Packet loss on the Plusnet network
Guess there's not much more to learn for now then 
I put a fault report in at the beginning of the week, on which there's been no progress since the immediate automatic tests, and I don't suppose there'll be any activity over the weekend.
One very unhappy bunny
I put a fault report in at the beginning of the week, on which there's been no progress since the immediate automatic tests, and I don't suppose there'll be any activity over the weekend.
One very unhappy bunny
So, yesterday I got a reply to the ticket I submitted on Tuesday. It had been checked as a speed throughput problem, when I was in fact reporting packet loss, so I'd been asked to perform a speedtest via the BT tester. I phoned up to explain that this would not help, since a) I wasn't reporting a speed problem per se, but a degradation in performance due to packet los and b) the tester would just show a speed within normal bounds (I'm seeing 5-6Mbit/s, which is normal for me).
I had a long chat with a very helpful chap, although we didn't resolve anything - he's just suggested switching interleaving back on, to see if it improves (this on a line that shows excellent parameters, and has had interleave switched off for years with no problem before). I'm of course happy to try anything to get my line back to normal performance.
The biggest problem I have is with VOIP applications (or maybe it's more accurate to say it's most noticeable, since if you talk and the person at the other end get's what you've said 10, 15, 30 seconds or even a minute after you've said it, then it's unuseable). It was using VEntrilo that I 1st noticed something was wrong - this was about a week and a half ago.
I've also noticed strange behaviour with my upstream in partcicualar. Where as downstream speed tests continue to give good numbers, and test as expected, upstream tests are less reliable. Upstream speed has been below what I've had in the past, and tests such as mybroadbandspeed.co.uk, just act addly - it often pauses whilst preparing the upstream test for like 30-60 seconds, then the progress bar moves as normal until the last 10% of the bar - and then it stops. Sometimes the test eventually completes (maybe a minute later) and some times is just hangs there or returns a file error. This just isn't right.
The results have got worse again this afternoon - and checking back on my previous comments, last Sunday afternoon/evening was especially bad. A coinicidence?
I've just managed to get a couple of speed test to complete, with these results:
As you see, downstream is fine, but upstream is often very poor.
I've also created a Thinkbroadband line quality monitor:
This clearly shows the 24/7 packet loss I'm experiencing and large variation in ping times.
From what I can see, the BT speed tester only checks your downstream throughput - or at least doesn't test the upstream if downstream looks ok (as mine does), so this test is useless in this situation. I guess I could Fraps mybroadbandspeed.co.uk running on my desktop to demonstrate the strange way it runs, as I cannot think of any other way to report this.
To be honest, I don't know what to do next. Yes, I need to see what switching interleaving back on will do, but I have little confidence that it will make any difference to the packet loss, since my line is not long (400m), and neither is it showing low quality issues. It's always been fine before, and surely it would affect both upsteam and downstream?
Does anyone else have a Thinkbroadband monitor set up, that they'd be prepared to share, to compare what you're seeing?
I had a long chat with a very helpful chap, although we didn't resolve anything - he's just suggested switching interleaving back on, to see if it improves (this on a line that shows excellent parameters, and has had interleave switched off for years with no problem before). I'm of course happy to try anything to get my line back to normal performance.
The biggest problem I have is with VOIP applications (or maybe it's more accurate to say it's most noticeable, since if you talk and the person at the other end get's what you've said 10, 15, 30 seconds or even a minute after you've said it, then it's unuseable). It was using VEntrilo that I 1st noticed something was wrong - this was about a week and a half ago.
I've also noticed strange behaviour with my upstream in partcicualar. Where as downstream speed tests continue to give good numbers, and test as expected, upstream tests are less reliable. Upstream speed has been below what I've had in the past, and tests such as mybroadbandspeed.co.uk, just act addly - it often pauses whilst preparing the upstream test for like 30-60 seconds, then the progress bar moves as normal until the last 10% of the bar - and then it stops. Sometimes the test eventually completes (maybe a minute later) and some times is just hangs there or returns a file error. This just isn't right.
