High Utilization Fix Dates

Anyone know why they keep delaying the fix dates all the time for high utilization faults... It seems to move on each time to at least a month later?

Do they have like ONE engineer working on this?

Are the dates moving, or they are upgrading and users are growing their data use to consume available bandwidth even after upgrades.

This has happened with BT Wholesale upgrades in the past, when incremental increases in capacity are not enough.

In reply to a post by Coogz:

Do they have like ONE engineer working on this?

Judging by the number of complaints I'd say it's more like NONE.

My own area whent to the dogs and I was given a date 3 months away so I downgraded from 50 to 10 as that is all it was delivering a lot of the time. Some sort of work was done and my IP changed but the TBB monitor still looked abysmal. As I'm now only on 10Mbps I can usually get that but I imagine gamers would be unhappy with the high jitter.

IMO VM are running scared of Infinity and releasing products which their infrastructure simply can't support - 100Mbps on 200Mbps pipes supporting a couple of hundred modems. There will be some upgrades for the speed doubling but imo that capacity is needed on what they are selling now.

VM marketing are clearly hoping to BS people into staying with them or joining on the back of headline speeds without the bandwidth being there to back up those speeds when they are used in anger.

The jitter is visible in the Ofcom testing too, particularly on the upstream

You'll always get jitter on cable because of the way the upstream works using requested timeslices. How bad it is depends on how many torrent freaks they've sold to in the local area who are smart enough to evade the shaping.

Here's my TBB chat from yesterday - http://www.thinkbroadband.com/ping/share/8b155e7a6ff...

Plenty of people have posted far worse (but only those on cable SFAIK)

Edited by kwikbreaks (Wed 08-Feb-12 10:10:29)

In reply to a post by MrSaffron:

Are the dates moving, or they are upgrading and users are growing their data use to consume available bandwidth even after upgrades.

This has happened with BT Wholesale upgrades in the past, when incremental increases in capacity are not enough.

Probably both.

VM wait until utilisation tops 90% before even starting to plan an upgrade. By then performance is severely impacted already and then the time from planning to relief been provided can be several months. During that time they are of course still signing up new customers and the upgrade may well only be a small incremental amount. It does seem from where I sit VM want and try to keep utilisation at around 90-100% and only act when thngs get extremely bad,

On my current high utilisation fault I am now on my 4th ticket, they just keep getting closed with no real reason.

To give you an idea how bad it is to wait until 90. 10% of a upstream pipe is only 1.8mbit/sec. 100mbit users can consume 60% all by themselves, 50mbit 30% and 30mbit about 17%. Jitter is likely noticeable on anything above about 40% utilisation.

Comcast a US cable provider consider 70% as high utilisation and thats on larger shared pipes as well.

Edited by Chrysalis (Wed 08-Feb-12 12:41:59)

In reply to a post by Chrysalis:

Jitter is likely noticeable on anything above about 40% utilisation.

Mid-60s.

In reply to a post by Ignitionnet:

In reply to a post by Chrysalis:

Jitter is likely noticeable on anything above about 40% utilisation.

Mid-60s.

it was noticeable here when I was told I had mid 40s utilisation. Also if utilisation is at say 40% and a 100mbit user does a upload speedtest or something, that would cause a burst of jitter would it not as it would burst it to 100% utillisation or close to it.

Edited by Chrysalis (Wed 08-Feb-12 21:46:21)

In reply to a post by Ignitionnet:

In reply to a post by Chrysalis:

Jitter is likely noticeable on anything above about 40% utilisation.

Mid-60s.

So in most areas just about any time you want to use it unless you're an insomniac.

well I am curious how he obtained this mid 60s figure. Has there been specific live testing eg. and how recent was it.

If it was from when he used to work for VM years back, back then conditions were different, mainly that upload speeds were much lower.

Even if we accept mid 60s as a jitter trigger point thats 30% or so below what VM consider high utilisation and probably below the average utilisation across their network.

