Broadband speeds - definitions:

[robgul]

My BT Business Broadband is running very sloooooowly - apparently there was a problem at the local exchange 2 weeks ago and according to BT Tech " . . will take a week to get back up to the contracted speed for my location"  [Which is 36-45 down,  5-7 up]

Running speed tests on the BT hub, independent tests give, as at today, about 24/25 down and 4/5 up.

What I don't understand is the result when I check the speed in my BT account about 10 mins ago - the results are:

BB Test (in account page)   Sorry. Can't reach 37.3 Mbps at the moment.
Broadband speed to your hub  Network Hub Device 10.7 MbpsDownload 4.1 MbpsUpload
Broadband speed to your device Network Hub Device 20.8 MbpsDownload 5.2 MbpsUpload

The Device icon in the first (slower) line is grey out.

I am testing from a PC to the BT hub via Wifi (with a repeater disc)

The definitions seem counter-intuitive - can anyone explain the different results please?  [BT's tech people have given conflicting replies during our several conversations]

[robgul]

. . . and as touch of total irony as I typed the OP an email arrived from BT telling everything was working* and asking me to rate the service!

*I assume that this was the replacement for the back up over the EE network where the "sub-hub" was replaced last week - it now is showing as "connected" but with fewer signal bars than it needs to actually fire-up should the copper/FTTC fail.

[Morat]

The first thing I'd do is used a cable to connect to the BT Router and run a speed test from there. Wifi introduces many variables and it would be best to eliminate those so you can be sure of which tree to bark up.

[robgul]

The first thing I'd do is used a cable to connect to the BT Router and run a speed test from there. Wifi introduces many variables and it would be best to eliminate those so you can be sure of which tree to bark up.

That's not really practical - an BT's own speed tests (when I speak to the techs) are around the same as I'm getting - it's the definitions on my results screen that don't make sense to me.

Network to Hub - to me would mean the speed coming through the wall to the router

. . . and the other measurement the speed to my device (i.e. the PC) from the ether

[Afasoas]

If your broadband still seems slow, I'd ask what is your latency like?

So bandwidth is how many simultaneous bytes can be downloaded. It's the number of lanes that the road has. (Crude analogly).
Latency is how long it takes to shift a backed from A to B, and back to A again.

From a PC, you can run cmd.exe, at the prompt type ping 8.8.8.8 and press the enter key.

This is from a Linux machine, but it gives an idea of what you should expect:

Code: [Select]

$ ping 8.8.8.8
PING 8.8.8.8 (8.8.8.8) 56(84) bytes of data.
64 bytes from 8.8.8.8: icmp_seq=1 ttl=58 time=19.8 ms
64 bytes from 8.8.8.8: icmp_seq=2 ttl=58 time=22.9 ms
64 bytes from 8.8.8.8: icmp_seq=3 ttl=58 time=17.9 ms
64 bytes from 8.8.8.8: icmp_seq=4 ttl=58 time=16.4 ms
64 bytes from 8.8.8.8: icmp_seq=5 ttl=58 time=19.1 ms
64 bytes from 8.8.8.8: icmp_seq=6 ttl=58 time=16.9 ms
64 bytes from 8.8.8.8: icmp_seq=7 ttl=58 time=17.4 ms
64 bytes from 8.8.8.8: icmp_seq=8 ttl=58 time=20.3 ms
64 bytes from 8.8.8.8: icmp_seq=9 ttl=58 time=20.0 ms
64 bytes from 8.8.8.8: icmp_seq=10 ttl=58 time=18.4 ms
64 bytes from 8.8.8.8: icmp_seq=11 ttl=58 time=21.4 ms
64 bytes from 8.8.8.8: icmp_seq=12 ttl=58 time=16.3 ms
64 bytes from 8.8.8.8: icmp_seq=13 ttl=58 time=19.7 ms
64 bytes from 8.8.8.8: icmp_seq=14 ttl=58 time=17.2 ms
^C
--- 8.8.8.8 ping statistics ---
14 packets transmitted, 14 received, 0% packet loss, time 13018ms
rtt min/avg/max/mdev = 16.308/18.844/22.939/1.904 ms
My latency is generally under 20 milliseconds.
It might useful to do this via an Ethernet connection and via WiFi, and compare the difference between the two. I'd expect no more than 5ms (and ideally less than 3ms) of extra latency via WiFi ... but with wider variation in the ping times.

