Little peak time latency humps
I was thinking less about the size of the packets, more about the number of them. Every customer pinging every second would be an extra 600,000 packets per second to handle. Or 1.2 million if you count both the requests and the replies. What unfortunate destination should we all ping to test this theory? 
I usually test pings on bbc.co.uk (well, I do pay a TV license fee to support them) 
Quote from: ejs .......... What unfortunate destination should we all ping to test this theory?
I usually ping 212.159.6.10 just to make sure it's working, but any one of the four would do
Slightly more seriously, as I don't have a TBB monitor running atm, I don't know if the attached helps, on pcl-ag07 20CN, the first is what I was seeing yesterday evening peak, normal pings in the past are more like the later part of the second graph. You can ignore the extreme peak at 2204-2206 & the strange one at 2238-2240 both due to local "activity".
Here's some ping graphs from where I was away (all PC's off - only the SamKnows monitoring box running). The dates are on the images. I've been on pcl-ag01 throughout.
This one I was away from mid afternoon onwards:
This one I was away from mid afternoon onwards:
I have no ping test history before February this year. But the hump is visible then. E.g.:
I was on pcl-ag01 all that day.
I was on pcl-ag01 all that day.
Quote from: jelv I have no ping test history before February this year.
Is the same for me, graph image just won't load before Feb. 1st this year. TBB must have lost them or something.
Well that's at least 4 months, and in that time what i wonder has actually been done about it,? I have submitted a ticket regarding this and the associated reduced/erratic throughput that occurs usually within 1hr of the peak time jitter fest or during it , and although i consider that i have give enough detail to support regarding this they don't agree, for a start ping-plotter pro images are a very useful way of providing information of tracert and ping combined plus they show jitter levels also, and a ping graph also
But they seem to want less detail ie the rudimentary windows tracert and ping
along with wire shark captures , as im behind a router does it mean that i would have to connect direct to the pc instead of via the router to show what traffic priority HTTP is being given during peak times, If the answer to that is yes then this would create a problem, as the gateway would change therefore the results i have taken wouldn't be current any longer ,
It seems to me there is far too many obstacles thrown in the way and hoops to be jumped through in order to get a known issue looked at , not impressed
But they seem to want less detail ie the rudimentary windows tracert and ping
along with wire shark captures , as im behind a router does it mean that i would have to connect direct to the pc instead of via the router to show what traffic priority HTTP is being given during peak times, If the answer to that is yes then this would create a problem, as the gateway would change therefore the results i have taken wouldn't be current any longer ,It seems to me there is far too many obstacles thrown in the way and hoops to be jumped through in order to get a known issue looked at , not impressed
