Virgin Media Community

[leemcg]

I'm on M350 in North East London. Every day without fail, on ethernet, I get a lovely 350D, 36U at about 7.30am (have just run a speed test now). But from about 9-4pm my upload speeds fall to 2-4Mbs. 

Regular as clockwork, and the same for 2 other people with Virgin in my street.

As it seems the only way, I'm still including all of my details. Maybe all three of us are configured wrongly...

My Monitor

[url=https://www.thinkbroadband.com/broadband/monitoring/quality/share/0ec5e249737eff958566b9d715d1e688680053e8]My Broadband Ping[/url]

(sorry, I've tried a few different ways to copy in the link, and none seem to work as a nice image. Happy to edit if someone can let me know what the right way to link is)

And from my Hub (which is in modem mode) status:

Downstream

Downstream bonded channels

Channel Frequency (Hz) Power (dBmV) SNR (dB) Modulation Channel ID

1	202750000	2.7	37	256 qam	9
2	210750000	2.5	37	256 qam	10
3	218750000	2.5	37	256 qam	11
4	226750000	2	36	256 qam	12
5	234750000	1.9	37	256 qam	13
6	242750000	2.5	37	256 qam	14
7	250750000	2.5	36	256 qam	15
8	258750000	2.4	36	256 qam	16
9	266750000	2.5	36	256 qam	17
10	274750000	2.2	36	256 qam	18
11	282750000	2	36	256 qam	19
12	290750000	2	36	256 qam	20
13	298750000	2.2	36	256 qam	21
14	306750000	2	36	256 qam	22
15	314750000	2	36	256 qam	23
16	322750000	2	36	256 qam	24
17	330750000	2.2	36	256 qam	25
18	370750000	1.5	36	256 qam	26
19	378750000	1.4	36	256 qam	27
20	386750000	1.4	36	256 qam	28
21	394750000	1.5	36	256 qam	29
22	402750000	1.7	36	256 qam	30
23	410750000	1.7	36	256 qam	31
24	418750000	1.7	36	256 qam	32

Downstream bonded channels

Channel Locked Status RxMER (dB) Pre RS Errors Post RS Errors

1	Locked	37.6	7960	34897
2	Locked	37.3	6694	31940
3	Locked	37.3	7782	32311
4	Locked	36.6	8577	33386
5	Locked	37.3	8045	32610
6	Locked	37.3	7805	28418
7	Locked	36.6	7252	30326
8	Locked	36.6	6578	30110
9	Locked	36.6	6152	27855
10	Locked	36.6	9081	30386
11	Locked	36.6	9868	32578
12	Locked	36.6	9574	24204
13	Locked	36.6	10210	24101
14	Locked	36.6	7005	22890
15	Locked	36.3	6800	25439
16	Locked	36.6	9669	24788
17	Locked	36.6	7427	23715
18	Locked	36.6	7837	21209
19	Locked	36.6	11220	25363
20	Locked	36.3	7253	26263
21	Locked	36.3	9642	25044
22	Locked	36.6	10161	23784
23	Locked	36.3	9142	19598
24	Locked	36.3	8156	20665

Upstream

Upstream bonded channels

Channel Frequency (Hz) Power (dBmV) Symbol Rate (ksps) Modulation Channel ID

1	25799757	39	5120	16 qam	4
2	46199985	41	5120	64 qam	1
3	39399758	40	5120	64 qam	2
4	32600063	39.5	5120	64 qam	3

Upstream bonded channels

Channel Channel Type T1 Timeouts T2 Timeouts T3 Timeouts T4 Timeouts

1	ATDMA	0	0	2	0
2	ATDMA	0	0	1	0
3	ATDMA	0	0	1	0
4	ATDMA	0	0	11	0

Network Log

Time Priority Description

02/02/2021 07:55:39	Warning!	LAN login FAILED : Incorrect Username / Password / ConnectionType;CM-MAC=**:**:**:**:**:**;CMTS-MAC=**:**:**:**:**:**;CM-QOS=1.1;CM-VER=3.0;
01/02/2021 22:39:46	critical	No Ranging Response received - T3 time-out;CM-MAC=**:**:**:**:**:**;CMTS-MAC=**:**:**:**:**:**;CM-QOS=1.1;CM-VER=3.0;
01/02/2021 11:40:22	Warning!	RCS Partial Service;CM-MAC=**:**:**:**:**:**;CMTS-MAC=**:**:**:**:**:**;CM-QOS=1.1;CM-VER=3.0;
30/01/2021 11:15:48	Error	DHCP RENEW WARNING - Field invalid in response v4 option;CM-MAC=**:**:**:**:**:**;CMTS-MAC=**:**:**:**:**:**;CM-QOS=1.1;CM-VER=3.0;

I've tried to report this, but haven't managed to speak to a person yet, hence trying here.

[Andruser]

When did you last reboot the hub?  If the BQM refers to your current IP address, then the hub is rejecting test packets that create the BQM, and a reboot usually solves that, also you've got high post-RS error counts, which could be accumulated over time.  You've also got upstream noise issues because one of the channels is dropping modulation.

But all of that may be incidental - what you describe sounds like a widespread problem for many VM customers at the moment or "over-utilisation", which is where there's too much traffic due to home working and home schooling.  That usually affects upstream traffic, and hits the upload speed and latency (response times) which affects gaming and video calls or voice conferencing.  If this is the problem, then the only cure is to wait for lockdown to end, or to get a new ISP (which might take until lockdown has ended).

