Home » ISP News »

UPDATE3 BT Reveal UK FTTC Self Install and Vectoring Broadband Trial

Posted Friday, April 19th, 2013 (8:58 am) by Mark Jackson (Score 9,718)
fttc bt mastersocket aaisp

BTWholesale has revealed the seemingly imminent launch of a new ISP trial to test the operators first wires-only self install superfast broadband service (no engineer required), which could dramatically cut the setup cost and contract length of UK FTTC lines. A trial of VDSL Vectoring (faster FTTC speeds) is also on the cards.

Currently anybody who buys a Fibre-to-the-Cabinet service has to, with the exception of some special ISP offers, pay a one-off installation fee that can cost around £100 (includes the cost of a VDSL Modem and engineer installation). The holy grail for most ISPs would be to make it so that customers could install the service themselves, which would remove the need for a costly engineer visit and potentially also make contract terms more flexible.

Last May 2012 BTOpenreach began the first step towards this goal by conducting a technical trial of Microfilter devices for its up to 80Mbps FTTC lines (here), which are a key part of the process. Shortly after that BT also began a PCP-only pilot that allows ISPs to take over responsibility of the connection hardware component by introducing their own VDSL capable modems to replace the Openreach kit (here).

Now the latest roadmap update from BTWholesale has revealed a timetable for the first customer trial of an FTTC Self Install product, which is set to begin by the end of April 2013 (Phase A) and could result in a commercial launch as soon as autumn 2013.

fttc self install bt timetable 2013

But the road to an FTTC Self Install product is not a smooth one and BTOpenreach has already confirmed that “the use of microfilters may result in reduced speeds when compared to an engineer-based installation“, which is due to a number of factors.

An engineer does quite a bit of physical work to ensure that interference is reduced to a minimum and the best line speed obtained, which is not possible when simply plugging in a Microfilter device into your phone socket. Expectations about the speed detriment vary but many fear that it could be quite significant.

Speaking of speed, FTTC Self Installs aren’t the only things on BTWholesale’s timetable. The latest update also makes mention of a “Vectoring Trial“. Vectoring is a vital technology because it’s designed to reduce crosstalk (interference) on copper based FTTC (VDSL2) telephone cables and can thus open the door to faster broadband speeds over greater distances

fibre broadband roadmap 2013

So far BT’s FTTC technology has relied upon simple profile changes to improve service speeds, which most recently saw the services headline download speed jumping from 40Mbps (Megabits per second) to 80Mbps because its spectrum allocation was raised to 17MHz. A similar profile boost in the very near future could potentially push this up to around 100Mbps but anything faster, or needing a greater reach, is likely to require something different.

Vectoring (e.g. the ITU’s VDSL2 G.vector G.993.5 or G.fast standard) is one possible solution and some tests have shown that it can deliver speeds approaching 200Mbps or potentially a lot faster when combined with additional line bonding (linking several telephone lines together), although the latter is likely to be too expensive for most home users to consider. However it should be stressed that different operators have different solutions and we don’t yet know enough about BT’s chosen method to make any reliable predictions about performance or reach.

On the surface the adoption of Vectoring sounds like a straightforward proposition but it’s unlikely to be one that comes without some form of cost, which may or may not be passed on to customers as part of an upgrade fee. Quite how much it will cost to upgrade BT’s kit for vectoring depends on many factors, such as their chosen solution and any relevant technical challenges.

The good news is that BT has at least put Vectoring Trials into their roadmap, which is a very encouraging sign that they’re keen to keep developing FTTC’s capabilities and despite the well documented shortcomings of copper line based internet access technologies (it’s fast when you live near to the street cabinet but slower when you live further away). But we’d take the dates given in the table above with a pinch of salt until BT can confirm.

As a side note the same BTWholesale update also saw the operator hint at the introduction of a Fibre Only Exchange (FOX) “strategic solution“, which would replace old copper line products with “ultra-fast” fibre optic phone and broadband (FTTP) services and has already been successfully trialled in the rural Oxfordshire village of Deddington (here).

BT now looks set to expand FOX into a more specific product offering and their roadmap suggests that we can expect to hear more about that sometime within the next few months.

UPDATE 3:00pm

As promised we have a statement from Openreach regarding the Vectoring trial, which is now confirmed as being due to commence during the summer. A further statement concerning the FTTC Self Install product is expected to follow.

A Spokeswoman for Openreach told ISPreview.co.uk:

Openreach is planning to test the capabilities of vectoring as part of a trial involving its CP customers during Summer 2013. We will provide Industry with more details on the trial over the coming weeks.‪

It’s far too early to say whether we will be deploying vectoring in our network as any decision will be dependent upon the outcome of the trial.‪ However, we believe vectoring has the potential to be a cornerstone technology of FTTC deployment in the future.

Customer feedback and the latest figures from Ofcom suggest that our fibre products are performing very well and, as you would expect, we’re continually evaluating emerging technologies aimed at further enhancing performance.”

UPDATE 3:15pm

Openreach has just furnished us with another statement, this time concerning the FTTC Self Install product.

A Spokeswoman for Openreach told ISPreview.co.uk:

Openreach is offering its communications provider (CP) customers the opportunity to pilot its “PCP Only” FTTC Connection Variant from this month. Additionally, we’re working with select CPs to trial the use of their own integrated modem/router devices. Subject to feedback from these projects, the commercial launch of the service is planned for Q3 2013/14. BT Wholesale has already confirmed that it will be participating in the pilot from the end of April.

This service will have a number of advantages for CPs and consumers. For the first time, CPs will be able to issue their own integrated modem/router rather than using the Openreach modem. This will reduce both the equipment and power consumption in the customer’s home, resulting in savings for CPs and consumers who order fibre broadband.

It will also offer CPs and their customers the option to “self-install” FTTC in the home, rather than having to arrange a visit from an Openreach engineer. This will give CPs and their customers more control over the installation experience. However, as home wiring can affect users’ headline speeds, CPs can still request that an Openreach engineer installs the product and they will carry out additional work to minimise the level of electrical interference within the home.”

UPDATE 24th April 2013:

The first two trial locations for Vectoring have been revealed (here).

Delicious
Add to Diigo
Add to Slashdot
Leave a Comment
91 Responses
  1. Anoyed Tax Payer

    You’ve read the timetable wrong in regards to the vectoring trial. The timetable is 2012 to 2013. If the trial was going to take place the timetable shows that it would have been in Q1 of 2013 and not in summer/autumn 2013. The timetable shows the trial as a greyed option, this is because the trial has not taken place because at this time the BT equipment is not compatible and would cost far too much to upgrade the cabinets and now it seems that FTTPod will cost less than Vectoring so at the moment Vectoring is way down the list of priorities for BT.

  2. Bob2002

    I’m glad I had an engineer install, he had to do some rewiring on top of installing a new master socket. Not something I would have known was necessary if it was wires only.

  3. ritchie

    Mark,

    As I understand it, the £100 charge to ISP’s covers BT’s desire to have all new FTTC subs ordered with a simultaneous new line and therefore they need an engineer visit. The cynic in me suggests this has been done in order to reduce problems they had in the early days with a new VDSL service being delivered over the existing line. FTTC services can be, and are, self-installed when a service based on a SMPF is ordered. It works well like this for a number of providers today.

