Kitz Forum
Announcements => News Articles => Topic started by: UncleUB on October 20, 2009, 08:27:35 AM
-
You can now find out what speed broadband your neighbour gets(Although at present the speedtest is down due to demand)
http://www.ispreview.co.uk/story/2009/10/19/test-snoops-on-the-uk-broadband-isp-speed-of-your-neighbours.html
-
I would like to know why ISP review thinks the smaller ISPs are not going to be as well represented when the street map example seems to illustrate the ISP's logo.
I do agree that it is necessary to include the package details for a better comparison.
However I am VERY disappointed that the myth of line length being a single cause for poor speeds is being continued, when much of that problem is caused by line quality.
It is the quality problem which is so hard to cure particularly with the fault procedures BT Openreach employ.
Kind rergards,
Walter
-
However I am VERY disappointed that the myth of line length being a single cause for poor speeds is being continued, when much of that problem is caused by line quality.
My next door neighbour has broad band and when I put his number into kitz line checker it comes back at 3000kbps.
When I was with BT mine was 1000kbps(it is now 2500kbps since I migrated to O2)
But both our lines come off the same telegraph pole and I presume then take the same route to the exchange
-
@ Uncle,
I think one of the problems with the BT line checks is that they rely on some historical fugure which all too often seems inaccurate and is often set far too high.
This might have been done originally to encourage people to get broadband, or possibly to show how "good" the network was.
I've recently been involved in a new broadband connection where the checker gave a figure of 6 mbps when the other line to that house had been "false greened".
After the usual wiring struggles we've now managed a steady IP Profile of 500 kbps - a minor miracle, or so the delirious occupant thinks !
Kind regards,
Walter
-
I'm not sure that the line check figure was deliberately set high on the BT checker, otherwise <insert name of regulator / competitor / minister here> would have come down on them like a ton of bricks for misleading the public etc.
This is pure speculation, but it may have been based on distance from the exchange and some coefficient to cater for going round the houses a bit. Unfortunately this probably couldn't take into account the way-longer-than-average routes that some lines take.
I also think that the checker does take into account the figure stored for lines to neighbouring properties, and it routinely seems to adjust itself based on neighbours. Perhaps the results are adjusted based on calculated / actual connect rates grouped by postcode area or property?
Not sure what happens with new lines now, but when I moved in here in 2004, the BT checker said 500kbps, red for 1Mb, since it was a new number. I managed after some pushing to get that raised up to 1Mb, and then 2Mb, and I'm now syncing at ~10Mb on ADSL2!
Whilst I was on BT max, I managed to sync at 8Mb pretty much all the time, yet the BT checker said 5.5Mb.
BT are their own worst enemies by not publicising enough information about how their network and planning works, and so what we "know" is really a collaboration of what people have deduced. However, whilst I'm certainly not questioning your findings, I'm unconvinced that false greens are a deliberate act by BT - they are more hassle for them after all. And surely there are enough conspiracy theories in the world already ;)
-
I actually find that the results of the BT line checker seem to change over time to track fairly accurately (with a lengthy delay) my actual line speed.
Before I 'sorted' my line (thanks to this forum), the BT checker showed my line as being similar or poorer than neighbours. Now it shows my line being much better than most of neighbours (which it now is). Recently, when I tolerated an appalling speed for a few weeks in order to get a DLM target reduction, that was also reflected in the BT checker.
I've never been sure whether that should surprise me or not, as I get the impression the BT checker is generally not thought to behave in that way? :-\
But if the BT checker does reflect true line speeds, then it's probably a far more effective way to snoop on your neighbours' performance, as it doesn't depend upon them having tun any optional speed tests.
-
@ MrChris,
Perhaps we just have to say this whole subject is steeped in the technological mysteries of the wonderful BT karma !
I wonder if Kitz or azzaka might like to add anything to this somewhat trivial saga.
To clarify I was using the direct BT tester currently on the Zen Web site, which doesn't include the extra data Kitz and SamKnows add re the exchange limitations.
I have just re-tested the two numbers that were involved, both having been attempted with Zen as the first and only ISP involved. The false green one last year and the other in September this year.
BOTH now say the lines are capable of 0.25 Mbps yet we know the one last year was false-greened at that time.
Kitz says they both have over 75 dB loss which I do believe as the line distance must be towards 7 km, and have radsl at amber.
The extrordinary puzzle is why did BT change the 6 Mbps estimate down to 0.25 when we all observe a steady IP profile of 500 kbps with sync speeds between 600 and 1000 kbps?
I should also add that when the order was placed on the 6 Mbps line BT discovered that it was part of a DACS circuit !
Kind regards,
Walter
-
The extrordinary puzzle is why did BT change the 6 Mbps estimate down to 0.25 when we all observe a steady IP profile of 500 kbps with sync speeds between 600 and 1000 kbps?
I just reallised, I should've said it seems to track my IP profile, not my line speed. Doesn't change the questions (and doesn't answer Walter's point), but worth clarifying.
-
Perhaps we just have to say this whole subject is steeped in the technological mysteries of the wonderful BT karma !
Definitely!
The extrordinary puzzle is why did BT change the 6 Mbps estimate down to 0.25 when we all observe a steady IP profile of 500 kbps with sync speeds between 600 and 1000 kbps?
Further to this and 7LM's comment, perhaps it tracks lowest IP profile achieved in the last xyz days, or even uses line training figures i.e. MSR/FTR rates? Who knows though??!
Maybe we should ask RAMBO? Maybe RAMBO be confused anyway! ;) Heh - I know it's a trivial thing, but it would still be good to know exactly how it worked!
Kitz says they both have over 75 dB loss which I do believe as the line distance must be towards 7 km, and have radsl at amber.