[Kim]

" . . will take a week to get back up to the contracted speed for my location"  [Which is 36-45 down,  5-7 up]

This likely means the Dynamic Line Management has been reset: The DSL will automagically resync a few times with different parameters until it determines the optimal balance between performance and stability.  It necesarily takes several days, because DSL line performance is affected by factors that can vary over a 24 hour period.

What I don't understand is the result when I check the speed in my BT account about 10 mins ago - the results are:

BB Test (in account page)   Sorry. Can't reach 37.3 Mbps at the moment.
Broadband speed to your hub  Network Hub Device 10.7 MbpsDownload 4.1 MbpsUpload
Broadband speed to your device Network Hub Device 20.8 MbpsDownload 5.2 MbpsUpload

The Device icon in the first (slower) line is grey out.

Well that's clear as mud.  Does that mean a test from the outside world inwards?  (Presumably not, as the download and upload would be the other way round.)  What does "your hub  Network Hub Device" mean, and how is it different from "your device Network Hub Device "?

I am testing from a PC to the BT hub via Wifi (with a repeater disc)

As Morat says, this will obfuscate the issue.

The definitions seem counter-intuitive - can anyone explain the different results please?  [BT's tech people have given conflicting replies during our several conversations]

Presumably this test isn't something us mortals with proper ISPs can see?  I note that https://www.bt.com/help/broadband/fix-a-problem/test-your-bt-broadband-speed wants us to install an app, which is the sort of rubbish you get when you let 12 year olds with business degrees run an ISP.

[robgul]

This may make the results screen clearer:



My expectation is :

- the first line is the speed along the wire to my router
- the second line is the speed right through to my device (which ia PC on Wifi - as memntioned cable is not a practical possibility - the router and PC are about 70 feet apart, diagonally opposite corners of large-ish house and on different floors)

BT Tech insisted that the results are the other way round . . .  WTF?

I'm waiting for an(other) answer from them.

[Afasoas]

I suspect the extra "speed" is explained by the first bullet point in the small print?
4G boost could mean any number of things.

But like I said, this is bandwidth that we are discussing - and whilst it's pretty poor, it's not so bad that you'd actually notice unless there are multiple users who are streaming high bit rate content or making large downloads. Even with two at-home workers, many Teams/Slacks calls and a fair amount of streaming, our average bandwidth utlisation is tiny.

How responsive your Internet connection feels (how long YACF takes to load a page, for example) will depend on latency  - which I've explained how to measure above.

As Kim said, it might take a few days for Dynamic Line Speed to settle on what it deems to be the fastest reliable sync speed.

[Kim]

I agree with your interpretation.

Yes, the results are the unintuitive way round.  Assuming that's not a one-off fluke (some transient issue at the time of one test and not the other), it could suggest that the Hub is just not up to handling that much data^[1], in which case it's not a very useful indication of anything.  (There's mention of something called "4G boost" which sounds like some kind of devious bonding thing, but presumably that's only going to work if the device has its own cellular modem.)



_{[1] These speed tests being a measure of how quickly a large file can be transferred over the network means that they're affected by the performance of the computing device at each end.  If you grafted a Fast Ethernet adaptor onto a Commodore 64^[2], you wouldn't expect it to be able to download files at the full 100Mbps, and the same is likely to apply to a router that's specifically optimised for forwarding packets from one interface to another while using as little electricity as possible, and wasn't really expected to handle more than the occasional firmware upgrade locally.)
[2] Is there a Rule 34 equivalent for this sort of thing?}

[robgul]

I agree with your interpretation.

Yes, the results are the unintuitive way round.  Assuming that's not a one-off fluke (some transient issue at the time of one test and not the other), it could suggest that the Hub is just not up to handling that much data^[1], in which case it's not a very useful indication of anything.  (There's mention of something called "4G boost" which sounds like some kind of devious bonding thing, but presumably that's only going to work if the device has its own cellular modem.)



_{[1] These speed tests being a measure of how quickly a large file can be transferred over the network means that they're affected by the performance of the computing device at each end.  If you grafted a Fast Ethernet adaptor onto a Commodore 64^[2], you wouldn't expect it to be able to download files at the full 100Mbps, and the same is likely to apply to a router that's specifically optimised for forwarding packets from one interface to another while using as little electricity as possible, and wasn't really expected to handle more than the occasional firmware upgrade locally.)
[2] Is there a Rule 34 equivalent for this sort of thing?}

The 4G thing is a standby connection which only kicks in if the main wired connection drops - it's a separate gadget that works either ethernet cabled or wifi from the router and has an EE sim in it to connect to the EE cellular network

I have tested the BB speed, using BT's test platform, with and without the 4G thing powered and there's no difference.    When it is powered on the 4G only shows 1 of 3 possible signal bars which is, apparently, insufficient for it to work anyway!

