FYI: http://forums.thinkbroadband.com/fibre/t/4134882-slo...

That's my experience. I graphed the results from Speedtest.net this morning and it's very clear. Between 9th May and 17th June results vary between 50Mb/s and 60Mb/s. After that (last week and a bit) they are all over the place but the bulk lies between 30Mb/s and 40Mb/s.

I haven't unlocked my modem yet but my profile remains at 76.76Mb/s.

In reply to a post by GaryPower:

I am now with BT with the up to 80/20 infinity 2 product.
Download speedachieved during the test was - 30.08 Mbps
For your connection, the acceptable range of speedsis 12 Mbps-69.83 Mbps .

Very similar to what I've been getting. Whereabouts are you? I'm in Brackley.

FWIW I've got nowhere trying to talk to my ISP. Basically if it's within that range it seems we just have to suck it up and live with it.

Edit:I see that you're in Luton. Not likely to be the same node then I wouldn't think. We hang off Birmingham last I heard. Although...hmm. My latency dropped at the same time last week as well. What latency do you get to London/Namesco according to Speedtest.net? I wonder if my routing changed from Birmingham to Milton Keynes? That might reduce latency and put us both on the same node. Dunno if that's even possible.

Edited by Andrue (Fri 29-Jun-12 13:55:21)

Andrue

At the risk of "teaching granny to suck eggs"!

You may have already done this in which case I would be interested in the results!

Have you tried streaming video to another PC ( or PCs) and then running the speedtest. If the speed remains the same this would indicate that it is not the line/connection but may be a PC issue or a router somewhere that is limiting individual streams to (below) a set value.( Same effect as contention but constant over time).

I believe that some of the speedtest issues on high speed connections are due to PCs abilities or configuration settings on PC's and across the network restricting individual PCs speeds. Some connections seem to be able to run 2 PCs both at the same speed as one PC.

We will see even more of these issues as people move to 330Mb. I think it is unlikely to affect most people's real world performance but will be visible when people use speed testers across the Internet.

In reply to a post by kitcat:

Have you tried streaming video to another PC ( or PCs) and then running the speedtest. If the speed remains the same this would indicate that it is not the line/connection but may be a PC issue or a router somewhere that is limiting individual streams to (below) a set value.( Same effect as contention but constant over time).

I'm not sure what you mean by that. I have a server in my house and PC to Server transfers are way above what even FTTC is capable of so I don't think it's my PC or router causing the issue.

Also there was a very definite speed drop as witnessed by the following speedtest.net results:

6/16/2012 7:47 AM,GMT,56.84,14.26,33,London,250
6/16/2012 9:37 AM,GMT,52.56,14.21,38,London,250
6/17/2012 7:37 AM,GMT,53.55,14.11,36,London,250
6/17/2012 7:42 AM,GMT,58.15,14.23,35,London,250
6/18/2012 4:51 PM,GMT,53.95,14.26,35,London,250
6/19/2012 4:47 PM,GMT,57.87,12.69,45,London,250
6/19/2012 6:13 PM,GMT,32.63,11.46,13,London,250
6/20/2012 3:21 PM,GMT,45.29,10.81,15,London,250
6/20/2012 3:22 PM,GMT,44.93,11.33,15,London,250
6/20/2012 9:10 PM,GMT,35.92,10.79,15,London,250
6/20/2012 9:11 PM,GMT,39.6,9.14,13,London,250

Also note the latency change (second highlighted column)

I understand the limitations of the 8xx routers and had allready checked the processor (and am mappting the processor utilisation via prtg) as well as checking it live when running the tests.

I have a minimal configuration currently, basic pppoe, nat etc and have not seen the processor above 50% at any point during the tests.

The speeds were also the same during install using laptop directly to the homehub and I have gone back to this to retest since

thanks for the suggestion though!

Regards,

Gary

Hi,,

Yes I am in Luton
Currently showing 16ms to the London Namesco test server on speedtest.

We would need a traceroute to see if our routes converge at some point.
I am guessing there is some major oversubscription somwhere.

At least I am still running far quicker than I used to!

In reply to a post by GaryPower:

We would need a traceroute to see if our routes converge at some point.
I am guessing there is some major oversubscription somwhere.

I'm kicking off one of those this evening. I'm not sure if it would show all the hops though. It could be the BTw/BTor routing that has changed rather than the 'customer visible' IP routing.

At least I am still running far quicker than I used to!

Me too and in fact my minimum speed is 1Mb/s faster than I got with Bethere

In reply to a post by GaryPower:

Hi,,

Yes I am in Luton
Currently showing 16ms to the London Namesco test server on speedtest.

We would need a traceroute to see if our routes converge at some point.
I am guessing there is some major oversubscription somwhere.

At least I am still running far quicker than I used to!

I've run some traceroutes and I'm seeing high levels of packet loss. Intriguingly I'm seeing 30% on the first hop into IDNet's network then a further 5% on the next hop. The first one could be down to BT's network I assume but the second one is presumably all down to IDNet. Either way it adds up to around 34% for almost every location I put in.

In reply to a post by Andrue:

In reply to a post by kitcat:

Have you tried streaming video to another PC ( or PCs) and then running the speedtest. If the speed remains the same this would indicate that it is not the line/connection but may be a PC issue or a router somewhere that is limiting individual streams to (below) a set value.( Same effect as contention but constant over time).

I'm not sure what you mean by that. I have a server in my house and PC to Server transfers are way above what even FTTC is capable of so I don't think it's my PC or router causing the issue.

As I read him he means have two, or even three PCs connected to the router, with one or two of them streaming direct from somewhere out there. Nothing to do with your inhouse server or LAN. Then do the speed tests on the remaining machine and see if they stay the same, or fall off due to the others using bandwidth.

Just out of interest, what does this test make of your connection ?