    An engineer visit is still necessary even with a self installed service, but only at the cabinet, and the indications I’ve seen for volume related savings at the cabinet suggest that this would achieve would only give a discount of c. 10-20% per circuit subject to a minimum number of lines being activated at once.

  4. boggits

    BT uses financial quarters not calendar ones, so Q1 is April/May/June and (from experience) when they say Qx its almost always the start of the last month (i.e. June in this case)

    • MikeW

      Rumours about the Vectoring trial seem to have surfaced more in the last 2 days. They’re fairly consistent on a May start.

      One rumour also suggested “up to 120Mbps” from late summer – which would tie with both the end of a Vectoring trial and with the capabilities of Vectoring.

  5. New_Londoner

    As posted on TBB, it’s good to see these developments are on the agenda, encouraging to know that there should continue to be plenty of headroom between my actual bandwidth needs and the capability of my FTTC line.

    I presume those that keep pushing the “FTTC is a cul-de-sac” line will be feeling more than a little foolish! Ditto those on here that said in reply to my previous posts about vectoring and new profile changes that neither would happen here.

  6. NJay

    As a typical financial year runs April to March I intepreted chart to be

    Q3 = Oct-Dec 2012
    Q4 = Jan-Mar 2013
    Q1 = Apr-Jun 2013
    Q2 = Jul-Sep 2013

    So putting vectoring trial beginning now for about the next 4 months?

    • I’m not 100% sure whether it’s a financial calendar year or a typical calendar year. But as Vectoring is greyed out then the only clear dates are for Self Install.

    • TheFacts

      Says 12/13 and 13/14 so financial.

    • Anoyed Tax Payer

      No it’s not a financial year chart. The 12/13 at the start of the start just shows that the Q3/Q4 is for 2012 and then Q1 is for 2013.

    • NJay

      It is a financial chart as there are other clues.

      13/14 header is between Q2 and Q3 which lines up with the qtr dates I originally put
      The OR price increase for installation was this month and so inline with it being in Q1 on the chart.

    • adslmax

      Indeed, NJay. It’s financial as many people have pointed out.

  7. NJay

    Just to add I am glad I had a Engineer install too (albeit two attempts). The first installer plugged the OR modem into the office extension on existing ADSL filter as quote “It won’t affect speed”. Hmm 51Mbits DL speed on a 64Mbits estimated line.

    Complained and 2nd engineer relocated master to office & fitted VDSL2 plate. Result its currently synced at 77.4.

    A no the internal wiring wasn’t that bad as it was originally holding a 4.1Mbit ADSL2+ connection at 61dB attn as I’m nearly 4miles from the exchange. Just shows how dramatic a difference wrong filter etc may have when FTTC self install comes about.

  8. Dean

    Vectoring and Profile 30. Any trial is only likely to be in small scale, mainly because the line cards in current FTTC are only Profile 17a MAX capable.

    Openreach supplied modems (both from ECI and Huawei) are also currently profile 17a MAX only though that “MAY” be fixable via a BT pushed firmware update. Homehub 3 would still be pointless for “trialling” a vectored and/or profile 30 product as that only has a single gigabit port (rest are 100Mbps) hardly ideal for testing a product that can in theory go 200Mbs.

    Profile 17a and its theoretical ‘Actual Net Data Rate’ max is 100Mbps Down 60Mbps Up. Profile 30a potentially can do 200Mbps Down. On short lines. The port density of the 30a linecards is only something like 16, compared to 48 on current line cards so they are not very cost friendly (to BT at least) either.

    Vectoring on its own achieves nothing for longer lines as it can only cancel or not cancel noise and can only, so to speak cancel noise to a certain limit, noise beyond a limit (depending on how line card and rest of network is configured) vectoring will do nothing to as it wouldn’t be able to cancel it out.

    As to wires only installs, i see no reason why it should not be an option. However an openreach engineer currently has to go to the cabinet and as they are normally walking distance to a persons home its kinda pointless, as he may as well come and fit your filter/faceplate and do any wiring you want while he is in the area. Saves BT having to pay what must currently be a fortune in fuel to all the contractors they have taken on though ;)

    I frankly fail to see what BT are trying to achieve by this announcement apart from another wee wee contest with Virgin about who can boast the fastest MASS consumer service.

    • BT didn’t make an announcement. We hunted the data down from their April ISP Forum, which is where ISPs often meet to discuss various matters and plans with BT.

    • Dean

      Ah right Mark that makes sense then. :)

    • MikeW

      Profile 30a seems unlikely for now – the SIN specifying modem capability for “wires only” requires only 17a support. They wouldn’t launch a service that was going to be imminently obsolete.

      Vectoring is indeed useful for longer lines – it doesn’t make VDSL2 go any further than its theoretical best, but it does get real lines up to those theoretical speeds, even out at longer distances. Realistically getting 100Mbps out to 400 metres covers 50% of lines (and it might work out to 500 metres), while 50Mbps out to 800 metres is a benefit to the next 30%.

      Yes, bonding is really required to get better speeds out over a longer reach. The VDSL2 cabinets ought to be capable of it in hardware & software – so it is just a matter for BT to make it work administratively, and at a decent enough pricepoint.

    • MikeW

      Mark – there’s a slightly different roadmap earlier in the same slides from the ISP Forum.

      That one puts the vectoring trial a little later relative to the launch of FTTPoD, and similar to the launch of 220Mbps FTTP – more like July (release AI, from the key below it).

    • The new update from Openreach confirms a vectoring trial for summer 2013.

  9. I hope to have an update/statement from Openreach shortly. Keep an eye out for that.

  10. adslmax

    It’s a financial calendar. Don’t listen to the troll.

    This is an improvement and will help more people reach higher speeds, however after this the next options are pair bonding, which is potentially problematic, and profile 30a, which requires line card swap outs and may require recabling and even additional cabinets as the current 30a cards have lower port density than the existing 17a cards.

    • Anoyed Tax Payer

      @adslmax Sorry I don’t understand your comment “It’s a financial calendar. Don’t listen to the troll”

      As far as I know nobody has commented on the calendar, just the project timeline, which if you’ve ever done project management you would know that they are not done on a financial calendar.

    • Ignitionnet

      Max – copy/pasting other people’s comments from other sites without giving credit or asking permission isn’t wise.

      ‘Posted by Dixinormous 8 days ago
      This is an improvement and will help more people reach higher speeds, however after this the next options are pair bonding, which is potentially problematic, and profile 30a, which requires line card swap outs and may require recabling and even additional cabinets as the current 30a cards have lower port density than the existing 17a cards.’

  11. MikeW

    Will vectoring cost any more to the end-user?

    On the face of it, I think not. For the principle of vectoring to work, it needs to be running for every line in a cable binder – irrespective of whether they want the faster speeds or not. Irrespective of whether they’re willing to pag for a gigbit-capable modem/router. And irrespective of what ISP they are with.

    That means BT need to include everyone – and so anyone on a current 40/10 or 80/20 package will need to be included (and will need a vectoring-friendly modem or firmware update), and so anyone will consequently benefit from vectoring… up to their current package limit.

    However, with widespread vectoring, BT will get the ability to start selling 100Mbps headline packages (or possibly even 120Mbps packages) – and people taking those certainly could be charged more. *These* people alone start to need a modem with better throughout capability – a fact which makes it less likely that there will be an across-the-board increase from 80/20 to 100/20.