I get jitter 24/7 on VM even at dusk, so that would suggest my utilisation at 4am is at least 65%, isps that show uncongested traffic levels usually show that traffic at the time of day is many multiples below peak demand which means peak demand on my port is probably restricted by congestion (and certianly by shaping).

We also now know that VM have used their traffic management to hold back upgrades.

They use a trigger point on utilisation for upgrades, traffic management reduces utilisation (to a degree) which makes it harder for this trigger point to be reached, so in turn traffic management has stopped upgrades been carried out. Unlike most other ISP's VM still have 2 seperate forms of traffic management running and cannot prevent congestion. They used to have 3 until they scrapped their FUP letters. In my opinion STM violates the ASA ruling which comes into affect april this year, so its possible that will be canned this year leaving just the protocol shaping in place. Although the STM is likely not having too much impact anyway as its very generous on uploading and doesnt even affect the top tier at all.

It would be nice for tbb to run a story of some sort on this as this site usually is quite open.

Story on what?

The ASA 10% ruling, willing to bet that the testers used for the assessment of the 10% are not the 10% heaviest users triggering the throttling back. In fact the signup pages for a tester indicate they want average usage figures, which may skew things.

We were one of the few to cover the other metrics from the Ofcom testing, and have visibility of some of this ourselves via other testing that we run.

story on VM giving fix dates that slip and slip, closing tickets without fix, that sort of thing.

or is it in your view a non event?

In regards to the STM, what I meant by that is that the ASA ruled that any throttling on unlimited products can be moderate only, STM cuts speed by 75% which makes it far more than moderate. Hence my view it conflicts,

Edited by Chrysalis (Fri 10-Feb-12 10:54:33)

In reply to a post by Chrysalis:

In reply to a post by Ignitionnet:

In reply to a post by Chrysalis:

Jitter is likely noticeable on anything above about 40% utilisation.

Mid-60s.

it was noticeable here when I was told I had mid 40s utilisation. Also if utilisation is at say 40% and a 100mbit user does a upload speedtest or something, that would cause a burst of jitter would it not as it would burst it to 100% utillisation or close to it.

My idea of noticeable and yours are probably different. My idea of noticeable is a minor affect to service rather than obsessively watching TBB graphs and looking for the slightest delay on an SSH session.

In reply to a post by kwikbreaks:

In reply to a post by Ignitionnet:

In reply to a post by Chrysalis:

Jitter is likely noticeable on anything above about 40% utilisation.

Mid-60s.

So in most areas just about any time you want to use it unless you're an insomniac.

Some, not most.

call it what you like. an effect is an effect.

think of it this way, if a tbb graph is showing a fluctuation the chances are gamers will notice it. you can call it obsessive as much as you like.

that facts are this tho.

if I dont notice a problem on my connection I will at most check my graphs and do speedtests maybe once a week. If I do notice a problem on my connection then I will start to do checks which will include checking my graph and speedtests.

the sad state of affairs is I notice problems every day on my connection. Even with these daily problems I dont do speedtests and check my graph daily. I cant help it if simply using my connection shows me problems.

you may have a low expectation of your connection but dont expect everyone else to.

so you cant clarify how this mid 60s % figure was determined?

also its worth pointing out I never said major or minor I said any problem. certianly my current issues compared to a year ago are minor, however they are excessive when compared to better VM areas and certianly to infinity. Although granted I have no personal experience of infinity yet and that could turn out to be a pile of turd in this area when it comes to town.

Edited by Chrysalis (Sun 12-Feb-12 06:47:42)

most web pages today taking 3+ seconds to respond. Although not as bad as 12 months ago the way its declining my area will be like that within a month or so.

In reply to a post by Chrysalis:

you may have a low expectation of your connection but dont expect everyone else to.

Perhaps it comes from years of experience of cable?

LOL they saying its very high again.

http://community.virginmedia.com/t5/Up-to-30Mb-broad...

so they went from extremely high to way below fault threshold to very high.

5th ticket now.

mrsaffron read that thread I posted, and you will see what I mean.

I've amused myself with a response to that one on the VM board.

They really are becoming a total farce - I can see them going completely down the tubes in a couple of years time.

