DLM on fibre?

talktalk are adamant that there is none but my speed has dropped over the week by 20%.

In reply to a post by ashmo:

talktalk are adamant that there is none but my speed has dropped over the week by 20%.

What do you mean by 'fibre'? If it's FTTP (ie; fibre to your property) then I can see no reason why there'd be DLM. Fibre either works or it doesn't so DLM would be daft. I thought only BT was doing FTTP though at the moment.

If it's the more normal FTTC (Fibre to your cabinet) then, yes, there's DLM. Everyone has to have it - even LLU operators like TalkTalk and Sky. I'd be with you on this - there's going to be DLM.

Edited by Andrue (Thu 13-Jun-13 15:33:33)

In reply to a post by Andrue:

In reply to a post by ashmo:

talktalk are adamant that there is none but my speed has dropped over the week by 20%.

What do you mean by 'fibre'? If it's FTTP (ie; fibre to your property) then I can see no reason why there'd be DLM. Fibre either works or it doesn't so DLM would daft.

However if it's the more normal FTTC (Fibre to your cabinet) then, yes, there's DLM. Everyone has to have it - even LLU operators like TalkTalk and Sky.

Yes FTTC talktalk lies again

As the others have said

And the first intervention of DLM is usually a drop of speed of 15-20%, unless you had plenty of capacity to be over 80Mbps anyway, and an increase in latency of 8ms.

Not lying as such, just ignorant.

Openreach run a DLM at the cabinet, so there is no way TT and Sky can avoid it. Where they are correct is that they don't have another one at their end, unlike BT Wholesale (so all other FTTC ISPs).

The BT Wholesale one doesn't control the line in the way their ADSLx one did/does, but it does set the IP Profile that you can see from a BT Wholesale speed test - Further Diagnostics, if not on TT/Sky. TT/Sky don't have a (published) IP Profile.

In reply to a post by WWWombat:

As the others have said

And the first intervention of DLM is usually a drop of speed of 15-20%, unless you had plenty of capacity to be over 80Mbps anyway, and an increase in latency of 8ms.

Just annoyed with TT support really they are basically saying your getting above the "BT Wholesale" estimate [censored] off were not going to help you :/

Estimate was 54mb Down 15mb Up

Engineer came and with his tool not sure what is called he said I should get 70mb Down and 20mb Up with no issues.

First week I was getting a solid connection between 70mb - 76mb Down and 17mb Up and a 15ms ping.

Then from yesterday I now get 58mb Down and 15mb Up with a 22ms ping.

I thought FTTC could in theory handle speeds way above 80/20.

I doubt there is much TalkTalk can do as their is no fault - you are indeed getting more than the estimate. As the line settles in it will find the programmed best combination of speed/stability. That balances the 2 - and with FTTC being early days there are limits to what ISPs or BT will do to squeeze the last ounce out of a connection (and even with the mature ADSL many ISPs will do little to maximise connection speed).

Also, it is possible that since you connected a lot of others may have connected (especially if it was a newly enabled cabinet) so you may now be seeing effects of crosstalk.

The engineer will likely have based what he told you on what he saw at that moment in time; in reality the line needs to adapt over time the the conditions that are present on your line and in your area.

The estimates BT provide will factor that in (and will be working with a larger comparable data set than an engineer would) - and quite often we see customers getting faster than the estimate for an initial period, before it gradually settles closer to the estimate. As another poster has mentioned, cross talk is a key factor - especially given a new cab won't see many for a little while - until uptake of FTTC services increases after activation.

Maximum speeds on VDSL2 can exceed 80Mbit/s - but potential speeds drop rapidly over distance and with crosstalk, so very few people could benefit from faster speeds if BT did choose to push it further. Vectoring will likely lead to improvements in crosstalk, but distance will always remain a factor with FTTC.

regards,
Phil.

In reply to a post by ashmo:

Just annoyed with TT support really they are basically saying your getting above the "BT Wholesale" estimate [censored] off were not going to help you :/

There is a very fine line to be drawn here.

