That's the thing though, I've even tried explaining how things like streaming, windows updates, just normal everyday file downloads are badly affected by this but they just refuse to accept it's their problem.

I've explained it to the point that it doesn't make sense to me that they can't understand it.

The best I get is someone new to the problem saying something like 'I agree this does seem like a problem', only for them to go away for a couple of days then come back with something along the lines of 'I've spoken to the networks team and they said the problem is off-network', at which point it all goes silent and I get ignored for another couple of months.

The cynic in me thinks they do understand what I'm saying quite well but are themselves being stonewalled from higher up. That's what it seems like at least.

I think it would be very interesting if you showed results per area for VM peak vs off peak, if you can get over this thing you have against small sample set's you can make TBB a very valuable tool for consumers.

VM have been like this for years, including the NTL days.

Generally speaking if you in a bad area, they will either tell you its acceptable service, or just give you a ongoing discount and say take it or leave it, when they offer the discount they also typically tell you a upgrade is forthcoming, but 9 times out of 10 it doesnt happen, typically the only upgrades I ever witnessed was the one's carried out to match rollout's of new packages. I dont recall ever having an upgrade done on my node for "capacity management" alone.

What seems to have happened with this vivid rollout is some more corner cutting, someone flicking a switch without upgrades or upgrades have been done to capacity but not to the same scale as the speed increase and as such increasing congestion, even to previously good area's.

Regarding single threaded I think isp's had a point 10 years ago when operating systems had static sized tcp buffers meaning if a OS wasnt tuned, then the buffer could be too small to max out a connection single threaded, and regional CDN's were not the rage so downloading was often done over the atlantic compounding the buffer problem, however now every modern OS auto tunes the tcp buffer so at least on the client side single threaded performance should be adequate. My experience is on a good isp, mainstream services such as youtube, netflix and microsoft updates can max out my FTTC line single threaded.

Now there may be a valid argument e.g. if on a gigabit connection a single thread might only reach a few hundred mbit, but that wouldnt be noticeable to the end user and not break anything, a single threaded speed low enough to break services (generally sub 10mbit), is an isp side problem most of the time. Isp's insisting on multi threaded speed tests generally have something to hide.

The danger of small sample sizes is even if the data is supporting end user experience someone will roll out an 'expert' saying the results are not statistically significant and thus can be ignored.

i.e. a fine line to walk, but will revisit over the summer and see if a regional split gives enough data. The problem is not the peak time, but getting enough repeat people testing at the off-peak times.

It may be that for some larger VM areas I have enough to do something useful.

In London, North West, I would fall below 10Mbps every evening. There must be statistics showing this... & that was 10Mbps on speedtest.net (multi threaded). There are sadly some very significant congestion issues on virginmedia.

Very hard to prove I know.

now you explained it this way I understand where you coming from, no doubt someone will try to discredit the results I agree.

Well I'm back in Brum this week, using the Virgin Media connection that for the past 3-4 years has been solid.....

Seems that bandwidth boosts in the are have taken their toll and now not just single thread slow downs, but multiple thread slow downs too...

I did hav to chuckle though, I went on the Virgin Community Forum's to see if anyone else was complaining and found they where, the funny part was where Virgin wanted somebody to switch to modem mode and thy the connection again...

However what I know is that switching to modem mode changes the allocation on Virgins systems meaning you can get assigned to a less or non congested part of the network...

In reply to a post by mlmclaren:

However what I know is that switching to modem mode changes the allocation on Virgins systems meaning you can get assigned to a less or non congested part of the network...

It does no more than rebooting the Superhub.

EDIT: As far as the cable network goes that is. It changes IP address as the MAC address of the device getting the IP has changed, it's no longer the Superhub itself getting it but the device behind it.

Edited by deleted (Thu 30-Jun-16 15:26:37)

In reply to a post by MrSaffron:

The danger of small sample sizes is even if the data is supporting end user experience someone will roll out an 'expert' saying the results are not statistically significant and thus can be ignored.

i.e. a fine line to walk, but will revisit over the summer and see if a regional split gives enough data. The problem is not the peak time, but getting enough repeat people testing at the off-peak times.

It may be that for some larger VM areas I have enough to do something useful.

@MrSaffron: Have you been able to find any more patterns on this issue? I've ended up switching ISP as VM still haven't even acknowledged the problem. A couple of friends nearby are still on cheap VM packages so I can still see nothing has changed e.g. 8-9Mb/s on 50Mb packages.

Not had time to code up the regions into the peak/off peak system.