[leemcg]

Do Virgin acknowledge that? Because it seems to me that they try to pretend all is well. One of those other people in my street was only installed 2 weeks ago, so that can't be helping with the problem, if it's just utilisation.

[Andruser]

@leemcg Do Virgin acknowledge that?

Let me put it politely: It is readily and obviously apparent that VM's network is in some/many areas really struggling due to Covid.  The company's response is too often one of obfuscation, denial and delay, and is causing huge damage to its own reputation.  I suspect many customers would be pretty understanding if VM came out and admitted where and when it had problems due to the exceptional demands of lockdown, but openness is not a quality the company possesses.  In theory VM don't sell new connections in areas of over-utilisation, on the other hand there's a lot of things VM do or don't do in theory. 

But absent a BQM it is difficult to be certain.

[tehwolf]

---

@Andruser wrote:

@leemcg Do Virgin acknowledge that?

Let me put it politely: It is readily and obviously apparent that VM's network is in some/many areas really struggling due to Covid.  The company's response is too often one of obfuscation, denial and delay, and is causing huge damage to its own reputation.  I suspect many customers would be pretty understanding if VM came out and admitted where and when it had problems due to the exceptional demands of lockdown, but openness is not a quality the company possesses.  In theory VM don't sell new connections in areas of over-utilisation, on the other hand there's a lot of things VM do or don't do in theory. 

---

i suspect that many folks affected by overutilisation would also be willing (even happy!) with some sort of traffic shaping being implemented to improve performance at peak times!

[Philts11]

I agree, consistently bad/ unusable upload speeds.  I would sacrifice download speed for upload  - i'm not watching TV or gaming during the day but I am trying to participate in video calls 

[Andruser]

@tehwolf i suspect that many folks affected by overutilisation would also be willing (even happy!) with some sort of traffic shaping being implemented to improve performance at peak times!

That wouldn't help, because it isn't really about inequitable distribution of bandwidth, simply that the combination of a hard coded 10:1 down/up ratio on the DOCSIS standards plus VM's long standing approach to contention ratios (that is, the "average" use assumed when matching customers to available bandwidth) don't work when so many people are doing upload intensive things (MS Teams, Zoom, Houseparty, all the home schooling etc).  Back in the good old days of 2019, relatively few office workers worked from home, kids (when not skiving) did the decent thing and went to school, and in the evening there were plenty of alternative ways to spend our time compared to in front of a computer, so a fixed amount of bandwidth could be spread across more homes without causing a problem.

Individually, most houses aren't attempting to use more than their contracted upload bandwidth (some are, just not many).  But the cumulative sum of all that upload activity is probably (guessing) around three times the assumed cumulative use, and even with some headroom in the original contention calcs, there's a bottleneck at the CMTS.  Short of booting some customers off the system, there's nothing VM could do to sort this out in any short term timescale.  Worth noting that other ISPs may also have been caught out by Covid, but their problems will more typically be on the downstream because of the way VDSL works, and even then Openreach appear to be more careful about contention as they are subject to closer regulatory attention.   

Whisper it, because I have a reputation to uphold, but I do have some rare sympathy for VM in terms of the problem occurring - nobody realistically expected this to happen with any probability that made it worth investing for beforehand.  Where I have no sympathy for them is their appalling handling of these problems, the deafening silence about the issue, the fact that their avoidably poor customer service has made things worse for customers, and that they then crimp out another inflation busting price rise.  

[tehwolf]

@Andruser I'll take your word for it - I know very little about DOCSIS.

in my (simple) head though, if they capped upload at 15mbit for everyone on more than that (an amount that should support 3 - 4 video calls going on at the same time in any given property) they'd surely have more to go around..?

accept that it's probably not that simple, and as i said, the sum total of my knowledge on DOCSIS could be written on the back of a postage stamp with a sharpie, so I'm just makin' it up 🙂

[Andruser]

Capping the upload at that level wouldn't help, because the "average" use would still be too high.  Say you're on a 1 Gbps contract with what, 50 Mbps upload...most people aren't coming close to that other than infrequent file uploads.  System stress will be just be the same because across the fleet of 1 Gbps connections most households will be doing the same sort of 1-3 video-call sessions as customers with slower speed tiers, so they'd be contributing as much to the problem before or after any 15 Mbps cap were imposed.  Rationing bandwidth with a cap might work if VM capped all connections to say 6 Mbps upload, but that of course means each household has to choose one decent quality video-conference stream, or two low quality ones (bearing in mind users often have no control over the settings).  Who gets the videocall capacity in a household with two working parent and two children?  How will that land with customers?  Can you see VM daring to do something that radical?  And as the over-utilisation varies from non-existent to extreme, you'd have different caps for every segment of the network creating immense confusion.

[tehwolf]

all valid points.

If VM has oversubscribed as heavily as you intimate (or estimate) then yes, it would have to be a severe cap to resolve it in the short term - the question, though, is would it be any worse than the current scenario in heavily overutilised areas? i mean, a lot of the most commonly used VC/collaboration platforms will automatically adjust bitrate of video streams according to the bandwidth available so for most people, it wouldn't need any manual configuration knowledge.. it would just translate to blocky video 🙂 

I don't care about it being a nightmare to manage though - that's their issue not the end-users - and comes down to the ability of their teams to manage and track change in their environment - that's all process and governance (i'd hope that they were better at this than customer relations, but i guess it's a pretty bold assumption nonetheless). 

it would go down like a cup of cold sick with some customers, no doubt - but i imagine that if there was an improvement in service for core WFH/School From Home, most reasonable people would be accepting of it.