Virgin Media - Cambridge CB2 - Fault / Congestion?

For the last week or two I've had pretty rough internet at times. Massive pings and packet loss.. I thusly reported this to Virgin Media who sent out an engineer to correct the power levels but still getting problems.

I have a ping monitor on my connection:

30mbit TBB ping

And Virgin seem to say they don't see a problem.

It's not my own use causing this. Just wanted peoples opinions. Don't normally have any trouble here in Cambridge.

Your link isnt working need to remove the extra http:// and % from it
Looks like congestion post on the virgin forums to confirm and good luck for an actual fix

Edited by deleted (Sun 11-Dec-11 12:43:06)

Yes congestion for sure and pretty bad. If you see a bloke with horns and a long pointy tail on ice skates there's an outside chance VM may get around to fixing it. If not look to moving on if at all possible.

Since they released 100Mbps unlimited the number of areas with severe congestion has increased for some reason nobody in VM seems to be able to pinpoint. In the meantime as they can't figure out what the problem may be they've decided to deny having one.

Ahh 100mbit. I see. I guess that makes sense.

Also the pingtest url is

Pingtest

I couldn't add it to the original post. I've posted on the VM forum but fear my post will go ignored like when I had issues before.

They'll answer it but it takes about a week these days - subscribe to the post.

For me it was - yes your CMTS is overutilised and we've raised fault nnnn (note there appears to be no routine monitoring - they wait for complaints). You'll then have to ask when a fix is scheduled because you can't track these faults in any way yourself - in my case it was three months away. Now some faults remain in place with ever delayed fix dates. In my case there was an unannounced outage of several hours and I was on a new CMTS which was marginally less overloaded than the first. On enquiry that was it - the fix was completed early and the new (still overloaded) connection didn't exceed their limits - over 80% utilisation for 10% of the time - any network running on at over 70% on the tiny local pipes VM use (200down 18up) at the speeds they sell is a horrible experience.

Edited by kwikbreaks (Mon 12-Dec-11 06:24:58)

Looks like classic congestion, though at a guess it is only one or two 50 or 100 Meg users causing it.

I called faults up in Mumbai? And they said they see high utilisation that they know about and that there's no ETA on it being fixed or anything.


But "don't worry about it sir" .. "it'll get fixed".


Just wish I could get someone from this country to talk to. :/

In reply to a post by MrSaffron:

Looks like classic congestion, though at a guess it is only one or two 50 or 100 Meg users causing it.

A single 100Mb user can use 55% of an upstream port's capacity. Sadly this doesn't work with the amounts of modems Virgin can sometimes have on upstream ports. They are hoping that upstream bonding will make a big difference, and it should. 2 x 10Mb users on a 36Mb channel works better than a single 10Mb user on an 18Mb channel for the same reasons as the datastream 'skinny pipe' issues way back when.

Probably the same answer from UK staff, ie. they do attempt to improve capacity, but it is a constant battle with the growth of traffic.

If the issue is 100 Meg users using a dispproportionate amount of traffic, then the curbs on that product will appear.

In reply to a post by MrSaffron:

If the issue is 100 Meg users using a dispproportionate amount of traffic, then the curbs on that product will appear.

As VM are driven by marketing and BT Infinity will soon be 80Mbps unlimited imo it's a whole lot more likely that there will simply be more congested areas unless the forthcoming changes to traffic management are pretty draconian.

When I used to work on the phones for Telewest, we had a tool to check the utilisation of all possible channels that a customer could be connected to. If they phoned up complaining of poor speeds (and they were nice ) we could switch them over to another channel that wasn't as bad. So I wonder if Virgin have these tools to do this...

This could also be done by the customer by logging into the modem config pages and adjusting the channel in there - I haven't checked on my Virgin modem yet to see if this functionality exists at the customer end still...

In reply to a post by MrSaffron:

Looks like classic congestion, though at a guess it is only one or two 50 or 100 Meg users causing it.

indeed, the system is very fragile.

