Strangely enough their "latest throttling technologies" only have any effect on the TBB tester. I did look at the Talk to the Staff thread but there was no definitive solution to why the TBB tester and only the TBB tester consistently delivers lower speeds than others which it most certainly does for me regardless of port and time.

It seems to be OK for at least some 30Mbps users and obviously any ADSL connection is going to be slower still but for me on 50Mbps it doesn't give results which are of any use to me.

Edited by kwikbreaks (Sun 21-Aug-11 16:49:00)

In reply to a post by kwikbreaks:

It seems to be OK for at least some 30Mbps users and obviously any ADSL connection is going to be slower still but for me on 50Mbps it doesn't give results which are of any use to me.

This was the thread I meant in case you saw a different one:
http://forums.thinkbroadband.com/newsite/f/4025053-v...

I've made it work on a 200 megabit office connection, and it was reliable, and many people post it works on FTTC.

VM is radically different in each region - the thread in TTTS was talking about possible problems with VM's Linx routers too. Capacity perhaps.

TBB's speedtest is the only one that shows the real throughput (watch the graph on the flash tester); whereas a lot of the others, including Ookla based ones (speedtest.net and/or bbmax) use the peak speed. If you have a 100 Mbps connection and peak at 90 but mostly run at 20 due to congestion, the ookla testers tell you you have 90. TBB will say 20.

Anyway the real test is to download real files from well connected sites, Oracle Database from Oracle.com is a few gig - or big downloads from Microsoft's download centre - and time the downloads. tbbMeter is also useful to show the throughput too, completely separate to the web testers.

Edited by jchamier (Sun 21-Aug-11 17:40:14)

There's a graph on the speedtest.net one too plus I ran a networx graph alongside the tests and posted links to the images above. They all clearly showed that the TBB test delivered less speed.

This is my last post on this topic as I know for certain that the TBB test is inaccurate for me although I don't know why. Just how well it works for other people is completely irrelevant to me.

Edited by kwikbreaks (Mon 22-Aug-11 06:28:36)

Post deleted by wanapoo

Edited by deleted (Mon 22-Aug-11 08:11:14)

its worth pointing out your office connection probably has a lower stable rtt than the typical cable broadband connection. The "it works for me" doesnt mean its not broken.

I am still waiting for tbb to say what congestion avoidance algorithm they using.

VM also do not shape at 6.30am, I have done numerous tests and I have not had anything indicating throttling at that time.

Answer from me, if it is me who you are waiting for is I don't know.

They're throttling again tonight - dunno why but I'm down to a quarter of normal speeds


Been like that for the last hour plus. I only d/l during the night, if I d/l at all, and have just been doing browsing and email today.

MyVirginMedia pages are down, call centre is busy, nothing short of pathetic yet again. Crazy thing is, my usual indicator is working - the VOIP service is on, strange, that's normally first thing to go.

I didn't notice anything out of the ordinary last night - and this morning is fine. Both times theTBB rest went through without the hesitations on the upload part .

just done one and bizarrelly faster than 6am.



I got a speedtester on a freebsd server which had the newreno algorithm (old poor one), freebsd 8.2-STABLE added modern ones and when I switched to cubic a 100mbit VM user went from 25-40mbit to 85mbit.