Er . . . what?

Reports of average bb speeds have been perfectly satisfactory until today. However ThinkBroadband has suddenly gone bananas,. reporting 1.59mb/s on its single D/L check and 9.61mb/s on average over six. U/L report says 14.79mb/s and latency 25ms.

BTW check carried out at the same time gives 36.37mb/s D/L and 12.88mb/s U/L with 25ms latency. Ookla shows 37.82mb/s D/L with 14.99mb/s UL and 17ms latency on a very local server.Pinging BBC returns 15ms and PLusnet 28ms. These have all been very typical reading on my 40mb/s Plusnet fibre service for ages, testing via ethernet as usual and running under Win8.1.

ThinkBroadband says the new anomaly may be due to RWIN, which means nothing to me even when reading through the voluminous notes obviously intended for techies.

I am not aware of any other changes apart from the fact that routing has morphed from 87.***.***.*** to 91.***.***.*** for no apparent reason.

Could someone kindly interpret this weirdie and tell me what to do about it please?

BTW, have now run ThinkBroadband three times after allowing time and rebooting. Similar oddity each time.

Edited by deleted (Fri 05-Feb-16 18:26:49)

This is one of your tests yes?
http://www.thinkbroadband.com/speedtest/results.html...

RWIN or congestion is the too reasons, and congestion looks most likely.

Others on PlusNet running the same test version around the same time are getting good results, so rules out our end or network we run
http://tbb.st/145469704137886334791
Have also reviewed through the other tests in the last day or two and you seem to be the one with bad results.


What does our other test at http://labs.thinkbroadband.com/speedtest and https://labs.thinkbroadband.com/speedtest report (one is http the other is HTTPS) and post a copy of the share link at the end.

Thanks for addressing this.

Congestion? First attempt was at approx 11a.m. Fourth and latest has just been done under web address after logging in displayed as
http://www.thinkbroadband.com/speedtest.html?utm_sou...
and shows almost identical results. But will try the other addy given.

Don't understand how tests in last few days look bad to you. My records of ThinkBroadband tests show average as expected this month - 37.29 D/L, 13.20 U/L and latency 20. No 'foreigners until today.

http://www.thinkbroadband.com/speedtest/button/14547...

And someone has suggested that it may be down to the change from 87 to 91. Is this one of those which are chased off via the USofA?

Although I rebooted pc, I haven't rebooted router. Is that worth trying?

Latency is too low to be going to the US and back.

A router reboot might be worth a shot, or if everything else is behaving as normal just ignore it

If could be that you are on a congested part of the Plusnet network, and as we are more sensitive to congestion in our testing it may show up before other tests.

Have also looked at others with similar IP addresses on Plusnet and people look ok

I believe I truncated the previous address I gave. You may care to try this

http://www.thinkbroadband.com/speedtest/results.html...

Being slow of study I cant understand why both U/L and results from elsewhere look OK.

Rebooted Plusnet/Technicolor router, which has reverted from 91.***.***.*** to 87.***.***.*** and the performance to as before, i.e. approx 37 D/L and 14 U/L.. So how strange is all that?

It is not the IP as such that is the problem, but you have probably joined a different gateway, i.e. Plusnet is so large that there is not just one way onto their network and out to the wider world.

Why other tests are not showing issues, some tests will seek out the one good route from an ISP i.e. pick from the best of multiple servers, which can give you good results sometimes even when things are bad, or in some cases pick a server on the providers own network.

Also the way the tests are written, some will use more downloads at once to try and over come congestion, we've settled on six as gets past most local issues such as wi-fi problems but still sensitive enough to pre-warn of ISP issues. Also why we store a copy of the graphs too, as looking at those can tell as much as the raw figures when used to reviewing them.

In reply to a post by MrSaffron:

This is one of your tests yes?
http://www.thinkbroadband.com/speedtest/results.html...

RWIN or congestion is the too reasons, and congestion looks most likely.

Or the infamous single/double threaded discrepancy issue. I've had two of those this year, the first in several months.

http://community.plus.net/forum/index.php/topic,1477...

Although it is normally a peak-time problem which doesn't quite fit.

Edited by Andrue (Sat 06-Feb-16 08:57:06)