Slow broadband - network problem?

[#1]

Hi @ec5778

It's definitely a strange one.

If you could try the other router just to rule out the Hub Zero that would be good.

After that point in time, we'd potentially need to escalate this back to our suppliers, though I'm unsure what value a further engineer would provide at this stage.

Best wishes

Dave

[#2]

I've now swapped to the other router.

These are the stats I'm getting from the router:

Uptime:    0 days, 0:15:43
DSL Type:    ITU-T G.992.5
Bandwidth (Up/Down) [kbps/kbps]:    1,179 / 9,275
Data Transferred (Sent/Received) [MB/MB]:    8.93 / 31.40
Output Power (Up/Down) [dBm]:    12.8 / 0.0
Line Attenuation (Up/Down) [dB]:    18.8 / 36.0
SN Margin (Up/Down) [dB]:    6.8 / 6.0
System Vendor ID (Local/Remote):    TMMB / ----
Chipset Vendor ID (Local/Remote):    BDCM / IFTN
Loss of Framing (Local/Remote):    0 / 0
Loss of Signal (Local/Remote):    0 / 0
Loss of Power (Local/Remote):    0 / 0
Loss of Link (Remote):    -
Error Seconds (Local/Remote):    18 / 0
FEC Errors (Up/Down):    0 / 0
CRC Errors (Up/Down):    2 / 26
HEC Errors (Up/Down):    1 / 35

I notice a small number of errors showing up there.

Nevertheless overall it is slightly worse than with the plusnet one - I had to refresh the speedtest site three times before it actually gave a reading. But when I did get a reading it was 3.29Mbps which is far less than what I had previously and far less than the sync speed as reported by both routers and the Openreach engineer who came yesterday.

[#3]

Two other suggestions that have come up from searching around:

1. REIN - although this doesn't have the classical characteristics of high number of errors it's still a possibility.

2. Equipment or capacity issues at the exchange - could this be why the speed test estimate for the line has been reduced?

[#4]

Hi @ec5778

VLAN is showing green across the board - I would recommend raising this as a new fault and we'll ask our suppliers to take another look. It may be that something is amiss with the line that the previous engineers have failed to spot, it may be REIN (though the symptoms so far make it unlikely) - either way, I don't feel that there's much else you (or I) can do at this stage.

If you let us know when you raise a new fault we'll be able to escalate this for you.

Best wishes

Dave

[#5]

Some things jumps out, well to me anyway

The engineer did come this time, but unfortunately he wasn't able to work out what the problem was, he said its something to do with Plusnet as his equipment was only showing him the same readings as the routers' sync speed and it wasn't taking into account the actual throughput when trying to use the internet.

I'm assuming it was lower than what you normally see? If this was the case, this could be a lazy OR engineer, but based on my experience, and my firm but fair attitude, when this happened to me, fairly recently, when the OR guy attended my property and he saw lower than expected speed results, we chatted about it and we both agreed that he would start to work back from my property to all the joints in the cable until such time he observed a greater speed. I would have thought your engineer would have done the same.....

The same engineer turned up as last time. He was not impressed to be sent back!

He ran all the same tests on the line and said there's nothing he could do. He said he's only responsible for the actual wires between the property and the exchange and as they appear to be in order the problem must be to do with Plusnet equipment.

Again, he could be lazy, but surely if he used his equipment and observed a lower than expected speed, again surely he should have worked backwards....

Now my connection is FTTC and yours is ADSL, but the same principles apply - low speed and ideally a well trained decent engineer would hunt down why.

[#6]

How do I raise a new fault? I tried the link in message 10 but it just says

Thanks, you've given us all the information we need to investigate your problem

A Ticket has been raised to our Faults Team with these details. We'll contact you as soon as we have more information and let you know what happens next.

---

Meanwhile:- my internet now seems to have slowed right down. I tried the same test from my work computer using Internet Explorer and I got a result that seems much more accurate:

Downstream Test:  -provides background information.


