Explanation of Best Effort traffic classes in www.speedtester.bt.com report ?

[#1]

Hi,
we're all being asked to run the BT speedtester to help resolve sub-optimal network performance, but the report it presents contains a section that is impossible to relate to what expected line performance should be for each of the many Plusnet broadband product offerings - both active and legacy.
I, for one, would be very interested in an explanation of exactly how the sub-BE, normal-BE and priority-BE proportions should be apportioned for each product offering under normal operating conditions.
Knowing the expected distribution ahead of any fault would allow better diagnosis by customers if it looks wonky, surely ?
Anyone

[#2]

I think the three values should normally be approximately 5%:20%:75%, but in the best tradition of the Performance Tester if relative values differ significantly it simply means they differed.
Traffic shaping should not affect Performance Tester measurements, congestion might well do but I'm not aware of any statement that this should affect the different BT performance classes differently.
A handbook on the Performance Tester can be found here. The tester has recently been changed to cover FTTC and FTTH, but there doesn't appear to be a v7 version (yet) to cover these sevices.

[#3]

Thanks spraxyt,
for the link to the BT Handbook.
I was unaware of it's existence. I was hoping that it might help with our long standing issue with the tester
See this thread :- " BT speed checker U/S for several days « 1 2 ... 9 10 » "
No such luck however !
Perhaps  BT have yet another version which, for some of us , does not work at all !!!
Rgds,
edgeways

[#4]

Link Section 2.1. says

• 802.11b wireless networks which can only at best support 5.9 Mbps IP throughput

Feed my ignorance! Does that mean it can hamper performance with higher speed ADSL MAX or
2+ connections? What 802.11 type is recommended for most users?

[#5]

lucerne,
802.11b wireless was the first wireless standard commonly used in the UK and supports a theoretical maximum of ~11 MB/s so you can see if you had an Up to 20 MB/s (ADSL 2) connection then it could stop you getting the most out of your line.
802.11g supports a theoretical maximum of ~54 MB/s
802.11n has versions that can do 150MB/s and go up to around 300 MB/s
But they are all reliant on a good clean signal, so if you are in a built up area with lots of other wireless networks competing, or if you have a long distance between you and the router, or if there are thick walls in between that will reduce the speed of the wireless network.
So you want G as the minimum but N is normally going to be better.

[#6]

Quote from: spraxyt

Traffic shaping should not affect Performance Tester measurements

OK - now I'm confused. Aren't ALL residential product offerings 'Best Effort'  with only Business broadband being outside this BE classification scheme ? After all if you've bought a product with a SLA and expect QoS management, then 'Best Effort' is meaningless, isn't it ? So if BT Wholesale can offer differing proportions of what looks like their own internal de facto traffic shaping system, then how does an ISP map these offerings to their own products. Do we have a situation where an ISP would typically set up every one of its residential products on the same BT BE proportion set ?

Quote from: spraxyt

A handbook on the Performance Tester can be found here.

Good stuff ! Thanks...

Quote

CP’s may choose to use these markings to prioritise traffic being sent to End User. For example marking voice traffic with NBE, web browsing with PBE and file downloads with SBE.

and is this what Plusnet do in addition to their  Ellacoya per-protocol shaping ?