  12. G1h

    For vectoring to work really properly it needs all those on a cable bundle still running ADSL2+ from the exchange to be forced to convert to FTTC from the cabinet. Otherwise the Dslam vector controller will not know about or be able to take into account/modify the effect from the ADSL2+ signals.

    • MikeW

      While this is true, in reality it won’t make much difference – I read it somewhere, but can’t find my reference.

      For a start, ADSL2+ only inhabits the first 2MHz of the spectrum, while VDSL2+ inhabits 17MHz.

      Next, the ADSL2+ signal is already at reduced power (and probably has lost some of the higher frequencies) when it reaches the PCP cabinet – so has less ability to add interference to a VDSL2+ signal

      Finally, the power maps applied to the VDSL2+ signals already hobble the amount of power being used by the FTTC cabinet at the ADSL2+ frequencies that are relevant to that cabinet’s location. That causes modem sync negotiation to markedly underuse those frequencies anyway.

      In practice, the FTTC service is deliberately under-utilising the frequencies that collide with ADSL2+. In which case there isn’t much to be gained, relatively speaking, from applying vectoring to them.

  13. Dean

    “…and so anyone will consequently benefit from vectoring… up to their current package limit.”

    Unlikely. Vectoring will only help those significantly (and i do mean significantly) on lines shorter than 400M, with only guaranteed (or as near as can be guaranteed) serious (and i mean serious IE 10Mbps and higher) performance increases for those at 300M or less.

    Once you get to 350M (ish) the results become more down to actual line quality. Some lines in this country which have been in place for years and suffer regular water ingress the benefit on any half normal distance line (IE around the 300M and longer point which id guess most in the UK are to the cabinet) are likely to see a much smaller if any benefit.

    Vectoring on a poor and even some average lines can actually be a bad thing as the system attempts to constantly adapt (for want of a better term) to all the snap, crackle, pop some creaky old line makes. In instances like that DLM which BT already use on its own probably gives a better benefit and certainly ensures the line is more stable, both in terms of staying connected and consistent thruput.

    Furthermore vectoring on a long line (around 800M and longer) can actually DEGRADE performance, there probably are not too many lines which are say more than 1000M to a cabinet in the UK but i imagine there are a few. On a line like that an NON-vectored VDSL2 line would perform better, infact for anything over 1000M the best solution for those would be a bonded VDSL2 solution, but thats even further off than daydreamy vectoring, profile 30 and 120Mbps FTTC products.

    I also doubt BT will be allowed to market a vectored VDSL product at 100 or 120Mbps as 10% of their user base would have to achieve that rate, according to ASA guidelines. Right now people which are under 300M to a cabinet can not get the full 76Mbps on a non-vectored product. I live in a street full of em ;) My own line is just over (and this is actual distance) 250M to the cabinet and yet my FTTC connection tops out at around 72Mbps, which i, considering i know the line is good quality consider the result to be very poor.

    The biggest stupidity of all this “fantasy new tech coming for faster speeds” dribble i find funny though is the current FTTC solution can manage up to around 60Mbps up rates, yet BT in their wisdom have capped everyone at 20Mbps max. Its not like vectoring is even needed when you have a product which has been throttled back from doing its max already.

    • MikeW

      Dean, you manage to write a lot about the vectoring technology, but it is entirely at odds about what you can read elsewhere. Have you actually read anything?

      1) “Vectoring will only help those significantly (and i do mean significantly) on lines shorter than 400M”

      Take a look at the Broadband Forum’s “An Overview of G.993.5 Vectoring” (MR-257 May 2012). Link: http://www.broadband-forum.org/marketing/download/mktgdocs/MR-257.pdf

      Figure 6 shows a wealth of information, but isn’t 100% applicable to UK wiring:
      - At any one distance, the blue crosses show the range of performance for the different non-vectored lines at that distance. At 300 metres, the range is between 65 and 98Mbps. A large difference – as seen in practice down your own street – the variability in crosstalk there proving the point of this figure!
      - BT’s current estimating tool seems to estimates speeds around halfway between the lowest blue cross and the lower triangle. It certainly needs to be pessimistic enough to match that bottom cross or be below it.
      - The red circles show the range for the same lines with vectoring. At 300 metres, the range is between 145 and 150Mbps. Considerably less variability as well as being a lot higher capability.
      - Note something about those red circles? Even at 1km range, they’re still at double the speed of the lowest blue cross. Vectoring is having a significant impact out that far still.
      - The figure itself shows that 100Mbps is perfectly good out to 500 metres. 120Mbps would go to 350m, and 140Mbps to 200m. A lesser 50Mbps would get out to 850m.

      That is for the Annex A, so figure 7 would be more appropriate for the UK (we’d follow the BB17 labels). However, that is still for 0.4mm wiring. I thought (but could be wrong) that we used 0.5mm wiring.

      I’m sure reality will not prove so good, but it is looking far more optimistic than your “serious” and “guarantee” statements. I guess that is what a trial is for…

      And BT wouldn’t go from their internal trial at Martlesham to a live CP trial unless their trial showed some promise. Sufficient promise to consider upgrading the hardware, and still be calling it “a potential cornerstone”.

      My personal cornerstones for this technology are:
      - The original FTTC
      - The 17MHz rollout
      - Vectoring
      - Bonding for longer distances

      2) “Vectoring on a poor and even some average lines can actually be a bad thing as the system attempts to constantly adapt… to all the snap, crackle, pop …”

      A reference? Everything I’ve seen suggests it works just fine. It can’t improve it, but it doesn’t make it worse.

      3) “In instances like that DLM which BT already use on its own probably gives a better benefit and certainly ensures the line is more stable …”

      Vectoring (aka DSM level 3) doesn’t replace DLM (aka DSM level 1). In fact, for the reasons you mention, DLM is still required.

      4) “Furthermore vectoring on a long line (around 800M and longer) can actually DEGRADE performance,”

      Do you have a reference? There isn’t a single paper from industry, analysts, or research that suggest this, or warns about the possibility.

      5) “I also doubt BT will be allowed to market a vectored VDSL product at 100 or 120Mbps as 10% of their user base would have to achieve that rate, according to ASA guidelines.”

      According to TBB, 20% of the UK’s lines are less than 200 metres. If they got results like the figure I discussed earlier, and they chose to, they could probably run it as a technical “up to 150Mbps” sync speed with a marketable “up to 140Mbps” throughput speed. I expect lower ambitions of either 100-120 (marketable)

      From the same source, 60% are within 500 metres (100Mbps target), 80% are within 800 metres (50 Mbps target) and 98% within 1500 metres (as you say, good candidates for bonding).

      6) “My own line is just over (and this is actual distance) 250M to the cabinet and yet my FTTC connection tops out at around 72Mbps,”

      My line is 400 metres of old copper, and I currently get 80/20. The attainable has gradually dropped from 90 to 77 over time (yes, I’d drop from 80/20 to 77/20 if I resync’ed now)

      Both of our lines go to show the variability caused by crosstalk. There is nothing wrong or bad with your line, or your service – rather it demonstrates why vectoring is a good thing!

      7) “all this “fantasy new tech coming for faster speeds” dribble”

      The thing is, neither vectoring or bonding are fantasy – they’re too real. A shame that BT aren’t paying much attention to bonding though.

      For fantasy tech, you need to be looking ahead, at FTTdp, G.fast and the phantom modes. Expect them to be standardised in the next year or two and considered for use in maybe 5-6 years.