 Download  Speed
 0.8 Mbps

   

0 Mbps 21 Mbps
Max Achievable Speed


 Download speedachieved during the test was - 0.8 Mbps
 For your connection, the acceptable range of speeds is 4 Mbps-21 Mbps.
 IP Profile for your line is - 8.18 Mbps

2. Upstream Test:  -provides background information.


 Upload Speed
 0 Mbps

   

0 Mbps 0.83 Mbps
Max Achievable Speed


Upload speed achieved during the test was - 0Mbps
 Additional Information:
 Upstream Rate IP profile on your line is - 0.83 Mbps


This test was not conclusive and further testing is required.This might be useful for your Broadband Service Provider to investigate the fault.

Its bizarre that it now seems to indicate my line should be able to support 21Mbps - I've never got that and would never expect to be able to get that.

I think this BT Wholesale tester is incrediby unreliable. Maybe its this peculiar fault that makes it unreliable in this case, but its results never seem to match what I expect or what I get from any other speed tester. Its either reporting more than double or less than half what the other speed test sites report. And when I take into account how my general browsing experience feels at the time i.e. how fast pages/images loads, etc the other sites are a far more accurate reflection of my experience than the BT Wholesale speed test.

[#7]

@DS I'm assuming it was lower than what you normally see

No not really - that's the problem. The equipment the engineer had and my router are reporting sync speeds of 9Mbps, which is close to what I've always had. I'd be fine with that. The problem is that in real world usage (and speed testers, ignoring BT Wholesale's one which is more of a hindrance than a help) I'm actually getting 0-4Mbps.

[#8]

Oh, okay.

Forgive me for asking, as it's probably already been answered, but are you in the test point still?

I would take the speed test result with a pinch of salt, but I wonder if Openreach are working at your cabinet, possibly upgrading the equipment. It is odd the tester has now recently jumped to 20Mbps, this could be due to you using Linux, but I doubt it. Do you live near it?

The BT DSL checker will give you the cabinet number you're connected to, then via google you should be able to find out where it is or https://www.telecom-tariffs.co.uk/codelook.htm

[#9]

Apologies for any confusion caused @ec5778

I can confirm you already have a fault open regarding this.

I'll get this updated for you shortly and we'll update you when we know more

[#10]

Just wanted to come back here and update everyone.

This is still ongoing and at the moment there doesn't seem to be any resolution in sight.

I've been offered a discount on Fibre due to the problems but besides from that I'm now being told that there isn't much more that Plusnet can do about this. Plusnet have offered to send further engineers but with the caveat that its likely to be a long drawn out process which in the end may not actually achieve anything.

As far as I can tell, the phone line is good. We've had 4 openreach engineers out and none of them have found any obvious problem. Right from the beginning I suspected capacity issues in my local area are behind the low throughput. Since then at every stage my initial suspicions have been reinforced although there's no real way for me to prove it. I've been told that if network congestion is indeed the problem here, nothing further can or will be done about it and the only option is to upgrade to Fibre.

So - the main reason I'm posting back here is to find out, has anyone else come across this before? Anyone else experienced this issue?

(I'm in North London area connected to the pcn-ir01 in the Colindate POP according to http://usertools.plus.net/@gateway/, if that's any help.)

[#11]

So - the main reason I'm posting back here is to find out, has anyone else come across this before? Anyone else experienced this issue?

Coincidentally yes I have

Mine all started with BT and their then ADSL service, but way back in 2008. I could have periods of low speed and a dropping connection. I had many engineers visits and a lot of work done on the line right back to the exchange.

Despite all the work done the connection was still very poor. I'd endured years of hassle and by now it was 2014 (thus 6 years of hassle). In the end BT suggested I move to fibre as I too was suffering from congestion issues. Now if you think about it, as others move to fibre you'd expect the burden on ADSL to diminish....? But I was running out of options.

I wasn't offered a discount and BT were confident my issues would cease after migrating to fibre (yes, of course they'd say that as I was giving them more money). So we migrated to FTTC and it was ....... worse than before!! In the end BT and I agreed to part company using the connection as the reason. This was penalty free, though even if they'd not agreed I was going to use the price rise 'get out of jail card' anyway.

I then went to Sky, but back on ADSL. I had a very reliable service from them, but alas their pricing became too high (imho) so jumped to Plusnet.

