Peak time latency spikes

[#1]

As per my previous request for an answer why is it that my upload speed has not really changed compared to my download speed?
jim:quote

[#2]

Quote from: englishrick

ok still on ptw-ag04...and its shot up for some unknown reason ..

Ok opened my mouth toooo soon. still on ptw-ag04 and its down again even worse than before !!!

[#3]

I have no idea why the upload hasn't been as variable.

[#4]

Quote from: englishrick

As per my previous request for an answer why is it that my upload speed has not really changed compared to my download speed?

If an exchange is congested it is always the download that is the problem. I can't actually recall an example of upload being affected by congestion (other than if the download congestion is so bad the Acks can't get through) in the many years I've been participating in these forums.

[#5]

ok still on gateway plc-ag01 which work fine last night, but tonight speeds dropping by about 10Mbps down, getting about 42Mbps but not as bad as it has been will check again later

[#6]

englishrick - if you do pathping bbc.co.uk from the command prompt, can you post the results?
Upload speed - I think most people only use a small percentage of their upload speed so there is always capacity there. Download speed is always going to be more heavily utilised.

[#7]

If it's congestion at the exchange then the links out of the exchange have the same capacity on the downstream as well as the upstream. As most people have faster download speeds than upload and download more than they upload it takes an awful lot to cause a throughput issue on the upstream. We get a report twice a week showing the capacity at each exchange and if there are any problems, at the moment your exchange isn't showing on that report so we're querying that and also seeing if we can speak to the other customers at your exchange that are seeing the same.
The normal planning rules are that if BTW spot an issue at an exchange caused by lack of capacity they'll upgrade it within 3 days, however there will be cases where that's not possible and we're trying to find if that's the case but we aren't going to get an answer before Monday unfortunately.
Once we've found out then we'll know how long it will take to fix but if we're right it should be straightforward fix.

[#8]

Quote from: AndyH

englishrick - if you do pathping bbc.co.uk from the command prompt, can you post the results? Upload speed - I think most people only use a small percentage of their upload speed so there is always capacity there. Download speed is always going to be more heavily utilised.

Tracing route to bbc.co.uk [212.58.251.195]
over a maximum of 30 hops:
  0  Downstairs [192.168.1.227]
  1  www.asusnetwork.net [192.168.1.1]
  2  lo0-central10.ptw-ag04.plus.net [195.166.128.194]
  3  link11-central10.ptw-gw01.plus.net [84.93.248.116]
  4  xe-10-0-0.ptw-cr01.plus.net [212.159.1.40]
  5  kingston-gw.thdo.bbc.co.uk [212.58.239.6]
  6    *        *        *
Computing statistics for 125 seconds...
            Source to Here  This Node/Link
Hop  RTT    Lost/Sent = Pct  Lost/Sent = Pct  Address
  0                                          Downstairs [192.168.1.227]
                                0/ 100 =  0%  |
  1    3ms    0/ 100 =  0%    0/ 100 =  0%  www.asusnetwork.net [192.168.1.1]
                                0/ 100 =  0%  |
  2  ---    100/ 100 =100%  100/ 100 =100%  lo0-central10.ptw-ag04.plus.net [195.166.128.194]
                                0/ 100 =  0%  |
  3  23ms    0/ 100 =  0%    0/ 100 =  0%  link11-central10.ptw-gw01.plus.net [84.93.248.116]
                                0/ 100 =  0%  |
  4  27ms    0/ 100 =  0%    0/ 100 =  0%  xe-10-0-0.ptw-cr01.plus.net [212.159.1.40]
                              100/ 100 =100%  |
  5  ---    100/ 100 =100%    0/ 100 =  0%  kingston-gw.thdo.bbc.co.uk [212.58.239.6]

[#9]

Quote from: dave

If it's congestion at the exchange then the links out of the exchange have the same capacity on the downstream as well as the upstream. As most people have faster download speeds than upload and download more than they upload it takes an awful lot to cause a throughput issue on the upstream. We get a report twice a week showing the capacity at each exchange and if there are any problems, at the moment your exchange isn't showing on that report so we're querying that and also seeing if we can speak to the other customers at your exchange that are seeing the same. The normal planning rules are that if BTW spot an issue at an exchange caused by lack of capacity they'll upgrade it within 3 days, however there will be cases where that's not possible and we're trying to find if that's the case but we aren't going to get an answer before Monday unfortunately. Once we've found out then we'll know how long it will take to fix but if we're right it should be straightforward fix.

Ok thanks Dave

[#10]

20CN  on  'pcl-ag01'
Not off to a good start ! 

[#11]

Hi, just to say my speed is now back to normal, thanks for sorting that out for me.
My ping is still quite high, about double the usual ping......but that's probably due to the current problems (which don't appear to be getting any better going by the look of things on this forum) 

[#12]

2 night in a row very good am a happy chap Ping 18 upload 6.89 that the best I am able to get

[#13]

i'm on gateway pcl-ag06
Uptime: 0 days, 22:53:16
since last dropped session.
Fibre speeds are holding around 70mbs ping is 10

[#14]

20CN on 'pcl-ag01'
I have done many speed tests in the last hour and the results are all over the place !
Even during each and every test, the instantaneous speed is unstable.
Look at the plot below !

Many of my speed tests have been at 1.5Mbps for much of the test then risen to 6.5Mbps for a short burst,
unfortunately these speed test programs are reporting the peak transfer bursts and not representative of the majority of the test period.
Usually my speedtests progress fairly constantly, usually with one small peak or trough, tonight the bursty nature is a new problem. 

[#15]

Not looking bad from here:

IP Profile is 6Mbps