      8) “yet BT in their wisdom have capped everyone at 20Mbps max.”

      Oh yes. Ironic too that, at long distance, it is the upstream that gives out first.

    • FibreFred

      Yes I think Dean-duction needs to do more reading

    • Dean

      1) That link also thinks the max speed of vectored (page 12) is around 170Mbps on a vectored line and they even have tagged that chart as “WORSE CASE” i don’t think so.

      2) Vectoring either cancels out noise or does not, on a line where it can not cancel noise the vectoring does nothing except hinder.

      3) I was not saying it replaced it, i was attempting to explain on a poor line DLM does a good enough job of making the line as reliable as possible already.

      4) Im talking real life. ADSL2+ in its early days ‘charts’ all said that would improve everyones speed also, the reality turned out those with a poor or long line are better off on regular ADSL. VDSL will be similar, if a line is rubbish it is rubbish no matter how much wizardry tweaking you throw at it.

      5) Do you have any link to BT figures that confirm 20% of lines are less than 200M? I assume the TBB link for this 20% at 200M is this… http://www.thinkbroadband.com/guide/fibre-broadband.html which i can find nothing to back up as being true.

      6) The IP profile for my line is 76.02Mbps the actual line length of copper to the cabinet (ive personally measure it, saw it installed,d google maps agrees and BT bloke that came out also confirmed the cable route) is 252 Metres. Actual speed i get is around 72Mbps, anyway you cut it that is poor IMO. It is as good as brand new cable compared to most of the UK (house was built around 10 years ago). It performs nowhere near any ‘lab like’ chart you can show though and neither will profile 30 or vectored tech.

      7) Fantasy as in wild promises for higher speed not the tech not existing.

      8) considering BT use DLM there is no need to RATE LIMIT the upstream to 20Mb, lines that can do more would get more, lines that can only do 20Mb or less would get less, no different to how download works.

      PS if you look at alcatel lucent specs (who BT actually have agreements with) vectoring is not as peachy as the 170Mb charts you have come up with. Which again goes to show if the experts can not even agree what the MAX speeds are than any chart like info you read is nothing more than hopeful theory.

    • Dean

      @Fibrefred

      With your everyone that has anything which can be seen as negative to say against BT hallucinations and that everyone is Deduction and out to get you. Seek help, im actually a BT customer, recent speed test for you to back up what i have stated in my posts about my line…

      http://speedtest.net/result/2657408685.png

      Hopefully that will shut your pie hole. Then again as a regular reader here but not an often contributor over the past month it seems you have thought a dozen or so people are Deduction. It is thus clear you are way beyond any type of mental help.

    • MikeW

      Ah Dean, thank you for leaving such well though out criticism to my post.

      1) “I don’t think so”.
      I’m pleased that the best efforts of many telecoms researchers has failed to convince *you*, but please try to persuade *us*.

      The graph isn’t labelled “WORSE CASE” – it just has a label telling you that “WC” means “worst case”. If you look closely, WC is used in the “99% WC” portion, which are the blue upside-down triangles.

      The graph actually shows real-life best cases and worst cases – the range of blue crosses. The triangle, below the crosses, is then the theoretical point for the “99% worst case”.

      The serious point here is that BT’s current estimates are close to that “worst case”, while vectoring lets them get up to the red circles.

      2) Thanks for the reference.

      Have you got anything to back up the fact that vectoring *hinders*. At any point, anywhere? Absolutely everything I have seen, while figuring out Vectoring’s performance with alien noise, shows that vectoring has no negative impact.

      Anything, beyond your gut feel?

      3) Agreed – DLM does make a good job of making a line reliable. Almost all of the time.

      However, DLM can only ever react to crosstalk – the problem of which is notably growing as take-up increases. Where it reacts, it makes the line slower and keeps it relaiable.

      Vectoring’s job is to take away the crosstalk, improve the SNR, and allow the speed to raise – back to near the theoretical best.

      DLM will still be needed to react to alien noise that vectoring cannot cope with.

      4) If you are talking real life when saying that vectoring degrades performance on lines > 800 meters, where is the evidence you are using?

      You must have someone’s experience for this. A link you can point to, right?

      Or are you not talking real life at all? Are you really just employing scepticism and wrapping it up to look like fact?

      If you want an example of real life, here’s a link to a report from Alcatel-Lucent on the experience from their trials over the last 2 years – which would be real life customer trials: http://www2.alcatel-lucent.com/techzine/vdsl2-vectoring-delivers-on-its-promise/

      Note how the speeds have improved over the course of the trials, as the techniques have been learnt and improved?

      5) That’s my source – and I don’t have a better breakdown, beyond knowing the “averages”…

      BT do presentation slides that gives us some snapshots of the access network:
      - “Typical D-side” is 420 metres of 0.5mm with 25m drop line, from 2007 http://www.bcs.org/upload/pdf/emacfadyen-200307.pdf
      - “Average D-side” is 500 metres with 30m drop line, from 2011 http://www.bcs.org/upload/pdf/sfisher-090311.pdf
      - “Premises average 500 metres from fibre network”, from 2012 http://www.techweekeurope.co.uk/news/bt-openreach-fttp-ultrafast-broadband-101079

      6) “The IP profile for my line is 76.02Mbps … Actual speed i get is around 72Mbps, anyway you cut it that is poor IMO.”

      At a sync of 78Mbps, Your line is two-thirds of the way down the set of blue crosses for 250m. It is acting as expected in the presence of crosstalk. My line (halfway down the blue crosses at 400 metres) likewise.

      Both are performing slap bang in the middle of the “lab-like chart”. Are you not seeing that?

      Can that be described as poor? I’m not convinced – trailing off the bottom would be poor, near the “99% worst case”.

      7) It is healthy to be cynical, but not *too* cynical.

      ADSL2+ gave us excellent “up to” charts, but it is people’s interpretation that determined their optimism. BT’s own expectation of ADSL2+ was that it would create a median speed of 6-10Mbps – Ofcom’s latest broadband report shows that BT’s ADSL2+ speeds average between 8 and 10Mbps.

      Yes, taking the highest figure and touting it as “average” is wildly optimistic. But denying reasonable evidence is being equally wildly pessimistic.

      My personal optimism level cuts in at the 100-120Mbps level, rather than the 140Mbps level.

      8) Upstream has different problems (including the fact that Openreach use it as a big differentiator in the price charged). However, it is the limiting factor for long reach VDSL2, and without vectoring, those long distance lines really need to shorter ones to run at a lower power (upstream power backoff). If openreach don’t limit upstream effectively, to something a lot lower than the short-range lines could achieve, those long-distance customers get crowded out.

      PS) Alcatel Lucent are one of the leading edges of vectoring, but they aren’t used in our FTTC access network. If BT have agreements with them, it is for something else.

      However, ASSIA also seem to be a leading part of vectoring – and that company provided the editor to the BBF document I linked to, as well as Huawei (who do supply BT with VDSL2 linecards).

      Yes, I only linked to one document, but I do follow what both Alcatel and ASSIA are saying on the subject independently. I chose the BBF document because it gives a very good picture of best & worst case, rather than 1 line of “up to”. It adds a dose of realism.

