Since my posting last night about the very slow ADSL service, I have been
getting some very helpful replies. As a result I have checked my speeds with
3 different speed tests. When I did Google I found many more speed tests.
So the problem is which is the best to use. Before I can say what is best I
must have a scale to judge the test, obviously accuracy of the speed
estimate is very important, but it seems that some tests give much more
information than speed alone, and this info. can be very important in
diagnosing problem areas.
First of all I started with www.bandwidthplace.com , you can check my
results for the week on a bar-chart at
http://users.powernet.co.uk/mkmarina/powernet_adsl.gif I liked this
because when I was getting a fast service it gave a high value and when I
got a crap service it gave a low value. So whether it was absolutely
accurate or not it reflected the perceived speed. Unfortunately it gives
very little other information.
So I followed up on one of the replies to my earlier posting and tried
Daniel Elwell's test at http://www.vantage.pwp.blueyonder.co.uk I liked
this test as it gave much more information which may help technical guys.
It's speeds were about 50% higher than those reported by bandwidthplace, but
they were of the same order and the two were near enough that it didn't
matter. There was one minor problem, it was geared up to specific ISP which
did not include my ISP, but that just slowed things down doing tests which
failed.
It was after this I did a Google and found a whole load of other broadband
speed tests and tried http://www.adslguide.org.uk/ mainly because it had a
"uk" in the URL. This looked good in some respects, it knew of my ISP. On
the couple of test I did it's results were much higher than the other two
tests, and that's why I'm writing this email.
I know that these speed values are constantly varying through the day, but I
can detect statistically significantly patterns in the readings that I have
already taken, and it seems that the latter didn't fit in with the other
two.
WHAT A WAY TO SPEND A SUNDAY :-(
It doesn't matter, I'm a sad old bugger who doesn't work, so I'm not
sacrificing my weekend, I don't have weekends. If you are an equally sad So
& So check out my websites at, http://users.powernet.co.uk/mkmarina
approaching half a million people a year check them out so they are
interesting to sad old buggers like me.
Back to the problems of ADSL - What's the best ADSL speed test to use, any
suggestions ?
Thanks guys,
Pete Loud
I would expect all the speed tests to be limited by the slowest point on the
network between you and them.
Remember that the route can change at any time and may be different for each
packet.
So the most accurate test will be the one closest to you. If you know the
address of the machine doing the testing you can get some idea of how close
by doing a "TRACERT hostname" This will show you a hop list and network
delay for short packets. If you have XP you can also use "PATHPING hostname"
which will use larger packets, generate more traffic, and so slows the
network itself down. (not all these tests do this to a certain extent but
PATHPING is I think among the worst)
> Thanks guys,
>
>
> Pete Loud
>
>
>
>
>
If you'd have read my post earlier, I sent you the direct link to that
site, would have saved you "finding a whole load of other broadband
test"!! Sorry I bothered..
From my earlier post,
" With the help of some nice peolpe in a newsgroup, I also changed the
MTU to 1458 and that seems to have made a huge difference. Might not
work for you but it seems to for me, unless it's a temporary blip!
If you haven't already, look here,under forums, it's a very active
and helpful site.
http://bbs.adslguide.org.uk/
HTH
Frank"
The problem with all these speed test sites is that you are not testing the
speed of your ADSL line but the performance of the networks between you and
the speed test site. By using different sites you will find varied results.
The most accurate way that I have found for a bandwidth test is by doing a
download from the ISP's public ftp server (if they have one) which should
only be a couple of hops away. Once you go beyond that then all sorts of
network performace / link utilisation factors come into play.
Chris.
You would think so, wouldn't you? Beware PIPEX users, the ftp server is
well off-site:
Tracing route to tornado.dial.pipex.net [158.43.192.19]
over a maximum of 30 hops:
1 15 ms 16 ms 16 ms 81-86-128-1.dsl.pipex.com [81.86.128.1]
2 16 ms 15 ms 16 ms 62-241-161-1.dsl.pipex.com [62.241.161.1]
3 15 ms 16 ms 16 ms POS4-0.GW1.LND9.ALTER.NET [146.188.63.105]
4 16 ms 15 ms 16 ms so-3-0-0.xr1.lnd9.alter.net [158.43.150.141]
5 16 ms 15 ms 31 ms POS3-0.cr1.lnd10.gbb.uk.uu.net
[158.43.150.97]
6 16 ms 16 ms 15 ms pos1-0.cr1.lnd6.uk.gbb.uu.net
[158.43.253.129]
7 31 ms 15 ms 32 ms tornado.dial.pipex.net [158.43.192.19]
Trace complete.
Oliver.
Well, I do a little bit better ..
C:\Documents and Settings\Chris>tracert ftp.karoo.co.uk
Tracing route to nemesis.kingston-internet.net [212.50.160.116]
over a maximum of 30 hops:
1 <1 ms <1 ms <1 ms collossus [192.168.3.1]
2 30 ms 29 ms 29 ms cust-kc2000.kingston-internet.co.uk
[213.249.144.137]
3 29 ms 29 ms 29 ms 10.102.224.33
4 28 ms 28 ms 29 ms 212.50.160.116
Trace complete.
Chris.