The results have got worse again this afternoon - and checking back on my previous comments, last Sunday afternoon/evening was especially bad. A coinicidence?
I've just managed to get a couple of speed test to complete, with these results:
As you see, downstream is fine, but upstream is often very poor.
I've also created a Thinkbroadband line quality monitor:
This clearly shows the 24/7 packet loss I'm experiencing and large variation in ping times.
From what I can see, the BT speed tester only checks your downstream throughput - or at least doesn't test the upstream if downstream looks ok (as mine does), so this test is useless in this situation. I guess I could Fraps mybroadbandspeed.co.uk running on my desktop to demonstrate the strange way it runs, as I cannot think of any other way to report this.
To be honest, I don't know what to do next. Yes, I need to see what switching interleaving back on will do, but I have little confidence that it will make any difference to the packet loss, since my line is not long (400m), and neither is it showing low quality issues. It's always been fine before, and surely it would affect both upsteam and downstream?
Does anyone else have a Thinkbroadband monitor set up, that they'd be prepared to share, to compare what you're seeing?
This is what mine looks like:
This is now going beyond frustrating My connection is unuseable for VOIP, which I use most days!!!
My packet loss continues to be consistent 24/7, and yet I have a stable connection which syncs at the best rates available to me.
A current snapshot of my packet loss is :
Last Friday I asked the following question about what it would cost to leave Plusnet.
Right now, with the lack of any useful help from Plusnet (when I spoke to them on the phone on Saturday, the guy said that he'd never encountered a situation of severe packet loss on a good quality connection), I'm thinking I'll be better to buying myself out of my Plusnet contract, and taking my chances elsewhere. Yes, I realise that the fault might be evident with a new ISP, but right now I have no confidence that Plusnet have any idea what is going on, or how to correct it.
I'm willing to test anything else at my end, but with the problem being evident on 2 different routers, plugged into the BT test socket, and on 3 different PCs (2 wired, one wireless), 2 different routers and cables, I don't know what else could be eliminated at this end.
I've had fastpath for a few years with no problems, and the router has always held it's connection for weeks/months unless I reboot for a specific reason (such as last week when the Plusnet Service Status was advising reconnecting to bypass the network problems they were having).
My connection is unuseable for VOIP, which I use most days!!!My packet loss continues to be consistent 24/7, and yet I have a stable connection which syncs at the best rates available to me.
A current snapshot of my packet loss is :
Last Friday I asked the following question about what it would cost to leave Plusnet.
Quote from: webwisehost When I signed up, I deferred the £25 activation fee, as I viewed this as a long term sign-up, and had no intention of moving again any time soon, so the 12 month contract was not a problem. Can you confirm that if I migrated to another ISP now, all I would need to do is pay the deferred £25 activation fee and any remaining monthly subscriptions fees due before the migration was completed?
Right now, with the lack of any useful help from Plusnet (when I spoke to them on the phone on Saturday, the guy said that he'd never encountered a situation of severe packet loss on a good quality connection), I'm thinking I'll be better to buying myself out of my Plusnet contract, and taking my chances elsewhere. Yes, I realise that the fault might be evident with a new ISP, but right now I have no confidence that Plusnet have any idea what is going on, or how to correct it.
I'm willing to test anything else at my end, but with the problem being evident on 2 different routers, plugged into the BT test socket, and on 3 different PCs (2 wired, one wireless), 2 different routers and cables, I don't know what else could be eliminated at this end.
I've had fastpath for a few years with no problems, and the router has always held it's connection for weeks/months unless I reboot for a specific reason (such as last week when the Plusnet Service Status was advising reconnecting to bypass the network problems they were having).
Are you on a static or variable IP? I wonder if you are being attacked in some way.
Good morning Matt,
Hmmm, well it seems that someone may have kicked something overnight. I'm still on fastpath, but the line seems to have rest twice in the early hours of the morning.