100mbit has been rolled out on a infrastructure which typically has 200mbit downstream shared capacity and 18mbit shared upstream capacity. 100mbit has no downstream STM so only is managed by protocol shaping (easily evaded) I also believe the upload has no STM either.

So a single 100mbit user can use 50% of the downstream capacity for their area which will have 100s of users sharing on it, and about 60% of the upstream capacity, just one single user can do that. Some areas if lucky get a extra downstream channel which bumps the shared capacity to 250mbit which % wise isnt a big deal.

The VM PR leakers have been claiming 400mbit capacity is coming to fix these issues, however that 400mbit will be for a new 200mbit product so VM havent learned their lesson.

Its unreal of the attitude VM have to congestion levels.

Some news has leaked they plan to double the speed of all tiers except the bottom tier which will actually be tripled.

so 10 to 30mbit
30 to 60mbit
50 to 100mbit
10mbit users with a old modem will only goto 20mbit.

This will probably happen with no extra capacity to cater for it.

Remember the bulldog days andrew when they sold 2mbit adsl on 2mbit ports? and how well that worked? This reminds me of that as they selling speeds on ports that are barely bigger.

Sadly the new regulation only requires 10% of VM customers able to reach line speed for it to be legal so VM can happily oversub many of their customers and make marketing claims otherwise.

Thankfully FTTC is now coming to my exchange in 2012, I never thought I see the day I be so eager to join BT. My particular VM port ha sbeen confirmed as "very highly" utilised, the original fault ticket vanished and I had to get it reraised, the likelyhood is there will be a drip feed of extra capacity which will probably make the congestion a bit less severe but not resolve it.

Some people have claimed VM recycle tickets every 2 months so customers cant complain to the regulator.

In reply to a post by MrSaffron:

Probably the same answer from UK staff, ie. they do attempt to improve capacity, but it is a constant battle with the growth of traffic.

If the issue is 100 Meg users using a dispproportionate amount of traffic, then the curbs on that product will appear.

The issue really is the following in my view.

1 - there is no proactive monitoring and upgrades done, instead upgrades are done in response to customer complaints. At which point its already too late since it can take VM half a year or even more to do the work at which point sales still happen and the situation gets worse.
2 - As my above post selling too much on too small capacity.
3 - the rigid sticking to 'unlimited' usage on all tiers.
4 - when upgrades do get carried out they typically drip fed upgrades. I think in an absolute best case scenario the user would get a 100% increase in capacity when a reseg is done to a new empty port splitting the users down the middle, that however isnt a typical upgrade, its much more likely to be shifting some users off to another port already full of existing users. A game of musical chairs.

Is it possible for VM to get on top of growing usage? in my view yes?

Back in march 2011 VM enabled extra capacity in my area in preperation for the speed uplifts, they brought the extra capacity online 3 months prior to the speeds been uplifted. For those 3 months I had a very good service which only started to decline the last 2 weeks of it. The day they flicked the switch to triple upload speeds however it noticebly declined and was another case of VM not having a clue on capacity management. That new capacity should have been considered congestion relief for 'existing' services and then there should have been a 'further' significant upgrade to support the higher speeds.

The size of the upgrade that fixed my service for 3 months I believe was a 4 fold increase in capacity in upload capacity. If there was also a node split then it was probably a 8 fold increase along with a doubling of downstream capacity. Even then it started seeing light congestion towards the end of the 3 month period, this gives an idea how severe VM let things congest.

Any responsible isp network admin would upgrade capacity before it even comes close to congestion, it would be done ahead of time. VM appear to wait until complaints get high enough and then also require a silly high utilisation threshold to be reached as well. I have seen customers reporting speeds below 1mbit/sec on 50mbit products been told their area is within acceptable limits so no relief work will be done.

FTTC is now coming to my exchange in 2012

Same here!

I'm not having problems with Virgin at the moment but I would have nowhere else to go if the service deteriorated. A bit of competition will give them the incentive to keep their infrastructure up to date as well as relieving any current congestion.

