topic Re: Affect on speeds of the Plusnet profile in Plusnet Feedback

Re: Affect on speeds of the Plusnet profile

ejs — Sun, 30 Mar 2014 17:35:21 GMT

I've started a plan to re-test the issue myself. Because I can't just directly change the plusnet profile myself, I'll have to:
1. increase my line speed
2. wait for plusnet profile to increase
3. decrease my line speed
4. measure the throughput before the plusnet profile decreases
So far I have succeeded with step 1, and have been waiting for step 2, which is guaranteed to happen in a timely and reliable manner (hashtag sarcasm). An unexpected by-product of the experiment so far is that my line seems adequately stable at the higher speed.
[tt]US Connection Rate: 882 DS Connection Rate: 4019
DS Line Attenuation: 57.9 DS Margin: 6.1
US Line Attenuation: 36.3 US Margin: 5.2[/tt]

Re: Affect on speeds of the Plusnet profile

ejs — Sun, 13 Apr 2014 06:17:59 GMT

So I finally got around to doing the test:
PN profile 3.5
sync 4059
throughput 403KiB/s - 158385999 bytes in 6m 23s
PN profile 3.5
sync 3739
throughput 392KB/s - 158385999 bytes in 6m 34s
I was downloading firefox-28.0-3.fc20.src.rpm with wget.
I didn't do some worthless ping test at the same time because that would be completely missing the point of this thread - that the PN profile seems to limit my speed to less than what the line is evidently capable of. Also because I don't particularly care about the performance of ICMP echo requests and replies.

Re: Affect on speeds of the Plusnet profile

eblakeborough — Sun, 13 Apr 2014 09:49:07 GMT

Based on those sync rates, that profile is the correct profile by the way.
Your ip profile is always 88.2% of your sync rate. our profile matches BT's (the majority of the time)
On the second one the profile hasn't updated yet.

Re: Affect on speeds of the Plusnet profile

ejs — Mon, 14 Apr 2014 12:05:06 GMT

Quote from: Ed
On the second one the profile hasn't updated yet.

Yes that was the whole idea of those measurements. I'm waiting to do the third test, once the Plusnet profile updates to 3.2. Then the throughput won't be the line speed of 392KiB/s, it will go down to I expect about 360KiB/s. The Plusnet profile may be set to the same number as the BTWholesale profile, but its affect on the speed is not the same.
392KiB/s is approximately 3.21Mbit/s which is just under the BTWholesale 88.2% IP profile for the sync speed of 3739.
Perhaps not all the layers of protocols encapsulated within other protocols are not being accounted for correctly when applying the PN profile, or perhaps Plusnet just decide to knock 5% off all ADSL customers' speeds to reduce Plusnet's bandwidth requirements in the hope people won't notice the slight discrepancy.

Re: Affect on speeds of the Plusnet profile

ejs — Wed, 16 Apr 2014 06:12:03 GMT

The PN profile eventually updates, so I can do the third test.
PN profile 3.5
Limited by PN profile: yes
sync 4059
throughput 403KiB/s - 158385999 bytes in 6m 23s (3.308Mbit/s)
PN profile 3.5
Limited by PN profile: no
sync 3739
throughput 392KB/s - 158385999 bytes in 6m 34s (3.216Mbit/s)
PN profile 3.2
Limited by PN profile: yes
sync 3739
throughput 368KiB/s - 158385999 bytes in 7m 0s (3.017Mbit/s)
Not quite as bad as I thought but the difference is still greater than the 100Kb rounding margin given. So, why can't the PN profile limit the speed to what the line is actually capable of? How do you get away with calling it Unlimited broadband?
And it's going to be difficult to try to explain away the difference by arguing that the speedtester is measuring the speed wrong, since I didn't use a speedtest website, so unless you are going to try to argue that it didn't actually take wget 26 seconds longer to download the same file. Perhaps a load of good sounding but meaningless technical gibberish about protocol overheads, which are exactly the same for all of the three measurements? Oh dear, I did all the tests over wireless, perhaps you would like to blame the wireless?

Re: Affect on speeds of the Plusnet profile

chrispurvey — Wed, 16 Apr 2014 08:33:53 GMT

Quote from: ejs
How do you get away with calling it Unlimited broadband?

It's Unlimited due to no usage restrictions not speed.
The PN profile eventually updates, so I can do the third test.