BUT, since the line reset I've seen almost no packet loss!!!! The external monitors have shown a couple - but nothing continuous, or significant and pathping.ping.com are showing normal results:
I think it's a bit early to say the fault has gone altogether, since a) it's been less than an hour since I saw the improvements and b) I've not had a chance to test VOIP yet. It does, however, seem that it is nothing to do with interleaving. This is good, since ping is higher than I'd like, and interleaving will only make it worse, although I understand why you wanted to test the connection with interleaving off whilst the severe packet loss was evident.
I'll see how it hold up for the rest of the day, but as there is no packet loss evident at the moment, it would be wise to cancel the request to switch interleaving on - since with no packet loss right now, it's not going to prove or disprove anything!
If you rerun your tests now Matt, what does the errored seconds look like?
Hmmm, well it seems that someone may have kicked something overnight. I'm still on fastpath, but the line seems to have rest twice in the early hours of the morning.
Password: ********
Type ? for command help
> adsl status
--------------------------- ATU-R Info (annex A) ----------------------------
Running Mode : G.DMT State : SHOWTIME
DS Actual Rate : 8128000 bps US Actual Rate : 448000 bps
DS maximum Rate : 8128000 bps US maximum Rate : 448000 bps
DS Path Mode : Fast US Path Mode : Fast
NE Current Attenuation : 9.0 dB Cur SNR Margin : 10.5 dB
DS actual PSDM(C) : 00000000 US actual PSDM(R) : 00000000
ADSL Firmware Version : E.38.2.23
-------------------------------- ATU-C Info ---------------------------------
Far Current Attenuation : 6.5 dB Far SNR Margin : 25.0 dB
CO ITU Version[0] : 00000000 CO ITU Version[1] : 00000000
DSLAM CHIPSET VENDOR : < Unknown >
-----------------------------------------------------------------------------
INTL, DSTREAM bytes/depth= 0, symbols/codewords = 0, latency= 0*0.25ms.
, USTREAM bytes/depth= 0, symbols/codewords = 0, latency= 0*0.25ms.
CODE, DSTREAM fast parity bytes= 0, fast codeword= 0.
DSTREAM intl parity bytes= 0, intl codeword= 0.
CODE, USTREAM fast parity bytes= 0, fast codeword= 0.
USTREAM intl parity bytes= 0, intl codeword= 0.
Setting>>
FDQ : Enable, TCM : Enable, EC : Disable, Framing Mode : 3.
Running>>
FDQ : Off, TCM : Off, EC : Off.
DS : Framing Mode 3, US : Framing Mode 3.
ADI ADSL Firmware Version: E.38.2.23
BUT, since the line reset I've seen almost no packet loss!!!! The external monitors have shown a couple - but nothing continuous, or significant and pathping.ping.com are showing normal results:
C:\Users\Neil Gardner>pathping bbc.co.uk
Tracing route to bbc.co.uk [212.58.224.138]
over a maximum of 30 hops:
0 Neils-QuadPC [192.168.1.14]
1 my.router [192.168.1.1]
2 lo0-central1.pcl-ag03.plus.net [195.166.128.238]
3 gi1-6-311.pcl-gw01.plus.net [84.92.6.64]
4 po4.pcl-gw02.plus.net [212.159.1.129]
5 te2-2.thn-gw2.plus.net [212.159.0.190]
6 rt0.thdo.bbc.co.uk [212.58.239.25]
7 212.58.238.149
8 virtual-vip.thdo.bbc.co.uk [212.58.224.138]
Computing statistics for 200 seconds...