I am willing to be the network admins are asking for upgrades, but as there can be cost implications you may find that only as people upgrade to better packages, i.e. more revenue that they will carry out the upgrades.

Another issue is whether the Ofcom speed results are truly representative of the actual customer experience. If the level of complaints increases then it could backfire in a big way, as people will look at the speed claims and ignore them.

well any financial side of it is down to VM, they choose to sell the products for the prices they sell at.

If look at their history, every single year since I signed up they have either given out a free speed upgrade or dropped prices or both.

last year it was the free speeduplift and 20 to 30mbit.

this upcoming year it will be the speed doubling and 200mbit launch.

if we look at the revenue side.

typical port 300+ users.
revenue per user probaby average £20 or so a month.

a single port alone can easily be generating 6k of revenue a month.

its very indicative that VMs cash cow is now broadband its their most profitable service.

In reply to a post by Chrysalis:

typical port 300+ users.

Something that's never been clear to me - can a CMTS port support more than a single set of 4or5 down channels and the 1 up - I'm guessing they must or surely it could never work. Assuming a port is not just the one set of channels do you have any idea what the typical user count would be on a 200/18 pipe?

from what I understand it will have a single set of downstream ports which is 3-5 on docsis3 and multiple sets of upstream ports, in uplifted areas usually 2 or 3 upstreams.

So if eg. there is 300 users and 2 upstream ports.

Then all 300 will share the downstream ports, and then they will be balanced across the 2 upstream ports approx 150 each if a 50/50 load split.

Now some areas probably havent got too many customers due to low takeup so may have much less customers. So when I said typical 300+ I meant VM will let it get that high if there is enough sales.

A area near to me where a friend lives has only 40 customers on it. That area still has 5 downstreams allocated to it and 3 upstreams, but because the sales have been poor they have a lower effective contention ratio.

VMs network is extremely unbalanced. Its really luck of the draw, its entirely possible for one segment to have 10% utilisation and the neighbouring one to have 70% and then for VM to not balance them out as the latter hasnt reached fault thresholds.

I also dont know if a single port can support more, I presume yes as the plan is to have 8 downstream ports for the 200mbit service. There is plans for upstream channel bonding as well, however I dont know if they just going to bond existing channels which means no extra capacity or add new upstreams so instead of eg. 2 single upstreams it becomes 2 bonded upstreams 4 channels total. In fact it could probably support many more channels as long as VM were able/willing to allocate them and modems were issued to support the extra channels.

Edited by Chrysalis (Wed 14-Dec-11 19:02:47)

Thanks. I somehow don't find any of that reassuring....

I had Virgin in the CB2 postcode. Beyond awful. Friends of mine a block or two away suffer as well. Recently I moved to the West Cambridge site of the University and had no choice but Virgin. A friend and colleague who is my new neigbour assured me his 10Mb broadband was flawless here over the better part of the past year. I have signed up here and have found the connection to be.. flawless. Similarly, other reports from someone who has Virgin in the CB3/4 area in an older residential area has no coplaints at all. For what it is worth, I tried O2 in CB2 as well and found what I took to be throttling to be severe. Good speeds initially, but i would plump for Bethere if I moved back to the area again.

I suspect it is simply a matter of the exchange perhaps. CB2 - avoid. CB3 - amazing.

There are no exchanges on cable - the smallest network segment is a node and that covers way way less than a three digit postcode area - typically just a few streets. It's the luck of the draw whether or not you get congestion on your node and VM don't seem to care whether you do or not.

The TBB monitor shows ever increasing average ping and jitter on my new port. As I downgraded from 50 t0 10Mbps when it first became heavily congested I do still get 9+Mbps down 24x7 but do see occasional upstream congestion. No idea what it would be like if I were still on 50Mbps but I don't think it would be brilliant judging by the TBB monitor.

The good news from my POV is that my scheduled Infinity avaiability has been brought forward from Dec 2012 to June 2012. I can soon be free of this awful company and their abysmal network management.

