiance of their deluded leader, who actually believes he is still at home!!! Recoil in horror as ''technician'' after ''technician'' ''resets'' something and fools the hapless farang into believing that things really are getting better!!!! Be amazed as farang after farang fall for the old ''No, no, the local media are really, really interested in our problems'' line!!!!! We have nothing to lose but our ... er ... well, we have nothing to lose!!!!!

(How Bad is True ADSL?, Let's Gather Hard Facts and get it Fixed will be published just as soon as I can email the manuscript to my publisher.)

This just in: Metro and Untamed Travel will be employed as consultants by the Council for Democratic Reform to sort out all farang Internet connectivity problems.

[Lannig]

People probably know my business is IT and also Telecommunications with IT.

I also know the Media read Thaivisa.com so if we want better service from True, then we need some hard facts so the everyone can report fact rather than opinion.

[...]

ping proxy.asianet.co.th -t -l 4096

/quote]

Start by using metrics that mean something. ICMP echo traffic is given very low priority by almost any kind of router that does QoS, and many do these days, even here.

There could be a gadzillion reasons for which an ICMP echo or echo reply packet gets lost that wouldn't apply

to normal TCP traffic. Including the fact that ICMP echo could well be rate-controlled at the server itself!

You're going nowhere with this. Maybe your business is IT, but I hope for you it's not in technical things...

--Lannig

[Deksan]

People probably know my business is IT and also Telecommunications with IT.

I also know the Media read Thaivisa.com so if we want better service from True, then we need some hard facts so the everyone can report fact rather than opinion.

...

ping proxy.asianet.co.th -t -l 4096

Start by using metrics that mean something. ICMP echo traffic is given very low priority by almost any kind of router that does QoS, and many do these days, even here.

There could be a gadzillion reasons for which an ICMP echo or echo reply packet gets lost that wouldn't apply

to normal TCP traffic. Including the fact that ICMP echo could well be rate-controlled at the server itself!

You're going nowhere with this. Maybe your business is IT, but I hope for you it's not in technical things...

--Lannig

No need to be mean or ironic everybody is far from knowing everything ...

Edited September 27, 2006 by Deksan

[Khutan]

...Please buy my new book....

Thank you for your constructive input.

[MarkBKK]

...Please buy my new book....

Thank you for your constructive input.

My pleasure. I'm always in support of tilting at windmills

Thank you for your constructive original post.

[Prasert]

Start by using metrics that mean something. ICMP echo traffic is given very low priority by almost any kind of router that does QoS, and many do these days, even here.

There could be a gadzillion reasons for which an ICMP echo or echo reply packet gets lost that wouldn't apply

to normal TCP traffic. Including the fact that ICMP echo could well be rate-controlled at the server itself!

You're going nowhere with this. Maybe your business is IT, but I hope for you it's not in technical things...

--Lannig

Very good point!

Since the first DoS attacks, the majority of all ISPs have a rate-limit configured at their boundaries, allowing a few kbps icmp traffic (which is enough for normal usage). Anything above that is dropped.

If 50 people would do the proposed test, it would show many drops. The interface policy on a router before this proxy server would show a high drop-rate.....

So indeed, this test is useless for gathering any statistics.

If you want to gather statistics on your line, use mrtg to graph bandwidth usage and delay times, or some similar software.

[Cromarty]

just out of curiosity Khutan, why do you ask people to do this?

"If you are very technical type this cd\ then press enter

then CLS and also press enter, otherwise don't worry"

[astral]

just out of curiosity Khutan, why do you ask people to do this?

"If you are very technical type this cd\ then press enter

then CLS and also press enter, otherwise don't worry"

Most Windows users do not know the command level at all these days.

[Indifferent]

Not sure a ping alone is a reasonable benchmark for quality of service.

I know most of my problems (if/when they happen) are due to slow speed ad certain times (presumably due to contention) and complete lack of access for periods (used to be days at a time in the past..its much better now).

Since most people are users and not techies if you are serious then you could supply a small task bar program that sits and downloads a small file from your website (say arbitrarily) once every hr and logs connect/disconnects. This would give very rough and ready statistics over time (which could be graphed to evaluate peak contention periods and the effect) and connection up time.

In fact. I used to have such a program before I reformatted the hard-drive that would accumulate the periodic benchmarks, upload the info to their web server and give a pretty display of all the statistics over time. I used it to prove to TOT that thier service was crap and I wasn't just a whining farang. Can't remember its name now though..If I find it again I'll post it.

[robhudd]

Hi All, my first post here after lurking for a while.

Just on comment on using ping to test the quality of the a connection, whateveryone has said is true about icmp being given a low priority etc. Also what you should be aware of is thaton any network there is a thing called a MTU which is the maximum size for a packet. On a DSL network this is usually around 1480 bytes. So if you send a packet of greater than this then it must be broken into smaller units and then re assembled at the other side (therefore done twice). So if you are trying to use ping dont make it a big packet size or it will give you worse results than you should really get.

Cheers

[Cromarty]

CLI is a dying area with windows, not so with linux.

But why change directory and then do a clear screen. Makes no difference to a ping. For those who want to be depressed with LoS speeds try Dan Elwells speedtest.

http://www.broadbandspeedtest.net/