In FTTC, most people do not suffer from physical line problems with the service, so there is little work needed to rectify such faults. It does happen to some people, but support staff don't get too much practice at dealing with problems like that.

In truth, we're all largely "early adopters" still, so we're getting put onto cabinets that don't yet have much take-up. The combination of early-adoption, and largely good quality lines means that most people are seeing speeds higher than the estimates. For now.

However, the biggest factor that will affect our speeds isn't a physical problem with the line, but is crosstalk - the signal for other subscribers that appears as noise to our modem. As take-up increases, the noise increases, and the effect on our modem increases too.

Unfortunately, the effect is random, and can occur at any time - totally out of control of the ISPs or Openreach. It depends on your neighbours, and what they do.

The BT estimate seems to take into account a large amount of crosstalk, though not quite the theoretical worst-case amount.

So... we all need to expect that crosstalk will arrive, and increase, as take-up increases (and there's room for a lot more yet). We need to expect that the high speeds we initially see, the above-estimate speeds, will not stay. We need to expect that FEC and interleaving will become the norm, and could be added at any time.

Hence the fine line. Your slowdown could be due to a fully-expected increase in crosstalk, for which there is no remedy as yet. Or it could be caused by a physical fault.

In one case, it is pointless sending an engineer, and a waste of resources. In the other case, it might be worthwhile.

How else can we judge which case it is? The estimate is one reasonable way to do so.

I also heard that BT will investigate if there has been a 25% drop in speed within a 7 day period. I'm not sure if this is a current threhold, or one that only applied sometime in the past.

I thought FTTC could in theory handle speeds way above 80/20.

Yes, but it is still distance-limited.

The top downstream speed for 17MHz VDSL2 is theoretically around 170Mbps very close to the cabinet, without crosstalk. 100Mbps is possible at 400-500 metres, again without crosstalk. Unfortunately, crosstalk can take away half of those theoretical speeds.

Vectoring is a big solution to these, which is being trialled later this summer. It holds good promise, but we should expect it to take 2-4 more years to upgrade the BT network, if it gets the go-ahead.

Why this reliance on what bt estimates say,?they are estimates based on little more than theory ,and not real life, just because someone has a speed that is higher or lower than what BT spew out should mean very little,
for instance they should not be used to determine if there is a fault or not as a way for an isp to avoid sending out an engineer

Edited by tommy45 (Sat 15-Jun-13 14:31:42)

You tell'em Tommy . Get nearly every ISP in the country to change it's attitude. We know you can do it!

I bet not all isp's will hide behind the often inaccurate bt speed estimates , only those who can't be bothered pushing bt openreach whilst on the other hand are quite happy to take your money each month ,
Getting bt to fix faults is part of being an isp, or it should be, and whilst some are happy to provide this there are some who aren't
But it's bang out of order where someone who has an estimate of say 50mbps but gets a higher sync or even the max 80mbps and the line has maintained this speed for several months or even a year, and their speed is suddenly cut due to DLM and the isp simply says "but you are still above the estimated speed for your line, theres nothing we can do sorry"
I in the past had a similar argument with BE support, they where trying the same spiel, all because they refused to accept or even investigate if there was a problem with trheir own llu kit

In reply to a post by tommy45:

I bet not all isp's will hide behind the often inaccurate bt speed estimates , only those who can't be bothered pushing bt openreach whilst on the other hand are quite happy to take your money each month ,
Getting bt to fix faults is part of being an isp, or it should be, and whilst some are happy to provide this there are some who aren't
But it's bang out of order where someone who has an estimate of say 50mbps but gets a higher sync or even the max 80mbps and the line has maintained this speed for several months or even a year, and their speed is suddenly cut due to DLM and the isp simply says "but you are still above the estimated speed for your line, theres nothing we can do sorry"
I in the past had a similar argument with BE support, they where trying the same spiel, all because they refused to accept or even investigate if there was a problem with trheir own llu kit

Exactly my point mate FTTC has not been around that long and I guess BT Wholesale testing for estimated speeds was not deeply investigated and ISP's just take whatever is given to them to which they don't care as I see it estimated speeds should but just ONE of the factors used not the ONLY one as talktalk are using if I was able to get 72mb for a week let me have it don't give it to me for a week and then take it away if it was stable to me, if I was only estimated to get 54mb give me 54mb in the first place and leave it at that.