The mid-60s % figure was derived from laboratory tests by hardware vendors based on controlled tests of minislot utilisation on TDMA cable networks with hundreds of test modems.

Along with this are my own observations as someone who, unlike yourself, has actually worked with this stuff in the real world over multiple operators and network deployments.

I'm sure your observations of your Think Broadband Quality Monitor trump this and you'll be happy to try and correct me with second hand anecdotal evidence based on out of date or flawed information.

I couldn't care less what your connection says, when will you get through your head that every thread on Cableforum, Think Broadband and wherever else that touches on this topic isn't about your connection, your experience is exceptional and does not reflect the average experience.

Yes, jitter will increase during peak time on contended TDMA networks, shock of shocks, if you've a problem with this return to ADSL where you've a synchronous and nailed up link with the ISP''s transport network or purchase yourself a leased line.

Your constant whining about your own connection along with your total cluelessness on anything remotely technical outside of your extremely narrow experience in managing web servers is getting boring. If you don't like your VM service tell them to shove their service up their hindmost, switch back to a more stable jitter ADSL given that's what you seen to obsess over, return to giving yourself callouses over a Think Broadband quality meter and stop moaning. Ideally try and get something resembling a life while you're at it, ideally it might distract you from Think Broadband quality meters for a while.

Either way stop trying to play a big boys' game pretending you have a clue about networks of any description. You very clearly, across tons of posts across different forums, really don't. Just another nerd wannabe making money out of other peoples' ignorance and bigging up his own ego trying to make out he knows what he's talking about on forums.

Edited by deleted (Mon 13-Feb-12 22:21:40)

In reply to a post by kwikbreaks:

In reply to a post by Chrysalis:

you may have a low expectation of your connection but dont expect everyone else to.

Perhaps it comes from years of experience of cable?

Vendor tests and personal observation.

Do bear in mind that this of course doesn't come close to Chrysalis' observations of his Think Broadband Quality meter along with out of date reports of historic load from technical support on the occasions he gets figures. Clearly far more reliable.

In reply to a post by Chrysalis:

most web pages today taking 3+ seconds to respond. Although not as bad as 12 months ago the way its declining my area will be like that within a month or so.

So what? What does that have to do with the subject in question? As per any excuse to whinge about your own connection. No-one's forcing you to use Virgin, given they evidently aren't fixing it in a hurry either change or stop hijacking every thread you post on in whatever forum that's remotely related to whine about your own connection.

In reply to a post by Ignitionnet:

Vendor tests and personal observation.

It seems that your personal observation coincides with my own observation of my VM connection - I really don't expect much from it at all these days and will certainly be moving on as soon as I can risk a phone line based connection again.

In reply to a post by kwikbreaks:

In reply to a post by Ignitionnet:

Vendor tests and personal observation.

It seems that your personal observation coincides with my own observation of my VM connection - I really don't expect much from it at all these days and will certainly be moving on as soon as I can risk a phone line based connection again.

Or wait for a fibre connection instead.

Many people are getting fed up with VM but they, VM, still trounce most BT telephone line broadband products.

BT are serious about FTTH. I was speaking to a cheerful Openreach network engineer the other day who expects to lose his job within 5 years along with many Openreach employees if FTTH takes off. Most of their work is based on dealing with faults to copper lines which will be a thing of the past in a FTTH world.

My area is planned for FTTC and I will switch to it when it arrives. My phone line had a fault which OpenReach failed to fix after 5 attempts which is why I switched to cable. Going back to ADSL would virtually certainly see me on the same line with he same old fault. With FTTC I only have the last line segment which with any luck will be fault free or diagnosed and fixed on install.

With FTTC there is still the same length of plain old twisted copper pair but it only gets used for voice. FTTH will involve high install costs so I reckon your friend from OpenReach will still be in work for quite a while.

Edited by kwikbreaks (Tue 14-Feb-12 22:21:01)

In reply to a post by Ignitionnet:

In reply to a post by kwikbreaks:

In reply to a post by Chrysalis:

you may have a low expectation of your connection but dont expect everyone else to.