      However, I also like some of the reports here, except that most graphs do not detail the profile of VDSL in the test: http://www2.alcatel-lucent.com/techzine/vdsl2-vectoring-delivers-on-its-promise/

      In that article, Alcatel-Lucent say “vectoring trials, conducted by Alcatel-Lucent with various CSPs around the world, have demonstrated that 100 Mbps is achievable over copper at 400 m — and even up to 500 m.” All their marketing material (including their latest press releases for deploying 1 million vectored lines) focus on this 100Mbps level. Focus *heavily*. They must be pretty confident about this, right?

      Funnily enough, the graph from the BBF document says *exactly the same thing*.

      And funnily enough, 400-500 metres also happens to be the UK’s average D-side. Except we use thicker copper than all those trials.

    • Dean

      1) I believe you meant to say “thought” out criticism. Regardless as said your chart as stated at the end of my prior post is nothing more than hopeful theory, nothing to do with research.
      I guess the bunch here http://www.ospmag.com/issue/article/vectored-dsl-rescue whos chart says vectoring can not even do 100Mbps is wrong as that conflicts with your opinion.
      Or maybe this one is right…
      http://www2.alcatel-lucent.com/techzine/boosting-vdsl2-bit-rates-with-vectoring/
      nah still not near the daydream 170Mb theory you picked out.

      2) you can go read the actual ITU spec so no it is not gut feeling….
      http://www.itu.int/rec/dologin_pub.asp?lang=e&id=T-REC-G.993.5-201004-I!!PDF-E&type=items

      3) And that is part of the issue Vectoring can only remove certain amounts of noise, you could have a line so noisy vectoring is unable to make any improvement or remove the noise. It is not a solves everything solution.

      4) That is not what i stated, i clearly pointed out it would be for long and POOR performing lines. Any chart you produce is likely lab test or best case.

      5) Those figures and the percents seem to conflict with TBB info, which is correct? I suspect none of it considering BT thereselfs often deem info like that commercially sensitive. Again nothing more than some blub theory on the internet.

      6) It is poor performing ive seen longer lines than mine which are also older perform better. That is real life and not based on some chart so you can ignore that again.

      7) Ill agree with you and the 100-120Mbps MAX for vectored VDSL2 guesstimate, anything more as your initial chart shows (IE OVER 170Mbps) is likely to be utter fantasy, even with a perfect line i doubt many will get near that figure, that is pushing on for 200Mbps which is what profile 30 (SUPPOSEDLY) does under good circumstances.

      8) Ill tell you what part of the problem with BT FTTC and upstream is and why people like myself which should get their full but rate limited 20Mbps dont get it. It is due to stupid QOS set on their locked down modem, disable that and you can instantly see another 2+Mbps upstream, ive seen that with a previously borrowed unlocked modem (same model as my locked down one :( ) Look around at all the speed tests people have done and its rare you see one over 18Mb. You are lucky to see one at 18Mb, the normal best is around 17Mb and thats if people are lucky. Why they have locked it to 20Mbps max and enabled QOS on the modem is a mystery though. Any guess again is just theory.

      As to the rest, as stated in (7) i agree 100-120Mbps MAX is likely what vectored will give on VDSL2… NOT OVER 170Mbps like your first chart shows. I assume now you have looked around at other charts you accept the initial claimed MAX speeds of over OVER 170Mbps is a daydream. If not why does all this so called “researched” MAX SPEED data conflict so much site to site???

    • MikeW

      Thanks Dean, you can have the “thought out” epithet without irony (and spelling mistake) this time – the links & sources you are using help immensely.

      1) I disagree. My chart stems from research, and offers the great advantages that it tells you what flavour of VDSL2 was involved (“Performance of G.993.5 for the VDSL2 Annex A 17a band profile”) and when the data pertains to (May 2012). It also lets you see all results – the whole range min to max, so you can determine the average yourself.

      (I go into more detals about how I rate/trust this graph at the bottom)

      The other chart I reference (on Alcatel’s techzine site “VDSL2 Vectoring Delivers on its Promise”) isn’t as good – it doesn’t show the flavour of VDSL2 but it does give dates: The article is dated Sept 2012, and shows a range of dates in the trials (July 2010 to April 2012). It doesn’t indicate whether it is showing average speeds, minimum speeds or maximum speeds

      That last article shows date is critical – recent results are more impressive as the technology (and the management of it) matures. Results therein in 2010 get 100Mbps at only 200 metres, while results in 2012 get 130Mbps at 300 metres.

      Early results are also restricted – the test equipment often couldn’t actually go higher than 100Mbps. Tests tended to top-out at that point.

      It proves that graphs must be chosen with care – and not just discarded because you don’t believe the results (good *or* bad)

      Your 1st “ospmag” reference has a few problems. There is no date, which means it may be presenting early results that we could suspect to be immature – the latest dated item in the references is January 2010, suggesting it is indeed old. I used ospmag’s search function, and it told me the article was dated April 2010 (I won’t mention April 1st).

      Likewise the graph doesn’t show the flavour of VDSL2 that was tested, the bandplan, nor the capability of the copper it was tested on.

      The first datapoint on that graph is at 300 metres, showing nothing shorter – you wouldn’t expect the highest results to be visible there either.

      The graphs here seem to match with what I’d have expected in early 2010.

      The author of this article has written later papers, with better graphs. See later (a lot later).

      Your 2nd graph (“Alcatel”, “Boosting VDSL2…”)shows better results than the first, but still suffers problems. The date is now September 2011 (and the results show some maturity on from 2010).

      However, the graphs don’t detail VDSL2 flavours, and don’t tell you whether the “typical gains” are minimum, average or maximum. It also tells us it is “near-optimal”. Other articles suggest that 98% vectoring is only enough to get 50% of the gain, so how much is “near-optimal”?

      Despite its flaws (and only going down to 100m), and the slightly older date, I’m still reasonably happy with this second graph. If we can get to this “near-optimal” in the real world, and get 100Mbps out to the 350 metres it suggests, I’d be pretty happy.

      I have another set of graphs to show you. These are from Ericsson, and very old (dated 2009) but they are fantastic at showing the details behind the tests:
      http://cms.comsoc.org/SiteGen/Uploads/Public/Docs_Globecom_2009/Vector_globecom2009_final_v2.pdf

      Page 7 shows the usual result for early tests, of 100Mbps over 1000ft (300m), on 0.4mm copper.

      Page 9 shows a result for 0.5mm copper, which is apparently what BT use.
      This shows 100Mbps over 2000ft, or 600 metres.

      Given that almost all tests show performance for 0.4mm copper, that last one is, at the very least, promising for us in the UK.

      2) Funny, mate.

      I love reading specs for the fun of it, but if it you want to prove that “where it can not cancel noise the vectoring does nothing except hinder” I’ll need some pointers. I accept the “does nothing” part, it is the “hinders” bit that is important.

      3) Going any further on this point is only worthwhile if (2) becomes apparent.
      Not much point arguing here.

      4) Nope. Your full text reads: “Furthermore vectoring on a long line (around 800M and longer) can actually DEGRADE performance, there probably are not too many lines which are say more than 1000M to a cabinet in the UK but i imagine there are a few.”

      No mention about poor performance in that paragraph.

      If you really meant to qualify it, then we’re back to discussing (2) again, I guess.

      5) Does it conflict with TBB? How does that make it a “blub theory on the internet”? (Whatever that is)

      TBB’s surrounding data tells you that 45% of lines are < 400m, and that 60% are < 500m. That suggests the 50%tile is somewhere between 400m and 500m, nearer 400m.