And as we wanted Plusnet TV, we had to go back to FTTC but the issue of the dropping connection came back, this wasn't too bad though it started in 2016. There were also fairly complex issues in getting the Plusnet TV working too, but I stuck with them. This year was when we had 6 or 7 (I forget now) Openreach engineers out and the last one seems to have fixed it.

So, you could stick with ADSL and work with Plusnet and you may not achieve anything, then again what ever is causing this might be found and fixed. You could try moving to FTTC and still have issues, though some do move over and their connections are then fine. You could be lucky (as I've just mentioned) and the issues go away. Let's say it doesn't fix it, then if you stick with Plusnet once your discounts run out you would more than likely be paying more (as FTTC is more expensive) or they may give you discounts for the while duration of your minimum term FTTC contract. I would have them put something in writing so that if you do decide to stay as is, then you can leave penalty free - but I'd do the same if you do decide to move to fibre too.

(also wait for other suggestions and/or opinions too :))

[#12]

Hi @ec5778 ,

Reviewing the whole topic, this does not look like a synch issue, or a REIN issue (error count would be very high).  One of the Plusnet staff reported that the VLAN was green, so on the face of it, this is not a back haul issue.

Are you on a fixed IP address or a dynamic address?

Which speed testers have you used?  What does the Think Broadband speed tester report?  Marked variations between speed testers has been raised as a concern in the past.

Exactly what is reporting the poor performance figures?  If you are looking at data transfer figures from file copying applications, can you confirm that (a) those metrics are in bits and not Bytes (it has happened before!!); (b) be sure that there is only no data transfer running (on this PC or other devices - e.g. phones, smart TV...) at the time the tests are being run and (c) can you be sure that there are no background (hidden) data transfer taking place?  If this were a Win10 platform, I would be suspecting that the OS is "quietly" running system update activities, but as it is linix I do not know where to look specifically, but I expect that similar concepts exist?

@Gandalf - did anything ever come of your investigation into BTw reporting that VLANS were "OK" … only to see the issue "disappear" after you pushed through a VLAN swap request?

Is a TPE likely to offer any potential remedy here?

Where is the BTw speed test "test head" - in the exchange or somewhere remote from the exchange (up the back haul)?  The better data throughput speeds reported by BTw speed test compared to other seems odd; if the BTw speed test head is in the exchange, then the difference might be accounted for by a VLAN issue, but if not, then all speed tests are using the same back haul … what might be the difference thereafter … "new" network routing?

Are there any continuing concerns about the variability of BTw speed tests results compared to others and the anecdotal observations that the browser used had some deleterious impact on the results reported by other speed testers?

If this user is on a dynamic IP, is powering off the router for an hour plus to force the allocation of a new dynamic IP address likely to be helpful?

Is there any data to suggest that this user is maxing out the US - we have seen issues with the 2704n under stress data transfer conditions!

[#13]

@DS Thanks for sharing your experience. Sounds awful! I hope mine doesn't drag on that long.

@Townman To answer your questions:

1. I'm on a dynamic IP
2. I have used every speed tester I could find - speedtest.net, thinkbroadband, which broadband speed test, AAISP, BT Wholesale. Initially I preferred Which Broadband for consistency but I find the ThinkBroadband one is the most detailed and hence recently I've preferred it.
3. See my ThinkBroadband speed test results
4. Everything is reporting poor performance - except BT Wholesale speed test and a couple of other speedtest servers e.g. there's one called Tele2 which bizarrely seems to be similar to BT Wholesale. But all the other speed tests above are reporting poor performance, as are good old fashion downloads (which are measured in Kilobytes per second, which I know needs to be multiplied by 8 to get Kbps), streaming video keeps buffering and the quality is degraded and the internet simply feels slow to use.
5. Past experience shows that other data transfers don't always affect speed test results that much, however to make absolutely sure I have run several tests every day and occasionally run tests connected to the router via ethernet cable and disabled wireless from the router's admin control panel to ensure nothing else could be using the connection at the same time.
6. Linux does have auto-update mechanisms but I have it set to prompt before downloading anything in the background. Due to the large number of speed tests I've run I'm reasonably confident that the problem cannot be attributed to background data.
   
