Little peak time latency humps

[#1]

http://community.plus.net/ping-graphs/
So, there's been a little noise over the past year about the little peak time latency humps we've been seeing on the TBB graphs.  We didn't tend to have them this time last year, so something has changed but we've not pinned down the cause exactly.  Sometimes they aren't there at all!
Being as a bunch of you guys are pretty obsessive at looking at these graphs and trying to decipher them, can you help us understand them?
At peak, if you notice a hump, can you do a couple of things for us.
1) Stop your traffic and keep an eye on the graph.  We know that the humps look very similar to active internet use.  We want to identify situations when the humps are appearing with no customer side traffic.
2) Check which gateway you are on.
3) Grab a screenshot of the graph, and pop on it your community username, gateway and highlight on it when you weren't using the connection.  Pop a date on it too.  (it's annoying that they don't have that on there!.
4) Open a dos prompt and do "ping -n 1000 8.8.8.8"  then copy the ping summary once it has finished.  We are trying to see if you are seeing impact from the hump on your performance
5) Post the image and the ping summary it on this thread.
Also, do you have any historical graphs showing this peak hump missing?  It would be useful for us to look at your connection and work out what gateways you were connected to at the time.

[#2]

Here's one for you,
From Yesterday 13/06

I know for a fact that there was no traffic in use at that time, as I watch tv then (after 19:30) and nothing was being downloaded. (OK modification there was a 128Kb streaming radio running  :))
Couldn't tell you which gateway, probably the one I am still on pcl-ag02.

[#3]

Hairy Biker:  You are connected to PLUSNET3-WBC.Faraday PCL-AG02 and have been for 6 days.

[#4]

Is that the kind of thing you are looking for Kelly?

[#5]

Yep.  I've added a ping test too to the request.

[#6]

Other thing to those who've got the TBB ping graph set up, do you know how far back the history goes? If you can scan back and see if there are any patterns (e.g. it doesn't happen on a Saturday or started in November or something)?

[#7]

Do you really want people to ping 8.8.8.8?  I thought the google DNS server  addresses were "magic" and so don't necessarily go to the same place every time.
Might you be better pinging to some known actual physical host? 

[#8]

I almost started a thread about this topic yesterday,so nice timing !  

My graphs have been pretty much flat minimum and average latency for months when I was on "BBYW-Opt3".

Having changed to "Unlimited" a week ago, I've had the 'hump' EVERY evening !

I believe both above graphs were on gateway "PCL-AG01".
The latency 'hump' since being on 'Unlimited' is NOT causing ANY issues, but is an interesting observation.
My 'SamKnows' monitoring results pattern also changed on product switchover, but as yet I do not enough results to report any meaningful observations.
Results from speedtest.net don't seem to be out of the ordinary.

[#9]

I only noticed it first time 2 days ago, which is why I started this thread: https://community.plus.net/forum/index.php/topic,115707.0.html (Keep in mind I've only been a customer for just over 2 weeks though)
I don't believe it happened before that as I would have noticed it while playing games

[#10]

Brad:  Let me ask that back.  Are you seeing 8-9ms latency on those pings during the day, which rises up to 8-30ms at peak hours (19-22:00 ish?)

[#11]

Kelly, that is correct. It only went up during those hours - and back down again afterwards.

[#12]

Can you test that tonight and Sunday night again and try not to disconnect your router and report back here.