Eugh, I'm in the CB1 area, getting near the middle of town, and I'm seeing some atrocious speeds at times. My housemate goes on iPlayer, and suddenly the entire connection turns into a dead dog.

Most days I can't even get iPlayer to stream properly. I've taken to downloading them (often slowly) first so I can watch them. This is meant to be 30Mbps! I'm not sure whether to blame it on the SuperHub though, since at times I can't even get DNS lookups to go through if too many devices are connected throughout the house.

Anybody had any luck in complaining? What are the initial checks Virgin ask you to do, so I can get them out of the way?

I'm in CB22 - I had a horrendous connection for the first 6 months of 2011 which got fixed after the CEO got involved and then went pear shaped again from Oct onwards. They've just fixed it again and is now the best connection I've ever had. The problem in my case wasnt classic utilitsation but was the number of modems connected to an upstream section of the network. Its designed for a certain number and in Cambridge many more were connected and this results in packet loss even if the speed you're getting isnt too bad. The stats that customer service look at dont look at this - they only look at total utilisation in which case the ends users connection may be no good for streaming, but everything at Virgins end says everything is ok.

I'll be very happy if my connection stays the way it is now but I wont hold my breath.

I disconnected all my wireless clients, then hooked up a laptop to the superhub with an ethernet cable, and was getting 30Mbps without a problem, so it's not the connection, unless it's based on the time of day. Switching the wireless speed form 300Mbps down to 145Mbps improved the speeds over the wireless to about 15Mbps per client, which is more acceptable.

I'll retest tomorrow evening to confirm, but it's looking much better now.

Mine remains unfixed here in CB24. Re-emailed the CEO again today as I've seen no result since doing it back in December. If anything it's actually got worse.

As it is I've now cancelled my direct debit. Over the past year I've had a total of a tenner refunded to me for my trouble. Bit of a joke. I figure if they can only provide me with 5% of a service, I'll pay them 5% of the bill until it's fixed.

While VM are very inefficient in fixing congestion issues you'll quickly find that they are very efficient in adding fees for bounced DD requests and also in passing debts over to debt collection agencies. I'd recommend that you go through the proper cancellation process rather than just cancelling the DD. Just call 150 and select thinking of leaving if you are unsure how to cancel.

I've complained on the forums and by CEO office and still nothing has changed.

LAG

My question is why do VM not upgrade the nodes ahead of schedule? They own the kit in the cab so what they effectively need to do is just add more fibre to link back their central office? It can't be that expensive to push more fibre through the ducts?

Are saying is that there is congestion on selected nodes but the VM core network has lots of spare capacity?

Cabinets with fibre will have multiple fibre in them already, so usually not even that that is an issue.

Think more about the costs of peering, i.e. links from VM to the internet, and the high ultisation of upstream, which can affect the cost of peering, and willingness for some to peer direct.

down to cost.

They seem to wait for customers to complain and then if the threshold is exceed as well then relief work gets done.

This relief work isnt guarantueed to be an upgrade it could just be a rebalancing act to slightly improve things and get usage down to just under the threshold.

Also if a ticket gets opened by someone at VM for congestion (high utilisation) and the utilisation temporarily improves its possible for the ticket to simply be closed without any resolution carried out.

Edited by Chrysalis (Fri 20-Jan-12 19:00:36)

In reply to a post by rwbarrett:

I'm in CB22 - I had a horrendous connection for the first 6 months of 2011 which got fixed after the CEO got involved and then went pear shaped again from Oct onwards. They've just fixed it again and is now the best connection I've ever had. The problem in my case wasnt classic utilitsation but was the number of modems connected to an upstream section of the network. Its designed for a certain number and in Cambridge many more were connected and this results in packet loss even if the speed you're getting isnt too bad. The stats that customer service look at dont look at this - they only look at total utilisation in which case the ends users connection may be no good for streaming, but everything at Virgins end says everything is ok.

I'll be very happy if my connection stays the way it is now but I wont hold my breath.