      Add some flexibility, as some of the statements are "typical" and "average", which we could interpret to be mean, median or mode.

      My personal interpretation is that "typical" is likely to map to median, and that "average" *may* map to mean. If TBB data is for median, it's on point. If it is for mean, it is out somewhat but not out of the ballpark.

      To me, it is hard to call the data "in conflict". You've gone down in my estimation for such obtuseness.

      6) At this point, you ought to drop into my "to ignore" list.
      Of course there are longer older lines that perform better than yours.

      That graph, that you dismiss so instantly, shows that the best (luckiest) subscriber at 600 metres can get the same speed as the worst (unluckiest) person at 300 metres.

      Longer lines with no crosstalk will work brilliantly. The lucky ones get no crosstalk.

      Short lines that are unluckily bound with a disturbing neighbour and a lot of crosstalk will suffer more hugely.

      It is all down to the luck of which customers sign up to FTTC, and where their copper pair is located relative to yours within the binder.

      If you don't understand this aspect, you most certainly aren't qualified to talk about vectoring. Truly – those blue crosses mean *a lot*.

      BTW – Older copper is not worse. It is older joints that have problems, not older copper. And joints can easily be remade (though hard to persuade Openreach to do).

      That doesn't include overhead drop cables though – there the joints and the insulation get to suffer the effects of weather more. But drop cables average 20-30 metres.

      7) ok

      8) You are so right there.

      There's probably a reason. But until I find out what it is, I have my QoS turned off.

      PS2) "I assume now you have looked around at other charts you accept the initial claimed MAX speeds of over OVER 170Mbps is a daydream. If not why does all this so called “researched” MAX SPEED data conflict so much site to site???"

      Simple answer:
      - I don't daydream about 170. I think you're the one who has become fixated on it.

      - Instead, the 2 questions in my head are: "How far out will 100Mbps reach?" and "Will BT headline at 100Mbps or 120Mbps?"

      - Why so much max speed conflict? It depends on the date especially (due to maturity of the algorithms), but also the the max capability of the test equipment (which has improved over time), then the VDSL version used, and the copper diameter.

      Longer answer:

      I haven't seen enough graphs that are a) new enough, b) annotated with vdsl flavour, copper diameter etc, c) give behaviour for below 300 metres, d) tell me what the graphed-line really represents, for me to have formed a solid opinion of what the *real* capability is, currently, at the very lowest distances.

      What I do instead is look at a series of graphs, work out the date, and work out what information is missing – that gives me an initial level of "trust" about them.

      Then I look in the graphs at where the lines crossover the 100Mbps level – the distance that occurs at. I use that factor to determine whether to treat the whole graph as fantasy, or whether it is worth scrutiny: 2010 graphs that cross around 300 metres are fair, while 2012 graphs that cross between 400 and 500 metres are fair.

      If I think the graph is relatively fair, even if it is missing information, then I try to work out what else it is telling me.

      The BBF graph I adjudge to be fair around 100Mbps (but I assume that other, single line graphs, plot their line at the *bottom* of the red circles, and mentally I do the same here). On this point, it agrees with current statements from Alcatel-Lucent – so seems to be the current state-of-the-art for 2012/2013.

      Although I adjudge the graph to be fair, I have had little corroboration of those extreme figures at the low end. I treat them as unverified, but not as impossible. I continue to look for recent information to help confirm the figures, or otherwise (see below)

      Further, I look at the location of all those blue crosses, and relate them back to reports I see online from other users (their speed reports and their distances). I look at the inverted triangles, and relate them to estimates I've seen live in the UK so far. All these items *do* seem to fit reasonably well with real life experience – giving the blue sections, at least, an element of trust.

      Conclusion:
      I'm more trusting of the section between 300 and 900 metres than other parts.
      The rest I rate as "interesting" and look for corroboration.
      I do give credence to the blue areas.

      Another article:
      I have another research paper – this one on the effect of vectoring in the presence of un-vectored lines (from another cabinet, for example). Date not clear, but looks to have been presented in May 2012.

      http://www.assia-inc.com/technology/knowledge-center/industry-articles/Vectoring-Compatibility-CISS-2012(Web).pdf

      This is by one of the inventors of vectoring, so I start with an element of trust on it.

      Figure one in there shows a theoretical Fext-free performance (a slight dashed line that is hard to see) that also comes up with 170Mbps out to 200 metres, and 100Mbps to 100 metres. It seems to be matching.

      It also shows the bounds of unvectored performance in real life (the lines with squares and circles).

      *** Note that one of the authors of this paper is the SAME as one of the authors of your 1st linked article on ospmag.com ? That really old one?
      *** That author is John Cioffi, father of DSL, inventor of vectoring.

      Still wondering why the newer graphs go to higher speeds?
      Is it possibly because we're better at it?

      The same information can be seen in a similar paper, on vectoring in multiple cabinets (April 2012):
      http://www.assia-inc.com/technology/knowledge-center/white-papers/FASTWEB-ASSIA_White_Paper_on_Vectoring_(April%202012).pdf

      So, in conclusion, and how I see my questions being answered…

      I'm open to accepting that 170Mbps is possible and accessible to academic researchers in the labs, knowing the limitations of comparing such labs to real life.

      I'm open to results from the likes of Ericsson that show 100Mbps on 0.5mm copper to 600 metres. I understand that these will still be labs, but more practical.

      I'm open to seeing real-world results from Alcatel-Lucent. Where recent results show 130Mbps at 300 metres. I want to question them about how much real-world noise got into the trials here.

      How does this apply to real life, here in the UK?

      I reckon that we'll easily see 100Mbps for some portion of customers, well above ASA guideline limits. It could be 500 metres, and just above 50% of properties, and it could be 400 metres, and just below 50% of properties. I think it *will* be more than 300 metres – Alcatel's full-on push on this very point persuades me, on top of the other evidence (not the fact they're pushing, but the *way* they're pushing).

      Will we see 120Mbps? I believe some could, but whether it is enough for BT to justify a headline package here is another matter. I think they want to though…

      I reckon it will still have an effect out to 1km at least. Subject to (2) ;)

      I'm off to watch this:
      2013 lecture on future gigabit internet access, by John Cioffi, father, inventor yada yada yada.
      http://geromedia.usc.edu/Gerontology/Play/c66d78ed8e22483599c47f84a63849411d

    • MikeW

      On (2)

      The ITU Spec itself seems to only talk about distances in the “scope” section, stating that “[Vectoring is] particularly beneficial with short cable lengths (< 1 km) and limited near-end crosstalk (NEXT), background noise, and FEXT from systems which are not a part of the vectored group (alien noise)."

      There is no mention of your 800 metre limit; no mention of what distance vectoring continues to be beneficial but less than "particularly"; no mention of where vectoring stops being beneficial; and no mention of where vectoring starts to hinder.

      "Limited near-end crosstalk (NEXT)" is a design feature of VDSL, so can be ignored.

      "Limited background noise" is an interesting statement and worthy of discussion… except the spec does nothing to articulate what "limited" means.

      "Limited noise from non-vectored systems" is a given for now, although being studied and handled.

      Got a more useful reference that you can add to the discussion? Otherwise we're back to assuming you have a gut feeling.

      In fact, I think the gut feeling is being expressed with the larynx in your colon.