   [I know some of these questions I'm answering were addressed to @Gandalf but here are my thoughts:]
7. Clearly the BT Wholesale speed tester is served from somewhere that doesn't involve going too far up the backhaul. AAISP actually state this explicitly on their speed tester and I notice that it tends to higher readings than other speed testers.
8. I haven't noticed any massive variation between browsers (I've tried Chrome, Firefox and Internet Explorer) but there is undoubtedly a big difference between the results being reported by BT Wholesale speed test and 'real world' usage to practically any other website.
9. I have had many different dynamic IP addresses because the connection drops a fair amount - at least half of the drops have due to someone changing the line profile/SNR targets (to no avail). I seem to get a new IP each time and mostly it doesn't seem to make a difference, though occasionally it seems to be a bit more snappy, until the next drop/resync.
10. Re upload saturation - been there, done that! I don't think that's the issue this time, especially as I did a few tests where I isolated all other devices by disabling WiFi on the router. Besides, I'm now using the Plusnet Hub One as they sent me a new router in the hope that might help. (It slightly increased the sync speeds but didn't do much to the real throughput that I'm attaining in real usage).

I went on holiday a few days ago somewhere rural where the maximum speed they could get was 3.5Mbps and browsing there felt so fast compared to supposedly 6+Mbps BT Wholesale reports that I'm getting at home.

[#14]

Thanks for getting back to us.

I'm sorry to see you've been experiencing throughput problems for so long.

I'd say that the drops in your connection are the cause for this: 

I'd personally hold onto that thought of fibre for now, I've reviewed the engineer notes and from what we can see apart from changing the pair between the exchange and cabinet no other work has been carried out on the line.

It's possible that the issue may just follow after upgrading to fibre depending on where the issue is.

Other engineers seem to have just stated they've not found a fault/the cause for the problem. Until an engineer advises something along the lines of the line is too long, I reckon we keep on investigating.

What I'd like to see next is that we request a lift and shift (or a tie pair modification - TPM) which involves an engineer moving you to a different port at the exchange. We should be able to arrange this without an engineer needing to visit you. 

Having said that I've discussed this with Josh and he said he'll call you later on today to discuss this further.

@Gandalf - did anything ever come of your investigation into BTw reporting that VLANS were "OK" … only to see the issue "disappear" after you pushed through a VLAN swap request?

This was around 18 months ago and since the new reporting tool I've not seen this issue again.

Is a TPE likely to offer any potential remedy here?

What's a TPE?

If this user is on a dynamic IP, is powering off the router for an hour plus to force the allocation of a new dynamic IP address likely to be helpful?

Not sure this is in any way related to an IP address, there definitely seems to be a problem on the Openreach infrastructure or maybe something affecting the infrastructure like REIN or SHINE.

[#15]

@Gandalf Thanks very much for your input on this, much appreciated.

I'd say at least half of those drops are when we had an engineer / when connection settings where changed on Plusnet/openreach side. But clearly it is dropping more than you'd expect.

But what this doesn't explain is why single-threaded connections are slow even when it hasn't dropped for a while nor does it explain why adding multiple connections increases throughput linearly until I reach the maximum throughput offered by the current sync speed (taking into account overheads) - or why BT Wholesale's speed tester consistently gives much higher readings.

If I download a ThinkBroadband test files it will download at speeds fluctuating between 70-90kB/s which approximately the same as the 0.7Mbps result shown by the green-coloured "Average TBB Download" figure in the test reports (see link to test result in previous post).

If, however I then download several of those files at the same time, each one continues at approximately the same speed, until it maxes out my connection when I'm downloading 10-11 of them, which offers a total throughput of 8-9Mbps. I've also tried with a download manager that makes 16 connections to the same file, and this also massively increases throughput. I then tried reducing the number of connections to 1, and guess what - back to 0.7Mbps!

Surely that can't be explained by REIN/SHINE or disconnection problems? (It also makes sense that different speed tests report different readings depending on where they are physically located i.e. what route the traffic needs to take to get there, and how many concurrent connections they use - notice that the Think Broadband HTTP x 6 test is giving me approximately 6 x - 0.7Mbps of the single-threaded TBB test).