SkyUser Speed Tester Hits 20k of Tests

[NewsreadeR]

Since our inception of the SkyUser Speed Checker, today we have reached the milestone of over 20k tests.

Our results are only counted and graphed, if you are on a UK ISP's IP address at the time of the test, and then only once per 10 minute period.

Obviously the test is more Sky orientated because the majority of our visitors / members are already on Sky Broadband or a Sky TV customer and so we would expect to see Sky have the most results against them.

The results are as follows as at today:

Top Ten ISP's - Speed Wise

1) Be*  	 7795 Kbps  	  833 Kbps
2) Sky           4388 Kbps  	  405 Kbps
3) Zen           3909 Kbps  	  341 Kbps
4) Bulldog       3852 Kbps  	  370 Kbps
5) Newnet        3295 Kbps        473 Kbps
6) Entanet       3285 Kbps        523 Kbps
7) Plusnet       3086 Kbps  	  318 Kbps
8) IDNet         3037 Kbps  	  355 Kbps
9) TalkTalk      3033 Kbps  	  280 Kbps
10) Virgin Media 3006 Kbps  	  329 Kbps

What is surprising is, although Be* Virgin Media and Sky, all offer speeds of 16Mb and over, they are all well below that actual figure as a result.

As for all the up to 8Mb ISPs the exact same is apparent, leading me to the conclusion that the UK's existing telephony lines are no where near capable of delivering true high speed internet for the majority.

We appreciate that this is not directly the advertising ISP's fault, as the speed is dependant on the line quality / length.

[HSNetwork]

Bud,

One alteration I would have thought of as far as logging and displaying the results go is to display sky LLU and sky Connect as seperate figures.

Plusnet no longer offer or use LLU from what my sources tell me, and BE is only LLU (if I remember correctly unless using the O2 brand in which case IPStream is available).

The reason I say this is because there is a clear difference between LLU and IPStream both in download and upload and I fear that on a few of the above providers the figures are being hindered by the use of IPStream so we are not seeing an Actual True figure of what the average for an "Up to 16Mb" really is. Same goes for Virgin, Bulldog etc.

Is this something you would consider so we could all see what the true results are. This could perhaps be done by adding a little question box at the beginning of the test, eg;

Does your product offer Up to;
8Mb
16Mb
20Mb
24Mb Download speeds.

Does your product offer up to;
288 kbps
488 kbps
768 kbps
832 kbps
1.3 Mb
2.5 Mb Upload speeds.

The ISP is automatically trackable though.

[NewsreadeR]

Hi Tom

The reason I say this is because there is a clear difference between LLU and IPStream both in download and upload and I fear that on a few of the above providers the figures are being hindered by the use of IPStream so we are not seeing an Actual True figure of what the average for an "Up to 16Mb" really is. Same goes for Virgin, Bulldog etc.

We thought about this and decided against it. Reason being, for example I am on MAX with a theoretical UPTO of 16Mb. However I can only ever receive 7Mb tops. So even if I were to add the button and select UP TO 16Mb, it would flaw them results straight away, as I would be dragging the average down if I were to test my speed in that category.

The other consideration, very few people would know what UPSTREAM they were contracted to and without checking their router stats before taking the test would just be guessing.

Bearing in mind this Speed Tester gathers the results and then averages them per ISP, I feel that this is a pretty good guide / indication of the true state of play. Afterall I could get 8Mb on IPStream or LLU and at the end of the day, very few ISP's actually market the two different technologies, they market Speeds etc.

We do capture the UP and the Down speeds and I suppose that we could then split the results to show above 468Kbs as an upload guide, this would give you automatically which ones are on LLU for Sky as an example.

The test as it is now is a true average of those people who have taken their speed readings and so I would say is a fairly accurate state of play and a true reflection of that selected ISP's results.

The fact that ISP's offer upto 16Mb and above, is incidental, because the average speed is there in black and white. If more people were able to actually get them speeds, then the results would reflect that.

People could not make an informed choice or see their potential speeds for that particular product based on breaking down the results further still, and remember also not everyone could get an LLU based product.

The way I look at it in my case, I am above the average for Sky on Upload and Download speeds, but I myself am probably paying for more than I need, as I get nowhere near 16Mb.

I can probably arrange for the results to be tabulated or selected to show different aspects of what we capture, and I do think that the selection of an UPTO speed may be a good idea.

One way of doing this is like you say, ask what Speed you are on before the test. But would everyone know? or would they just guess? If a guess it would flaw the results straight away. An example of which is UPTO 2Mb which a huge majority of the UK are on as their ISP's have not MAXXED their lines, would they know that? or would they select UPTO 8Mb.

The simplest way of course to find out would be to add every single ISP's packages to a questionnaire at the end, IE We have detected that you are on Sky, which package are you on? Base, Connect, Mid or Max, and then collate them that way. This could then graph what you are asking. I can see if that is possible?

[HSNetwork]

The package idea would be better and I dont believe that the idea would be flawed. As you say your on up to 16Mb but you can only ever get 7. But this is the point, you are on this package and your receiving 7, obviously as far as results go line conditions etc would still have to be considered. But it gives a far better picture as apposed to the current 4388 kbps avg when sky offer up to 16Mb. I believe that just on Max products the avg would jump to around 9000 kbps or perhaps even 11000 kbps. which gives a far better picture of the avg speed for customers on the product having done the test.

My 2 cents worth.

[NewsreadeR]

You have got me thinking now.

Out of 15688 tests recorded with a Sky / Easynet ISP only 2040 people actually record over 8Mb as their result.

23 of these recorded speeds of over 16Mb
The average UPLOAD for these 2040 IP's is 592.99Kbps
The average DOWNLOAD for these 2040 IP's is 10.325Mb

But there are just too many factors to take into account, ie Bandwidth Overheads in the test etc

I think the way to go is for a person to select a package after or before completion.

[HSNetwork]

Yep

I reckon that would be a better idea. That then would at least alow you to split down the products with sky and give a better reading on what the avg speed is. If any other ISP wanted this doing I suppose they could sponsir you to have it done

[NewsreadeR]

Breaking this down further.

VM had 17 speed tests with the average speed over 10Mb out of 1052 tests taken, which is about 1.6%