[robgul]

. . . my current strategy is to see if I can get them to change my current contract (runs to 12 June 24) to FTTP at no cost/less cost.   The FTTP cabling is already at the pole across the street.

[Kim]

. . . my current strategy is to see if I can get them to change my current contract (runs to 12 June 24) to FTTP at no cost/less cost.   The FTTP cabling is already at the pole across the street.

That seems eminently sensible.  Even if you don't need the performance, FTTP seems to be a major win for reliability.  Power outages aside, I think ours has gone down twice since it was installed, and one of those was a firmware upgrade of the ONT.

[vorsprung]

The test results say that the speed from your phone or pc or whatever to the box the isp gave you "the hub" is 25Mbps and the speed from "the hub" to the outside word is 7.3Mbps

Which makes sense but those are terribly slow speeds for 2023

(25Mbps is just over 3 megabytes per second)

[Feanor]

The test results say that the speed from your phone or pc or whatever to the box the isp gave you "the hub" is 25Mbps and the speed from "the hub" to the outside word is 7.3Mbps

Which makes sense but those are terribly slow speeds for 2023

(25Mbps is just over 3 megabytes per second)

Yes.

I have taken the view that if you can't get either full-speed VDSL or FTTC then the only game in town is Starlink.
It Just Works.

I live in rural-ish Aberdeenshire, and many of my cow-orkers have complaints about the availability of decent broadband. When I suggest Starlink, all I hear is complaints about the cost. Yes, it costs. But you need to balance that against your requirements.

I have double costs, because I use the Starlink as an L2TP Tunnel to my actual ISP.

It just comes down to what you are prepared to pay for.
I'm in a situation that I don't have much choice.

[robgul]

This may make the results screen clearer:



My expectation is :

- the first line is the speed along the wire to my router
- the second line is the speed right through to my device (which ia PC on Wifi - as memntioned cable is not a practical possibility - the router and PC are about 70 feet apart, diagonally opposite corners of large-ish house and on different floors)

BT Tech insisted that the results are the other way round . . .  WTF?

I'm waiting for an(other) answer from them.

The speed in the image was a bit if freak - the average later in the day was around 23 instead of the 7

[Kim]

The test results say that the speed from your phone or pc or whatever to the box the isp gave you "the hub" is 25Mbps

That would be a sensible and useful thing for it to do.  But it should Shirley indicate that by greying out the 'Network' globe and dotted line between it and the Hub.  And not call it 'Broadband speed'.

[robgul]

UPDATE - BT Tech's own testing when I was talking to them got similar speeds to my testing - and a "High Resistance Fault" at their end, suggesting something at the cabinet, in the external box on my house (wire is overhead from a pole), or the internal box that the hub is connected to. 

An Openreach engineer is coming tomorrow . . . we'll see what happens.

[vorsprung]

The test results say that the speed from your phone or pc or whatever to the box the isp gave you "the hub" is 25Mbps

That would be a sensible and useful thing for it to do.  But it should Shirley indicate that by greying out the 'Network' globe and dotted line between it and the Hub.  And not call it 'Broadband speed'.

Yeah, the magic speed test screen is junk

But then again Internet providers don't make it clear that the "internet speed" depends on so many external factors that providing a number (like they do) is meaningless

They should be forced by law to provide a standardised set of metrics for things like
 - local loop latency
 - local loop bandwidth
 - ATM network route speed, your exchange to ISP endpoint. If it can route to alternative, that too
 - ISP to a real address in US/EU Central/Asia pacific latency and bandwidth
 - ISP to various shopping basket of providers like BBC, Google, Facebook (taking caching into account)

Stats should be max,min,average,variance,99 percentile etc

[robgul]

Another (final?) update: 

Openreach man came this morning, a bit of testing and there was a fault on one of the cables in a manhole round the corner in the next street - that's been fixed (replaced with one of the "spare cables" in the trunking) - speed now at  57 down, 7.9 up - the down being better than our contracted 37.5.

So all good then - hopefully.  But still planning to go to FTTP when current contract expires in June 24 . . . if I can get the price down.