    • MikeW

      And finally, on (2) again…

      I’ve been reading TR-197, the Broadband Forum’s document on “DQS: DSL Quality Management Techniques and Nomenclature”.

      This document is the one that gives you hints and tips about how to deal with all aspects of DSM and DLM.

      The best that this document can come up with is: “However, other
      impairments such as impulse noise, radio frequency interference (RFI), and crosstalk from other services such as ADSL, are considered as alien disturbers, which vectoring does not address.”

      “DQM methods for dealing with these alien disturbers, such as DLM/DSM Level 1 and
      DLM/DSM Level 2 are still valuable when used on vectored lines to ameliorate impairments from these alien disturbers.”

      Again, there is nothing whatsoever about maximum line lengths where vectoring may be effective. There is nothing about limitations to the background noise level. There is nothing about REIN, PEIN, SHINE or RFI interference save the quote above – which recommends that you continue to use DLM.

      There is nothing that shows vectoring gets hindered at any time.

      There is nothing that shows the DLM alone can be better than vectoring. At any distance, or any noise level.

      If this information isn’t in the ITU’s spec, and it isn’t in the Broadband Forum’s guide to DSM techniques, just where is it going to be?

    • Dean

      1) SO short version, you can not prove your chart is right you just believe it more.

      2) As explained in the ITU spec which clear mentions there are limitations to what vectoring can do. You go on to quote it even later in your posts

      3) Statement also proven in ITU spec, as you have later highlighted

      4) Now you are just selectively quoting rather than reading in full see prior paragraph i type which clearly states… “Vectoring on a POOR and even some average lines can actually be a bad thing….”

      5) Percents quoted on TBB and your other stats conflict, so either A) You can not prove the TBB info is correct or B) its just a internet theory. I suspect a bit of both and its more stats just like item one you believe as it suits your mission, whatever that may be.

      6) I can only hope you do ignore me. Somehow though guessing who you are i doubt that will be possible. I also do not care what any graph says with regards to my own connection, i am the one sat in front looking at what speed i get and have sat in front of another persons machine looking at what they get, my line which is shorter and in theory better quality performs worse. Thats the problem with charts, they do not match real world experience, something i doubt you have much experience of.

      7) im glad we agree 100-120Mb is likely what Vectored VDSL2 will give so im at a loss to the rest of your post and your still insistent belief in your 170+Mb chart.

      8) If i were BT id advertise it at 100Mb as that is likely far for achievable for many over 120+Mb. However the likes of BT, Sky, Virgin all like to have childish wee wee contests so BT are more than likely going to advertise it as 120Mb to so called “match” Virgins top speed.

      As to conflicting data, it is nothing to do with the date but down to the rest of the infrastructure VDSL is running on, you highlighted an example early on that being cable gauge as just one of the many things. Type of line card used would be another and there is so much more, which is yet another reason you can not believe some chart.

      As to your first assia link surely that blows your theory on date out of the water if suddenly you are going to believe an chart which is older. Date a chart was produced is irrelevant. The actual speed a connection protocol can make is just that, we do not get better or worse at it, if we did using that logic ADSL2+ by now should be running at something like 50Mb LOL And similarly the HONEST people market that as around 16Mb the less honest try to claim its 20-24Mb, guess which of those rates a larger percent of people get?

      THEORY on what a INTERNET CONNECTION can go at and what it ACTUALLY performs at are 2 very different things.

      170Mb VDSL2 is nothing more than theory, if it is not we can look forward to BT marketing their product as 170Mb. In actual fact its likely to be marketed at 120Mb to match Virgin or they are likely to try and get away with selling it as 150Mb to beat them in the wee wee contest. Either way 170Mb is chart driven fantasy.

    • New_Londoner

      @MikeW
      Good background knowledge and references, an interesting read. NB the TBB stats on line lengths from cabs were in turn sourced from a report published by Ofcom around a couple of years ago.

    • Dean

      That would be an interesting report to read, do you have a link?

    • New_Londoner

      @Dean
      See the following link:

      http://stakeholders.ofcom.org.uk/binaries/research/technology-research/asses.pdf

      The report is a little dated, but the basic stats on line length should still hold true.

    • MikeW

      @New_Londoner

      Thanks, that’s of great benefit. It also confirms that we tend to have 0.5mm wiring (as stated in SIN 351), which means we stand a chance of having markedly better vectoring performance than many of the theoretical/trial graphs show.

      The document matches what I’ve seen before on total distance to exchange, but it is the first time I’ve seen the data on just the D-side wiring.

      The only negative point I can make is that BT tend to distinguish the D-side wire from the drop wire when talking of the distances. The report’s authors aren’t clear that their method (attenuation of 300kHz from cabinet) goes to the property, rather than the DP.

      However, I found a different Ofcom report (Analysis Mason, dated 2011)
      http://stakeholders.ofcom.org.uk/binaries/consultations/wlr-cc-2011/annexes/line-length-cost.pdf

      This report refers back to the one you linked, and states that it *does* include the drop wire. Unfortunately (for us) it only details the newer findings for total line length (and focusses on lines that happen to be ADSL2+ customers).

      There is an interesting (and uncorroborated) quote: “it was indicated that
      Openreach considered that copper local loops in urban areas are using thicker copper cables (higher gauge).”

      Conclusion:
      - 15% of lines are <200m
      - 47% of lines are <400m
      - 70% of lines are <600m
      - 83% of lines are <800m
      - 89% of lines are <1km

      Seems TBB were in the right ballpark after all.

      @Dean

      Truly, I am still seeking evidence to build into deciding my trust level on all areas of vectoring. If you have evidence of the problems you describe, I *want* to know. I *want* to be able to discard the fancier evidence that I have seen, and downgrade my beliefs.

      But so far you haven't offered anything that alters my trust in other people. You've only succeeded in getting yourself dismissed as a knee-jerk crank.

      In all the discussion over specs, you insist how they help you, but without evidence. I quote the evidence from the specs that prove you are misstating things.

      1) I can prove it to the same extent that you can disprove it.

      At which point it comes down to trust of other people's experience.

      I think my level of trust is based on considerably more investigation, understanding, and thought than your level of mistrust. Which appears to be closer to knee-jerk.

      My level of trust isn't absolute though.

      2) The ITU doesn't clearly state anything about the limitations.
      It states there *are* some, rather like VDSL2 itself states that there are limitations for its use.

      But the very limited statement in the spec (that it is of "particular benefit" out to 1km, a distance that is of benefit to 89% of the UK) is *nothing* like as detailed as the restrictions that you are inventing, and touting as absolute fact.

      There is no statement anywhere in the specs that vectoring stops being useful at *any* distance. Just the inference that it becomes less than of "particular benefit".

      The spec notes that vectoring works with limited background noise, but doesn't state those limits. They may be the same level of noise that would prevent DSL from working at all (so no practical restriction at all), and they may be less than the noise you get from inter-stellar particles from the big bang (so no point in performing vectoring, ever).

      Clearly: The. spec. just. doesnt. help. at. all.

      I'm sure there *are* limitations. I look for those limitations in the papers I read. I don't rate your *unfounded* numbers as being close to the evidence I would seek to alter my trust levels in any of the other evidence.

      3) As (2).

      Except there is also no statement in any spec anywhere that DLM alone does a better job than vectoring with DLM.