+ if talktalk gave me a reasonable excuse in the first place like there have been new customers added to your cab or DLM has detected errors not just there is no reason it just is what it is basically the guy was arguing with me on the phone he didn't know anything about FTTC he kept referring to the dam exchange ha do they even have ADSL in India lmao

Edited by deleted (Sat 15-Jun-13 16:36:52)

As crosstalk has been mentioned, here's an example of 'probable' crosstalk on my connection:-

http://i.imgbox.com/acoPwHgr.gif

Back in December 2012, my connection could (& did) sync at almost 30Mb / 6 Mb.

Since then, my QLN graph has deteriorated, indicating a 'noisier' connection & Bitloading & SNR (not SNRM) have worsened.

My ISP, Plusnet, have recently arranged 2 separate engineer visits to investigate why my connection now syncs at below 21Mb / 5Mb when my speed estimate was 30Mb downstream (I can't now recall the upstream estimate).

BT's response was to lower my speed estimates to 23.3Mb / 5.4Mb!!!!

It does appear that increased crosstalk is the cause of my reduced speeds & the latest engineer confirmed that my cabinet is now almost at full capacity (thus my connection is highly likely to be experiencing reasonably high levels of crosstalk), although his tests confirm that it is in pretty good physical condition.

So, until vectoring is introduced, it seems that I'll either have to live with lower speeds, or move house to be closer to a cabinet.

BTW, I live around 1000m or so from my cabinet.

In reply to a post by ashmo:

... do they even have ADSL in India lmao

The company I work for has an office in India that has a 20 Mbps internet connection. It's in one of the few areas where up to 100 Mbps business broadband is available. Generally residential broadband out there is up to 750 Kbps or up to 5 Mbps for business broadband.

In reply to a post by simon194:

The company I work for has an office in India that has a 20 Mbps internet connection. It's in one of the few areas where up to 100 Mbps business broadband is available. Generally residential broadband out there is up to 750 Kbps or up to 5 Mbps for business broadband.

I know of people in Bangalore who at home had cable services with similar offers to Virgin Media (this was 5 years ago) of 10meg, 20meg, and 30meg, with around 1meg upload. I suspect that might be only in the area with high amount of IT workers of course.

The BT wholesale speed estimates are not just plucked out of thin air.
The estimate you were given was 54Mbps. This was the speed they thought you would get after DLM had kicked in etc. It takes into account crosstalk over time etc. The fibre estimates are far better than the ADSL ones.

It may seem stable to you however you are not constantly monitoring things - DLM is. The line was clearly not sufficiently stable, whether that's errors building up or whatever but ultimately your speeds were reduced.

It's probably down to say an extra 25 customers getting fibre in that week when you had better speeds. This would cause a gradual increase in crosstalk, followed by gradual deteriation of your line, error build up and ultimately DLM would have kicked in before matters got worse.

Sadly as uptake increases these things only get worse.

The wholesale estimates are the very best estimates that we have. They are realistic. So many people sign up and see great speeds and think damn that estimate was unrealistic, I'm getting so much better... These people nearly always end up around the estimate over time due to crosstalk.

The estimate is what you should really be looking at. That higher sync is a pipe dream and unfortunately the only way for you to get significantly better speeds is to stop everyone around you getting fibre.

You could swich ISPs but they all suffer the same crosstalk on FTTC. That's no solution.

The ISP does not have DLM on fibre. The fibre cabinet does run DLM though which is ISP independant. So whatever it does to a talktalk line, it would do the exact same if that was a BT line.

ADSL DLM was different accross different ISPs in many cases.

So they are right and wrong to say they don't have DLM. They don't but the technology used to deliver the internet to you has DLM built right in. That's not ISP provided though.