[#13]

Some times i get a reduced throughput a short time before the usual 8pm-11.30pm peak time (jitter hump) here is  a speed test result from a short time ago http://www.thinkbroadband.com/speedtest/button/137123391207758818034.png you can also see the fluctuation in speed to, which causes a yellow trace on the think broadband quality meter, because the latency rises during the download, but if i download the same file at after 12am i get full throughput and there is only a very small  yellow trail created infact sometimes there isn't any, because the latency remains at the normal base level, So IMO the two are in some way connected ,
For  me  Saturday 8th June13 was Free from the usual jitter hump  and iirc there was no reduction in throughput  that night  Sunday's and Mondays tend to be more severe  
Also if i use utorrent  which i don't frequently use, although not maxing out the line upstream or downstream at any point ,it appears to generate a single spike of packet loss
Whilst it's possible that it's down to my router getting swamped and failing to respond to ICMP due  the number of connections p2p generates , there may be another reason for this , does similar happen to anyone else ?
http://www.thinkbroadband.com/ping/share/dea32a342610415e0b8514bec49d6fb1-01-06-2013.html Saturday 1/6/13 jitter present  GW pcl-ag06
http://www.thinkbroadband.com/ping/share/d750bb528e17d8e830650f0f49ab6a2c-08-06-2013.html Sat 8/6/13 same GW
http://www.thinkbroadband.com/ping/share/7e35f7de1d0257f0649efeccfafd9639-09-06-2013.html Sun 9/6/13  I got 2 jitter humps, some  those small yellow spikes at other times have been common ,not actual usage
http://www.thinkbroadband.com/ping/share/5595a77dbab086d10751db8909f20e45-10-06-2013.html Monday 10/6/13  same GW   throughput was ok even though there was more Jitter
http://www.thinkbroadband.com/ping/share/fcd3b46d097f548b2cb94d8f8e170a2b-13-06-2013.html yesterday , throughput was affected but only by a little,
Just done another speedtest  http://www.thinkbroadband.com/speedtest/button/13712378995474574895.png Normal throughput again , BTW i have also tested throughput by using other speed-testers and actually downloading several  file's (HTTP)

[#14]

Only wandered past to check on the latest excuses with regards to webstats (or should I say lack thereof), but while here I just happened to stumble across this, so here's my 2p worth ...
This has been happening for ages and quite regularly too although not everyday that's for sure. I haven't correlated any of the problems I seem to have to any specific gateway and/or particular data centre to date ... but then again I've just about given up with it all TBH. If my fully hard-wired connection is 'acting up' in some way then I just wander off and leave it to b*gger around as it sees fit these days unless I'm desperate. I've wasted way too much time and effort trying to persevere with it or attempting to get to the bottom of whatever it is that causes problems but ISN'T down to me or my modem etc.
Here's a typical TBB graph for when I was several hundred miles away from my totally unoccupied premises for instance:

and here's another random and particularly bad one but this time with perhaps some intermittent light browsing going on but nothing of any significance if indeed anything at all. I tend to avoid doing anything of importance in peak times whenever I possibly can these days or end up giving up and trying again later.

In both cases, apart from routine and very slow mail checks at 5/10 min intervals or whatever, there's no real activity whatsoever. The 'humps' are quite often considerably worse than that as well but sometimes only just detectable. It's quite variable for no obvious reason other than presumably (over)loading somewhere.
Add to that the problem with breaks in service leading to having to terminate PPP but then experiencing problems with automatically re-establishing a connection due to LCP issues almost every time I'm DLing iplayer (or other heavily managed data BUT at no other time) much like this:

in addition to the peak time 'hump' and there clearly is a problem somewhere beyond my front door and it's clearly been around for absolutely bl**dy ages. But I've pretty much given up worrying about it now. Just accept it as being all part of the so-called 'service' and take the easy option I'm afraid. I either babysit my connection when I absolutely have to or simply don't use it when it's being stupid. Ridiculous situation really when paying £22/m for very limited BW but it keeps the old blood pressure down and all that  
If anyone's running a book ... my money is on BT primarily responsible although with PN the merest hint of a nose behind  Not that I'm hedging my bets or anything like that of course    Best of luck trying to pin it down to something or someone Kelly, I bet it's going to be a right old PITA to say the very least !

[#15]

Well @9.54pm data flow more or less stopped so not possible to even browse , (no bandwidth) for nearly 3mins  also a lot of packet loss and increased base latency , followed by reduced throughput  55-60 mbps  did something fail ?