Source to Here This Node/Link
Hop RTT Lost/Sent = Pct Lost/Sent = Pct Address
0 Neils-QuadPC [192.168.1.14]
0/ 100 = 0% |
1 0ms 0/ 100 = 0% 0/ 100 = 0% my.router [192.168.1.1]
0/ 100 = 0% |
2 141ms 0/ 100 = 0% 0/ 100 = 0% lo0-central1.pcl-ag03.plus.net [195.166.128.238]
0/ 100 = 0% |
3 151ms 0/ 100 = 0% 0/ 100 = 0% gi1-6-311.pcl-gw01.plus.net [84.92.6.64]
0/ 100 = 0% |
4 130ms 0/ 100 = 0% 0/ 100 = 0% po4.pcl-gw02.plus.net [212.159.1.129]
0/ 100 = 0% |
5 153ms 0/ 100 = 0% 0/ 100 = 0% te2-2.thn-gw2.plus.net [212.159.0.190]
0/ 100 = 0% |
6 148ms 0/ 100 = 0% 0/ 100 = 0% rt0.thdo.bbc.co.uk [212.58.239.25]
0/ 100 = 0% |
7 144ms 0/ 100 = 0% 0/ 100 = 0% 212.58.238.149
0/ 100 = 0% |
8 127ms 0/ 100 = 0% 0/ 100 = 0% virtual-vip.thdo.bbc.co.uk [212.58.224.138]
Trace complete.
C:\Users\Neil Gardner>
I think it's a bit early to say the fault has gone altogether, since a) it's been less than an hour since I saw the improvements and b) I've not had a chance to test VOIP yet. It does, however, seem that it is nothing to do with interleaving. This is good, since ping is higher than I'd like, and interleaving will only make it worse, although I understand why you wanted to test the connection with interleaving off whilst the severe packet loss was evident.
I'll see how it hold up for the rest of the day, but as there is no packet loss evident at the moment, it would be wise to cancel the request to switch interleaving on - since with no packet loss right now, it's not going to prove or disprove anything!
If you rerun your tests now Matt, what does the errored seconds look like?
Good morning Ben - and thanks for picking this up.
Since the problem seems to have resolved itself overnight, is there any point in doing a Wireshark capture? I'm more than happy to do so, but if there are currently no symptoms of the problem, will this achieve anything?
Furthermore, will switching interleaving back on at this point prove anything either way, since there is nothing to improve? The external monotors I've been running show nothing abnormal on my line with respect to packet loss, after the line reset at 2am this morning.
As I explained when I phoned on Saturday, I was happy to have interleaving switched back on temporarily, to test if this improved the situation - but would not want this to be a perminent change unless it was unavoidable, as I have it enabled for the gaming latency benefts.
Incidentally I've had ping -t bbc.co.uk running for over an hout now, without on lost packet!!!
Could we please cancel the request to have interleaving switched on in the meantime, on the assumption that the fault has resolved itself (or maybe it was a generic problem somewhere in the exchange/BT network that was addresses as part of other work overnight).
I'll leave it a couple of days before closing the ticket, just in case this is a horrible joke the BT network is playing on me, and it comes back
Cheers, Neil
Since the problem seems to have resolved itself overnight, is there any point in doing a Wireshark capture? I'm more than happy to do so, but if there are currently no symptoms of the problem, will this achieve anything?
Furthermore, will switching interleaving back on at this point prove anything either way, since there is nothing to improve? The external monotors I've been running show nothing abnormal on my line with respect to packet loss, after the line reset at 2am this morning.
As I explained when I phoned on Saturday, I was happy to have interleaving switched back on temporarily, to test if this improved the situation - but would not want this to be a perminent change unless it was unavoidable, as I have it enabled for the gaming latency benefts.
Incidentally I've had ping -t bbc.co.uk running for over an hout now, without on lost packet!!!
Could we please cancel the request to have interleaving switched on in the meantime, on the assumption that the fault has resolved itself (or maybe it was a generic problem somewhere in the exchange/BT network that was addresses as part of other work overnight).
I'll leave it a couple of days before closing the ticket, just in case this is a horrible joke the BT network is playing on me, and it comes back
Cheers, Neil
Static IP activated free of charge for you, it'll take effect when you next reboot / drop connection 