Quote
PN profile 3.5
Limited by PN profile: yes
sync 4059
throughput 403KiB/s - 158385999 bytes in 6m 23s (3.308Mbit/s)
PN profile 3.5
Limited by PN profile: no
sync 3739
throughput 392KB/s - 158385999 bytes in 6m 34s (3.216Mbit/s)
PN profile 3.2
Limited by PN profile: yes
sync 3739
throughput 368KiB/s - 158385999 bytes in 7m 0s (3.017Mbit/s)
Not quite as bad as I thought but the difference is still greater than the 100Kb rounding margin given. So, why can't the PN profile limit the speed to what the line is actually capable of?
And it's going to be difficult to try to explain away the difference by arguing that the speedtester is measuring the speed wrong, since I didn't use a speedtest website, so unless you are going to try to argue that it didn't actually take wget 26 seconds longer to download the same file. Perhaps a load of good sounding but meaningless technical gibberish about protocol overheads, which are exactly the same for all of the three measurements?

I'll flag this up with Dave, he's not in today though.

Quote
Oh dear, I did all the tests over wireless, perhaps you would like to blame the wireless?

Regardless, it is always more accurate over a wired connection.

Re: Affect on speeds of the Plusnet profile

chrispurvey — Wed, 16 Apr 2014 10:21:29 GMT

Can you try those downloads again as you're on one of the recently added gateways and we can see if there's a difference?

Re: Affect on speeds of the Plusnet profile

ejs — Wed, 16 Apr 2014 11:31:24 GMT

Quote from: Chris
Quote from: ejs
How do you get away with calling it Unlimited broadband?

It's Unlimited due to no usage restrictions not speed.

Even that is factually incorrect. Calling it "Unlimited Broadband" requires both speeds and usage to have no restrictions.
I can't do another test right now, probably the next convenient time will be early tomorrow morning, plus I already put by sync speed back up.

Re: Affect on speeds of the Plusnet profile

ejs — Wed, 16 Apr 2014 21:01:29 GMT

I don't know what gateway I was on when I did the test early this morning, I did not record it, although I was on ptn-bng01 shortly afterwards when I posted the results on the forum.
After reconnecting and ending up on pcl-ag02, I saw that the previous file download was not running at a completely maximum speed 100% of the time, so I had to download something else that did.
[tt]wget http://download.microsoft.com/download/6/E/F/6EFA2958-EF2B-4DD1-B3BC-B891CB7C83DC/Windows6.0-KB947821-v33-x86.msu
--2014-04-16 21:34:50-- http://download.microsoft.com/download/6/E/F/6EFA2958-EF2B-4DD1-B3BC-B891CB7C83DC/Windows6.0-KB947821-v33-x86.msu
Resolving download.microsoft.com... 23.61.255.25, 23.61.255.17
Connecting to download.microsoft.com|23.61.255.25|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 150518130 (144M) [application/octet-stream]
Saving to: ‘Windows6.0-KB947821-v33-x86.msu’
100%[======================================>] 150,518,130 370KB/s in 6m 38s
2014-04-16 21:41:28 (369 KB/s) - ‘Windows6.0-KB947821-v33-x86.msu’ saved [150518130/150518130][/tt]
I was not planning to mess about testing every single gateway, moving my sync speed up or down, and then waiting a few days for my PN profile to go up or down. And I already reproduced the same thing back in August 2013, which presumably pre-dates your new gateway.