Perhaps it comes from years of experience of cable?

Vendor tests and personal observation.

Do bear in mind that this of course doesn't come close to Chrysalis' observations of his Think Broadband Quality meter along with out of date reports of historic load from technical support on the occasions he gets figures. Clearly far more reliable.

thanks for clarifying

Ignition, you need to stop the personal comments.

You used to work for ntl but you dont anymore. So in terms of what is the average experience at current how are you backing that up?

For a start its pretty evident VM themselves cannot accurently determine user experience, as they think anything below 90% provides a good user experience. Its logical to assume you still have colleagues within VM who you worked with you feeding you information hence you able to leak stuff, however its also logical to assume that they will have an optimisitic point of view on what is a good and bad utilisation.

You have also contradicted yourself, in the past you have moaned about VM using 80% as a fault threshold, yet its now 90% and you suddenly now have no point of view or rather you dont want it aired in public. Even more so you have resorted to personally attacking me claiming I am some unusual event and its only a minor amount of VM customers with performance problems. Ofcom seem to disagree with you and they have samknows data, they are due to release some more this year and you may found it shocking or you may not, but those figures will show a large chunk of VM customers with congestion.

Your reply to me states the following.

1 - you think oversubscribing is acceptable sales practice.
2 - you take offense at me reporting problems with VM as you have not reacted like this to other people who have reported as much as me.
3 - you have completely ignored the farcical support I have posted, again whats your agenda.

So here is a few facts for you.

1 - its not just about jitter. the tbb graphs are a way of monitoring latency and jitter, when these increase it indicates congestion which can cause other problems. You know this but you are playing dumb. Would you be happy if you cant stream content, you download at abysmal speeds, web pages take several seconds to load, latency senstive applications break or at least perform poorly, games are unplayable. The list goes on.
2 - there is congestion and there is excessive congestion. Your reply indicates you think there is no such thing as excessive and people should just up with it.
3 - This is widespread on VMs network, its not just me alone and 99% of VM customers dont have or dont care about this. This is point I have to make because your posting pattern on forums lately you are starting to look like a PR person for the company, pretty much all the info you happen to leak is things like speed upgrades and capacity upgrades, never the bad stuff eh? dont think its not obvious because it is. You now trying to play down what I have posted as some minor jitter that I am obsessing over, come on you know its not just a bit of minor jitter. Maybe next month you will leak some information that VM are upgrading another package and doing some 'massive' capacity upgrades that will keep things running smooth. Then you will continue to ignore negative posts or shoot them down.

I dont make any money out of problems on VM or any broadband isp's so this has got to the point you trying to completely slander and discredit me so people ignore what I say.

Edited by Chrysalis (Wed 15-Feb-12 03:44:28)

If you would be good enough to inform me where I described VM's capacity planning methodology as being adequate or fit for purpose it would be appreciated. Taking what I said and stretching it to an extreme doesn't make it what I actually meant.

I do not agree with VM's fault threshold however this is not the level at which capacity upgrades are undertaken it's the level when a fault ticket is raised for a priority fix.

I couldn't care less about the support you've received which is why I didn't comment on it, again it's not about you and your ongoing issues over a number of years are irrelevant to me.

I regard a level of oversubscription as being absolutely acceptable sales practice. All broadband networks are oversubscribed to some extent. I do not, of course, condone excessive oversubscription.

I take no offence at your reporting problems, I am amused at your constant complaints over a period of years about cable yet you still are a subscriber. I can only assume you are masochistic in some way, and enjoy having a cause for complaint. I am puzzled how you seem to think you can speak authoritatively on cable though given your information is a BQM and occasional loading information from support.

Cable networks are in 99% of cases a TDMA medium. Time Division Multiple Access. When TDMA networks load to any kind of level they show jitter. It doesn't mean they're excessively loaded. To entirely avoid jitter needs each and every modem to have transmit opportunities immediately upon request which isn't feasible. A level of jitter doesn't indicate an oversubscribed network, it's a limitation of the technology. A jitter-free cable network is the exception rather than the rule and requires insanely low loading.