Edited by ukhardy07 (Sat 15-Jun-13 23:58:56)

+ if talktalk gave me a reasonable excuse in the first place like there have been new customers added to your cab or DLM has detected errors not just there is no reason it just is what it is basically the guy was arguing with me on the phone he didn't know anything about FTTC

Sadly this is the reality when you use an ISP which very often gives out free internet and has rock bottom prices. It's not the agents fault he's just not had the required training. You either pay more for better support at another provider (and usually a signficant amount more) or you can rely on us guys to give you this kind of info

even been close to a cabinet is no gaurantuee.

I can probably throw a stone at a FTTC cabinet from my front door, its very close.

Problem is its not the cabinet that serves me, the coverage is not a circular area.

I also do take issue with a comment that says openreach is powerless, the company that owns and maintains the lines that send the signal is not powerless, just they have chosen to employ policies that restrict possible resolutions to the problems related to crosstalk. eg. engineers doing pairswaps or testing neighbouring lines when doing new installs to make sure not adding new crosstalk, or replacing low quality cables when enabling an area, or reducing density of cables plus more.

BT installed a new dropwire to my property, my attainable jumped from high 60s to over 100mbit. Then they proceeded to move my line back to the old (shared with neighbour) dropwire on the basis of 'BT policy'. I asked for a quote to keep the dedicated dropwire but heard nothing back so gave up on it now. I was pretty shocked BT went to the expense of undoing their work just to enforce a policy.

crosstalk isnt gradual, it seems most people think whenever a new FTTC line is enabled in the cabinet it adds a small affect of crosstalk to existing lines.

Whats nearer the reality is many lines may add no crosstalk at all (not close proximity) but then other lines may add a ton of crosstalk in one go as they close in the bundle.

Its complete luck what happens. Someone could order FTTC and have speeds way above estimates with no to little crosstalk all that time, or someone can be the first on a cab and have heavy crosstalk after a week.

In reply to a post by Chrysalis:

I asked for a quote to keep the dedicated dropwire but heard nothing back so gave up on it now. I was pretty shocked BT went to the expense of undoing their work just to enforce a policy.

If the drop wire only has 2 pairs, yours and your neighbours, then perhaps the answer is to order a second line, they will have to put in a new drop wire to supply it Once it's in cancel the other line..

In reply to a post by Chrysalis:

crosstalk isnt gradual, it seems most people think whenever a new FTTC line is enabled in the cabinet it adds a small affect of crosstalk to existing lines.

Whats nearer the reality is many lines may add no crosstalk at all (not close proximity) but then other lines may add a ton of crosstalk in one go as they close in the bundle.

Its complete luck what happens. Someone could order FTTC and have speeds way above estimates with no to little crosstalk all that time, or someone can be the first on a cab and have heavy crosstalk after a week.

This whole explanation is correct, but it totally invalidates your previous post (where you ask Openreach to perform random pairswaps or to test neighbouring lines to ensure that crosstalk isn't being added)

Unfortunately, when you add one FTTC line into a cable, you do cause crosstalk somewhere. Exactly where is indeed random, but it will be somewhere.

If an Openreach engineer had to check every pair combination at install time, it'd take forever. Remember that, on addition of a further line, it isn't enough to check just the remaining idle pairs. The whole set should be re-checked.

In essence you are asking an Openreach engineer to do statically what a Vectoring linecard has to do millions of times per second..

I agree.

What would be best is if the install engineer has a device that shows all current syncs and snrms on existing lines in the cabinet fed from a hub somewhere which itself gets data from the cabinet.

He puts in a line and activates it, the device sets of an alarm if one of the following happens.

another line drops and loses more than 10% sync speed in process.
another line drops and stays down.
another lines loses more than 2db margin and line was previously only on 6db (so no buffer to absorb crosstalk).

even if there is no pair swaping to fix the problem, these events can be logged as crosstalk so if affected users log a fault then openreach know right away the reason.

Edited by Chrysalis (Tue 18-Jun-13 00:35:14)

2 weeks later now 53mb down :/