Re: Affect on speeds of the Plusnet profile

ejs — Thu, 17 Apr 2014 06:03:04 GMT

Well apparently I got kicked off pcl-ag02 at about 22:47, so when I re-tested the download speed this morning I noticed a tiny difference, and then saw that I'm now on ptw-bng01
[tt]--2014-04-17 06:42:26-- http://www.mirrorservice.org/sites/dl.fedoraproject.org/pub/fedora/linux/updates/20/SRPMS/firefox-28.0-3.fc20.src.rpm
Resolving www.mirrorservice.org... 212.219.56.184
Connecting to www.mirrorservice.org|212.219.56.184|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 158385999 (151M) [application/x-redhat-package-manager]
Saving to: ‘firefox-28.0-3.fc20.src.rpm’
100%[======================================>] 158,385,999 366KB/s in 7m 2s
2014-04-17 06:49:29 (367 KB/s) - ‘firefox-28.0-3.fc20.src.rpm’ saved [158385999/158385999]
--2014-04-17 06:49:37-- http://download.microsoft.com/download/6/E/F/6EFA2958-EF2B-4DD1-B3BC-B891CB7C83DC/Windows6.0-KB947821-v33-x86.msu
Resolving download.microsoft.com... 23.61.255.25, 23.61.255.17
Connecting to download.microsoft.com|23.61.255.25|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 150518130 (144M) [application/octet-stream]
Saving to: ‘Windows6.0-KB947821-v33-x86.msu’
100%[======================================>] 150,518,130 366KB/s in 6m 41s
2014-04-17 06:56:19 (366 KB/s) - ‘Windows6.0-KB947821-v33-x86.msu’ saved [150518130/150518130][/tt]
And I watched those downloads to check they ran at full speed the whole time, which they did. I conclude the difference between gateways so small that it is only just measurable and cannot possible account for the at least 22 KiB/s lost due to the PN profile.

Re: Affect on speeds of the Plusnet profile

Kelly — Thu, 17 Apr 2014 09:15:57 GMT

EJS: Feel free to bite my head off, but I'm wondering if it might make sense to start a new thread with your test results so far on this. I really want to get to the bottom of this, but there a lot of noise on this thread, and with a new one, we could ask the mods to be quite aggressive about keeping it on topic.

Re: Affect on speeds of the Plusnet profile

Oldjim — Thu, 17 Apr 2014 09:21:16 GMT

@ejs
If you want to start this in a new thread it might be better to split off the relevant posts from here
If you can advise the ones you would like splitting off either in this thread or by PM I will do it

Re: Affect on speeds of the Plusnet profile

ejs — Fri, 18 Apr 2014 09:06:15 GMT

I supposed it would make sense to split off replies 65, 69, 70, 74, then 77 onwards to a new thread.

Re: Affect on speeds of the Plusnet profile

Oldjim — Fri, 18 Apr 2014 09:38:27 GMT

Done
If you would like the title changing please advise
I might add that this thread brings everything back to my first post in the original thread - and no I am not going to try to sort it out

Quote from: Oldjim
Why does the Plusnet profile knock so much off the line speed
Example
Sync 4,000kbps BT IP profile 3,528kbps Plusnet profile 4,100kbps

Sync 4,000kbps BT IP profile 3,528kbps Plusnet profile 3,500kbps

I don't want some anodyne answer I want a proper explanation and a reason why it can't be fixed

Re: Affect on speeds of the Plusnet profile

ejs — Sun, 27 Apr 2014 21:03:31 GMT

I suppose I should approximately add the TCP and IP headers in an attempt to make the numbers more readily comparable to the IP profiles. Each IP packet has a 20 byte IP header, a 20 byte TCP header and up to 1460 bytes of payload data. 158385999/1460 = 108484 IP packets. 158385999+(108484*40) = 162725359 bytes including the TCP and IP headers. That makes the results:
PN profile 3.5
Limited by PN profile: yes
sync 4059
IP layer: 162725359 bytes in 6m 23s = 3.399 Mbit/s
PN profile 3.5
Limited by PN profile: no
sync 3739 (88.2% gives BT Profile of 3.297 Mbit/s)
IP layer: 162725359 bytes in 6m 34s = 3.304 Mbit/s
PN profile 3.2
Limited by PN profile: yes
sync 3739
IP layer: 162725359 bytes in 7m 0s = 3.100 Mbit/s
Those figures look almost as if the plusnet profile is the BT profile, rounded down to the nearest 100k, and then has 100k subtracted.
(I also just looked at some of the statistics in wireshark, but wireshark indicated 164244640 bytes which looks like it was including the 14 byte ethernet layer on each packet.)