Given your apparent expertise I'd have thought you'd have noted that downstream jitter on cable is low, while upstream jitter is higher. Good reason for that, the upstream is TDMA and works on a request - grant - transmit methodology, the downstream has the benefit of scheduling and standard transmit schemes used in synchronous connections.

I made no comment on levels of acceptable congestion or contention, I said that jitter on TDMA networks will increase at peak loading periods. This is a statement of fact.

I am not aware of Ofcom disagreeing with me. You obviously regard a few ms of upstream jitter as being a problem but then you are apparently superhuman and can notice such things. You may want to actually check out gamers and speak to them, their clients are only allowed to update the server periodically, maybe 30 or 50 times per second, 1000ms / 50 = 20ms of jitter being acceptable to avoid warping - which is an effect of one upstream datagram from a client arriving after a subsequent datagram and the server having to compensate.

The same Ofcom report that you claim disagrees with me informs that VM average upstream jitter around the 6ms mark and downstream of sub-1ms. Significantly higher than DSL but not high enough to cause issues with usage. These are the figures you think prove your point - they don't. VM can't call themselves the best for gaming but it certainly cannot be claimed that Ofcom's report describes their service as being unacceptable for gaming as a whole.

I appreciate for you, again, any jitter at all indicates congestion and should immediately be pursued until such a time as VM have all ports operating below the extremely low level where no jitter is evident at all. It is extremely low too, sub-40%.

This sounds like a lot to be operating at 6am say but as I'm sure you're aware, being a broadband expert, upload loading on broadband networks changes far, far less over the course of the day than downstream as the majority of upstream traffic is unattended bulk transfer. Downstream may swing by 75% depending on ISP policy, upstream it can be as little as a 25% increase. Note I said 25% increase, not 25% of the bandwidth but 25% of the lowest load period.

While we're talking about such things I'm also sure, being a cable expert, that you know how modems that have constant flows of traffic in their buffer will request bandwidth in with the traffic while modems that are handling small amounts of non-constant traffic like, say, a Broadband Quality Meter, don't do this so have to go through a request - grant - transmit cycle for each burst of data while ongoing flows will go request - grant - transmit + request - grant - transmit + request. This is an efficiency gain but can mean that the P2P monster with his full buffer can to an extent crowd out a BQM.

There are people with problems, a fair few, however they are it should be remembered part of a customer base of 4.3 million. What the actual numbers of ports that are in a bad state neither you nor I know. It's more than it should be, without a doubt, that still doesn't make it the typical experience.

I've spent enough time on this anyway. As you may have noted my activity elsewhere on cable has dropped to zero, no reason for it not to do likewise here.

Good day sir.

ignition I dont care about this debate with you anymore.

I will just say I have never claimed to be an expert in cable networks, and one doesnt need to be an expert to know when things are oversubscribed. Also my comment on the quality of tech support I have receieved doesnt require me to be an expert either.

Where you have dissapointed me as I used to respect you a lot is the personal attacks and the disrepect to what I post. I am not complaining about a minor increase of jitter at peak, its more then that and that all I will say to you now.

This is a thread regarding high utilization fix dates started by someone else and I simply participated in the same subject.

I am still a subscriber as the alternatives are very poor, a dodgy adsl connection that will be unstable and error prone, or something like 3G which is expensive to use. when FTTC arrives its likely I will be off at that point. Jumping ship is only simple when there is credible competition.

Good day to you also.

Edited by Chrysalis (Thu 16-Feb-12 00:42:34)

From reading his post Ignitionet has moved on and won't be responding to this or any other cable issues.

I will miss his knowledgeable input but certainly not the attitude he has displayed of late. Personal attacks are bad form and never justified. If a poster annoys you the answer is to ignore them as I have done with a certain VM staff member on some boards which support the feature.

Virgin Media are just play stupid games........bangin on about a 120mb service when there network cant even cope with [censored] 50mb. I am currently on the 100mb service which in peak hrs 5-10pm is just pathetic. They need some serious upgrades otherwise they gonna lose many customers