I'm in CB22 and my connection has gone downhill over the last month, I've had 3 engineers visit with no success and the VM forum team keeping telling me it's due to high-utilisation. The engineer tells me the utilisation issue was fixed last year, which tallies with your post. Are you still getting a good connection?

My connection is still good but I got moved onto a different UBR last September which is when my problem came back and this one is now ok. If you go to http://tools.virginmedia.com/ take a look at the second part of your Client Host Name and if its cmbg15-2 then youre on the overloaded UBR which as far as Im aware hasnt been fixed.

I was under the impression that VM's core network was still underutilized with lots of spare capacity.

As with all CATV networks the upstream is the limiting factor meaning we can only assume that with VM marketing greater speeds that it become more and more saturated.

Even on DOCSIS 3.0 there is limited allocation for the upstream channel presumably meaning dedicating more fibre or network re-segmentation work is the only way to help to reduce congestion.

There does seem to be a worrying trend of VM only acting once enough people complain and/or CMTS/UBR utilization reaches 95%.

In reply to a post by Kiggs:

I was under the impression that VM's core network was still underutilized with lots of spare capacity.

Core network isn't the same as access network. Similar things occur with BT's Virtual Paths on WBC ADSL and with LLU ADSL (see Sky's current problems) - due to the delays in increasing backhaul.

There does seem to be a worrying trend of VM only acting once enough people complain and/or CMTS/UBR utilization reaches 95%.

This sounds worryingly like money related.

Ah you've re-started a golden oldie thread from Dec 2011. Well my own are was stuffed then and still is. Fortunately relief is on the way for me at least - http://www.idontbelieveit.me.uk/bt-infinity-coming-a...

With any luck you'll be able to make the same move in the not too distant future. VM have blown it by pushing for headline speeds which their network cannot sustain at the local level if they are used and by selling unlimited they ensure that those speeds will be used pretty much constantly as their shaping is being bypassed by the use of VPNs. To be fair there really isn't much point in a limited 100Mbps service though because anything past a few Mbps won't make any discernible difference to browsing or gaming.

if you are on cmbg15-2-0, escape if you can. VM won't fix it. Unless some miracle occurs and they do ACTUAL UPGRADES for the speed double we are scheduled for this month/next month.

If I want to escape VM and have internet access I have one choice: Move home. ADSL in my pokey town is appaling. It can maybe do 3 or 4 mb at good times in terms of speed but has latency orders of magnitude above VM (I'm less than a km from the exchange!)

BT have no plans to hit the area with Infinity. My neighboring town gets it soon but I can't afford to move home.

I complain and complain to VM, only to be told different stories from different people. CEO's office, after 2 months of nagging, said it was too many modems on the line. No fault reference provided. I tried to go back through the forum to get one only to get different stories. Now they say there's an occasional utilisation spike but nowhere near enough to raise a fault.

This is for a connection that is unusable most evenings and weekends since November 2011, not to mention for the months of January to July 2011.

I cancelled my DD and told 'em I'll pay my bill as and when my bank account isn't over-utilised. As amusing as that is I still have to settle it once in a while.

I am in Trumpington in Cambridge and i've been moved to CMTS 17-2 now however things are still [censored]. Anyone know whats going on in the area?

In reply to a post by Mhisani:

I am in Trumpington in Cambridge and i've been moved to CMTS 17-2 now however things are still [censored]. Anyone know whats going on in the area?

Im in the same area. When i spoke to vm last week about the ongoing issues i was told the high utilization had a fix date of May 2nd.
Dont hold your breath though as the previous date was Feb 1st!

In reply to a post by Mhisani:

I am in Trumpington in Cambridge and i've been moved to CMTS 17-2 now however things are still [censored]. Anyone know whats going on in the area?

There's work being done for sure. According to a VM tech on the VM forums resegmentation work is being carried out. They are also adding capacity it seems as last week I went from 4 downstream channels to 6.

No speed double yet, which is due March/April for our area.