Peak time speed loss (19:00 to 22:00)
Since getting fibre, I have been seeing large speed drops in the SamKnows (SK) whitebox speed tests at peak times (images 3 and 4 are fairly typical). Yesterday (25th) I decided to run a few BT speed tests at randomly selected times that did not clash with other activity and two results (date/time in image name) are in image 1 ("normal" speeds; within 2%) and image 2 (the worst result); the "test3" option failed; probably a DNS problem 
The four "dips" in the SK down speed graph are (L2R) 21:00, 21:00, 20:00, 21:00; the three up are at 20:00, 20:00 and 21:00.
At the moment, I have not identified anything that switches/powers-up/creates noise only during the period in question*.
Apart from getting faults involved, any suggestions on the way ahead?
Phil
* Nor at other times
The four "dips" in the SK down speed graph are (L2R) 21:00, 21:00, 20:00, 21:00; the three up are at 20:00, 20:00 and 21:00.
At the moment, I have not identified anything that switches/powers-up/creates noise only during the period in question*.
Apart from getting faults involved, any suggestions on the way ahead?
Phil
* Nor at other times
@Chris,
Yes, all tests wired - I rarely use wireless (apart from infrequent "on-demand" TV
)
I've raised a fault ticket to see what happens
Phil
Yes, all tests wired - I rarely use wireless (apart from infrequent "on-demand" TV
)I've raised a fault ticket to see what happens
Phil
Right, tests have shown the set up to be running normally, with "no fault found" on the BT network ; a 80/40 "mismatch" somewhere has been corrected and I've been asked to run "routerless" to see if that eliminates the disconnections (which will require cable adjustments as I've not completed my revised infrastructure layout and will be done before 18:00 today). The former is not a surprise, if the tests were not run between 19:00 and 22:00, whilst the latter is, as the router has reported continuous connections since moving to fibre apart from "forced" drops, such as a group of about 4 when trying to get the BT speedtest "test3" to operate (about 2.5 days ago!). Bear in mind that I reported a 40 to 60% "speed/throughput" reduction during a specified time interval and not session drops.
Any possibility of seeing the Radius data in case the modem resyncs without the session dropping?
Ooops - should have checked the ticket!
Phil
Notes:
The SamKnows "drop" at 21:00 last night was 11%, but I have had smaller drops previously (see graph 3 above)
I declined the offer of a TG582n at no cost/postage as the TP-Link has been reliable and has Gb ethernet; also, as a last resort, I have a USB powered 11n300 "portable router", with a single WAN/LAN connection, that could be pressed into action
Edit: PPPoE connected at about 1450
; a 80/40 "mismatch" somewhere has been corrected and I've been asked to run "routerless" to see if that eliminates the disconnections (which will require cable adjustments as I've not completed my revised infrastructure layout and will be done before 18:00 today). The former is not a surprise, if the tests were not run between 19:00 and 22:00, whilst the latter is, as the router has reported continuous connections since moving to fibre apart from "forced" drops, such as a group of about 4 when trying to get the BT speedtest "test3" to operate (about 2.5 days ago!). Bear in mind that I reported a 40 to 60% "speed/throughput" reduction during a specified time interval and not session drops.Any possibility of seeing the Radius data in case the modem resyncs without the session dropping?
Ooops - should have checked the ticket!
Phil
Notes:
The SamKnows "drop" at 21:00 last night was 11%, but I have had smaller drops previously (see graph 3 above)
I declined the offer of a TG582n at no cost/postage as the TP-Link has been reliable and has Gb ethernet; also, as a last resort, I have a USB powered 11n300 "portable router", with a single WAN/LAN connection, that could be pressed into action
Edit: PPPoE connected at about 1450