They're doing well to be stable at 500kbps - well done sir :)
I should also add that when the order was placed on the 6 Mbps line BT discovered that it was part of a DACS circuit !
Classic :lol: lol - 6Mb down a DACS line would be some achievement indeed!!
-
It can depend on whether you query the line using postcode or actual telephone number.
~ If its postcode, it will attempt to give you an average of what local lines in that area are expected to receive.
~ If its on the actual phone number.. it will if possible return a figure based on the IPprofile.. or IPprofiles for that area. This can fall down if theres been a number change, or if theres some erroneous figure in the database.
Each of these checks can return different results. If you put in both phone + postcode, then phone number will get priority as it is generally likely to be more accurate. If a line is LLU then for the line rate data it will return average info for the area.
Personally Ive always found the guestimate more likely to be conservative in their figures. This is likely due to not everyone having optimised their lines.
BT certainly do appear to use stored IPprofiles in their calculations.
However these figures can only come from lines whose services are provided via BTWholesale and theres no record base for the LLU lines.
They are usually far slower at increasing a line rate and it requires a very long period of stability to show any increase. Drops tend to show fairly quickly. The timescale seems to be a bit of an unknown quantity.
When on maxdsl, bar something like stuck profiles and electrical storms I had a constant sync of 8128 for 3+ years. The database for my own line showed 6500 and it only updated to 8Mb after I'd had a constant 8128 for a year.
When I went LLU it reverted back to 6500 within a couple of weeks. Its still reporting 6500 to this day - despite me being on a supposed 24Mb line.
-
>> 've recently been involved in a new broadband connection where the checker gave a figure of 6 mbps when the other line to that house had been "false greened".
Out of interest was that a new line?
I strongly suspect that the line data may relate to a re-allocated phone number. If the number was last used by someone who lived near the exchange, then this would most likely be the cause of the higher than physically possible data.
-
@Kitz,
No both lines have been at this address for several years. However the 6 Mbps one is a foreign group - i.e. NOT one of the local ones we're used to and the 01483 Guildford area is quite large.
That suggests the number was migrated to THCN at some time in the past from another sub-area.
Re the orignial subject of this news note (Sorry)
I've been having a go with this new speedtesting service, admitedly in the middle of the evening peak and comparing it with TBB on my Virgin Media 20 Mbps service.
The new one gives 5.2 mbps down and 0.7 up whereas TBB gives 18.2 mbps down and 0.7 upload.
This suggests there are still some teething troubles and I know which one I'd rather believe !
Kind regards,
Walter
-
It seems to be hopelessly inaccurate. First, it couldn't detect my ISP (Newnet, which isn't on their list) so I said I was with Zen. The test ran and gave me a download speed of 9.3 Mbps. Not bad for an IP profile of 4500 kbps. ???
-
Apart from the gross inaccuracies and inconsistent results, this approach seems of very questionable validity to consumers and providers alike.
E.g. given that Be and O2 are using the same equipment, the fact that O2's samples have been on poorer lines than Be's makes the whole thing seem a nonsense.
Below are two screen pictures from two test runs with the slower one taken immediately after the faster one.
(Note from above this is on my 20 Mbps Virgin Media service which TBB measured at over 18 Mbps.)
Kind regards,
Walter
Edit to correct the order and then to add the line in parenthesis.
[attachment deleted by admin]
-
Even if the results were accurate (and I'm happy to believe that they're not), they seem to be restricted to the results of other people who have run that test on that website. That's a self-selecting sample which largely invalidates any statistical conclusions that could be drawn.
Most obviously, it may be used by people who have an active interest in broadband, many of whom will already have spent time on improvements like I-plates, faceplates, or bell-wire snippers. But these people (us) are not typical of the overall population.
Still, I guess it's already fulfilling its design aim. Thanks to the publicity it'll get, we've all clicked on the link, as will have other people on other forums, hence generating traffic to a site that's cluttered with commercial advertising.
-
Hi
Tried it and gave slightly lower results than Speedtest.net DL 5400Kbps to 5800Kbps . UL both ~ 400Kbps .
Showing my DS to be Highest(ZEN) . Guess which is the lowest ?.
Regards Jeff
-
Even if the results were accurate (and I'm happy to believe that they're not), they seem to be restricted to the results of other people who have run that test on that website. T
I hadnt run it before as I wasnt too sure about the privacy aspect, and wonder about the IP to location thingy
Anyhow thought Id give it a shot & Ive just found out how inaccurate the tester appears to be
my results
* Download speed
* Upload speed
*
* You7.8
* Be
* 7.8Mb
* 0.2Mb
Ooops not very good are they? So I ran another one
Provider Download speed Upload speed You7.1
* Provider
* Download speed
* Upload speed
*
* You7.1
* Be
* 7.1Mb
* 0.9Mb
*
Compare with a quick visit to the TBB speedtester
Speed Down 16683.91 Kbps ( 16.3 Mbps )
Speed Up 2173.56 Kbps ( 2.1 Mbps )
:no:
-
Still, I guess it's already fulfilling its design aim. Thanks to the publicity it'll get, we've all clicked on the link, as will have other people on other forums, hence generating traffic to a site that's cluttered with commercial advertising.
I think you may be right.
-
@ Kitz,
The top10-broadband.co.uk site is so grossly inaccurate that ISPs would probably have a good case for banning the site.
It is hopelessly inaccurate and seems to suffer from capacity / bandwidth problems as well.
This morning it has just reached the giddy heights of 10.7 Mbps on my 20 M Virgin service which regularly reports between 16 and 18 Mbps on the TBB site.
I.e. anyone with a service of over 10 Mbps MUST be warned of the inccuracies and a statement to that effect must be included on the results page and all such results must be disregrded and marked as such.
How to get this done ?
Kind regards,
Walter
[attachment deleted by admin]