Re: Affect on speeds of the Plusnet profile

Anonymous — Sun, 27 Apr 2014 21:53:01 GMT

I have a question.
If the customer can set their home gateway MTU value to anything they like (up to 1500 on PPPoA), this potentially means the PPP frame could contain anything from zero padding to whatever the worst case value is.
Does the 88.2% value assume that MTU is set to give optimal throughput (i.e. zero padding) - in which case downloads should occur at profile speeds for home gateways set to ideal MTU values such as 1430 or 1478, at the risk of the ISP sending too much data to the exchange if the MTU causes the frame to contain padding.
Or does the 88.2% limit throughput for connections as if they had the worst case padding, so as not to send too much data to the exchange in case the home gateway MTU is not configured to an optimum value ?, and therefore connections with optimal MTU will never achieve profile speeds because the percentage is too low.
I'm sorry if this makes no sense, because I've had more than a drink or two , but I think my question is valid.
Unfortunately I'm too far gone to even consider trying to do the maths

@'ejs' - have you optimized your MTU, and checked for zero padding using a tool such as DSLReports : Tweak Test ?

Re: Affect on speeds of the Plusnet profile

ejs — Mon, 28 Apr 2014 05:58:26 GMT

Which part of that tweak test tool do you imagine checks for zero padding of ATM cells? Fiddling with the MTU makes no difference to the transfer efficiency reported by that tool. That tweak test is going to have no idea if your packets are being chopped up and stuffed into ATM cells or not unless you tell it so by choosing pppoa after the test. I think I got a transfer efficiency reported as 93% or 94% regardless of a MTU of 1478 or 1500 because there were a few packets lost reported during the each test I did just now. I expect if I had zero packet loss during a test it would report 100% transfer efficiency regardless of MTU setting.
The vast majority of the overhead that the 88.2% IP profile allows for is the 5 byte ATM cell header, 48/53 = 0.906.
I did try a MTU of 1478 but found it made very little difference last time I checked, so all the above tests were done with a MTU of 1500.

Re: Affect on speeds of the Plusnet profile

adamwalker — Mon, 28 Apr 2014 14:18:43 GMT

@purleigh

Re what you asked in reply #15 I've deferred that question to someone who should be able to get us an accurate answer. We'll get back to you ASAP.

Re: Affect on speeds of the Plusnet profile

ejs — Mon, 28 Apr 2014 17:38:53 GMT

On further consideration, for 21CN at least, it may be that the impact of the Plusnet profile is the reason why a MTU of 1478 makes no difference (or even makes things slightly worse, smaller packets, more packets, more TCP/IP headers) compared to 1500.
Based on this, I'll try to represent the protocols as:
[tt] PPPoA over ADSL PPPoE PPP over L2TP over IP (over E?)
Me <------------------ MSAN <-------- MSE / BRAS <---------------------------------- Plusnet
[/tt]
For 21CN, the only section that involves ATM cells is the ADSL link between the end user and the exchange. If Plusnet send packets towards the MSAN slower than what the line is capable of, it doesn't matter if the MSAN has to pad some of the ATM cells with zeroes, because the bandwidth of this section isn't being fully utilised anyway.

Re: Affect on speeds of the Plusnet profile

ejs — Sat, 10 May 2014 06:55:33 GMT

I suspected the ancient dslreports tweak test java applet wasn't working well with Linux default TCP window scaling.
default:
[tt]# cat /proc/sys/net/ipv4/tcp_adv_win_scale
1
# cat /proc/sys/net/ipv4/tcp_window_scaling
1
# cat /proc/sys/net/ipv4/tcp_rmem
4096 87380 6291456[/tt]
disable tcp window scaling and increase rwin (could not be bothered to calculate some exact supposedly perfect value:
[tt]# echo 0 > /proc/sys/net/ipv4/tcp_adv_win_scale
# echo 0 > /proc/sys/net/ipv4/tcp_window_scaling
# echo "40960 873800 6291456" > /proc/sys/net/ipv4/tcp_rmem[/tt]
Result: zero retransmissions and 100% transfer efficiency reported by tweak test tool, with 1500 MTU.

Anyway, it should be fairly straightforward to test if the profile speed discrepancy is due to rounding down and then subtracting 100k. All it needs is for a couple of other people, preferably on 21CN ADSL, with different profile speeds (higher or lower) than mine of 3.5, to do some speed measurements. Just download a fairly large file, that takes at least 5 minutes to download, and watch the download to ensure it runs at full speed the whole time. Results needed:
1. IP profile
2. File size in bytes
3. Time taken to download
4. MTU if you've decreased it, or I'll just assume the default MTU of 1500 and MSS of 1460
Or someone from Plusnet could just confirm or deny that the profile is rounded down to the nearest 100k and then has another 100k subtracted (just to make double extra sure that too much data never ever gets sent down your line, of course).