      Clearly: The spec does not back you up.

      4) But that's a prior paragraph. Your statement at this point makes no reference to whether you have chosen to apply it to poor lines, good lines, or either.

      But still, poor line or otherwise, there is nothing anywhere to suggest that vectoring degrades performance.

      5) I think we've sufficiently debunked your beliefs on line lengths, and the internet conspiracy theories surrounding them earlier.

      Mission? I am primarily seeking evidence about the use and potential of this technology. I am keen to get decent, well thought-out information out about how vectoring will be of benefit in the real world (not the classrooms, not the labs, not commercial over-statement, and not over-monitored trials).

      But when I come across someone who is keen to dismiss the technology with unfounded data, and does so by masking that data as facts, I do find that I have gained a mission.

      That mission isn't to prove them wrong, or to get them to change their mind – I am well aware there is no point in that.

      My mission is merely to make sure that no-one else believes their tat, and finds the real information for themselves. To give them the sources to let them make their own minds up.

      6) I'll only start to ignore you when you stop attempting to push uncorroborated nonsense as facts.

      You can sit in front of as many lines as you wish, but you are wrong on the theory.

      Theory says that your line *can* be worse than a longer one. It doesn't depend on you, the quality of your equipment, the quality of the copper or the quality of the joints. It just depends on the way individual twisted pairs sit within the bundle of 100+ pairs in the ground, and which of those pairs are also subscribers with VDSL2 turned on.

      That is what makes vectoring both fantastically useful, but fantastically complicated. Both in theory and practice.

      And while you insist that your line *must* be better than a longer one, as if length is the only factor, you continue to prove to the world at large that you aren't qualified to talk about anything more complicated.

      7) I'm not just of the belief that 100-120 is achievable, but that it is achievable out to quite a long distance.

      If I believe that to be true, then it just so happens that, theoretically 170Mbps is plausible to some (very short) distance. After all Gigabit ethernet already gets 250Mbps down 1 twisted pair at 100 metre distances.

      It also just so happens that the graph that best shows my belief on 100Mbps distances for vectored, and best matches current practical experiences of variability of non-vectored, and best matches the way BT set their current estimates *also* shows the stupidly high possibilities at short distances.

      I don't discount those possibilities. I also don't believe that the extremes will happen in the real world. I also don't believe that BT will attempt to set a headline speed, and market at that speed. Frankly, I don't care about 170Mbps.

      In fact, I don't mention 170Mbps. You do, as a way to dismiss my post. And I respond to your accusations.

      Dates) Unfortunately, the date that a graph is produced on is significant indeed. Beyond belief.

      The physics behind the communications theory may not change, but our understanding of the theory can always improve over time (which is what the academics do).

      Our choices over turning theory into practice alter all the time – which is why specs get updates, or why new specs get written – such as G.fast, taking copper speeds to over 400Mbps. The ITU authors, from commercial company research arms, are doing this all the time.

      Our implementation of the both the theory, and of the specifications, can change over time. Our programming of DSPs improves, and the capabilities of those DSPs improves. Our ability to make the electronics quieter improves our ability to separate signal from noise, and all these improvements happen to the chipsets even when the theory and spec haven't changes. Our ability to scale this to ever-larger systems improves all the time. The engineers developing real systems are working at this all the time.

      Those researchers and engineers that try this out in the lab start with existing equipment – such as a stack of VDSL2 modems that were only ever built to cope with 100Mbps. As the lab tests go on, their equipment gets updated with those newer chipsets, with newer DSP programming. Their results improve over time.

      Those telcos (like BT) who have sufficient resources (such as Martlesham Heath) can try this out in circumstances close to reality. Their results feed back into all the previous stages, and make the implementation better.

      The end result is that practical, realistic results improve. All the time. Therefore graphs of real practical results are going to improve over time. Always.

      If there is no other graph that demonstrates this (plain, simple) fact, it is figure 1 in here:
      http://www2.alcatel-lucent.com/techzine/vdsl2-vectoring-delivers-on-its-promise/

      ADSL2+)

      You are comparing apples with pears here. You are mixing up marketing statements with technological capability, and throwing in public expectation to stir it up.

      The first ADSL2+ systems were sold as "up to 24Mbps" when there were zero people who could actually achieve the top speed. But that "up to" speed was the absolute top speed of that spec. It can't go any faster than that, ever – no matter how much we improve the copper, remove the noise, never.

      Even if you added vectoring to ADSL2+, it just can't go faster than 24Mbps. More people might get higher speeds, but it still couldn't beat that number.

      [ADSL2+: 480 downstream tones, 14 bits per tone (a symbol), 4000 symbols per second: Absolute theoretical top speed = 26.8Mbps before removal of guard or pilot tones, and before allocating any of the bitstream for framing or control information, and certainly before accounting for any noise or attenuation]

      It was a marketing failure, not a technology failure. It was a failure to deal realistically with expectations.

      VDSL2 on the other hand has been sold at a more reasonable headline rate, well within the capability of a lot more lines. That automatically means it deals better with people's expectations.

      It was first sold at 40Mbps on 7.05 MHz, but some people could have gone higher if BT chose to uncap it. Obviously 40Mbps wasn't the technological top speed of VDSL2 that 24Mbps was for ADSL2+.

      It was sold second at 80Mbps on 17MHz, but some people could have gone higher if BT chose to uncap it. Obviously 80Mbps wasn't the technological top speed of VDSL2 that 24Mbps was for ADSL2+.

      It will be sold at 100Mbps (or more) on 17MHz with vectoring. Again, this won't be the technological top speed of VDSL2 either.

      It could be sold at 150Mbps (or more) on 30a. Again, this won't be the technological top speed of VDSL2 either.

      So what is the real top speed that VDSL2 can actually go at, if presented with ideal copper conditions?

      From Ericsson I can get the theoretical top speeds that the VDSL2 protocol can handle in the absence of control, framing, noise or attenuation. Obviously that is a speed we'll not see in the field.
      http://www.ericsson.com/ericsson/corpinfo/publications/review/technology_update/archive/2009/issue_2/articles/vsdl2.shtml

      This tells us:
      "This gives the following total system capacity: 4000 DMT symbols/s * 4095 tones/symbol * 15 bits/tone = 246Mbps to be shared between the two directions."

      [I calculate that the 998 band plan used here allocates 2916 tones to downstream, so the downstream direction would be 4000 * 2916 * 15 = 175Mbps. Other bandplans could be more or less.]

      It also tells us:
      "In the same way, the VDSL2 profile for 30MHz (30a) has an effective symbol rate of 8kHz using 3478 tones, which gives 417Mbps."

      We don't have a band plan for the UK, so can't calculate the downstream portion.

      The apples-and-pears difference for VDSL2 is that the marketers stopped trying to sell the technology using an "up-to-the-theoretical-best-ever" speed.

    • Dean

      Your link or should i play along with you and say New_Londoners link (I guess you had to use that ID as i made the FibreFred one look dumb with his Deduction accusations) proves nothing about line length. That report to use their own words is nothing more than “THEORETICAL” and “ASSUMPTIONS”.

      Also BT wires are not 0.5mm. Standard wire gauges around that size are. 0.455mm (AKA AWG25) 0.511mm (AKA AWG24) 0.573mm (AKA AWG23).

      What BT actually use externally (IE to cabinets, from poles to your home etc) is 0.405mm (AKA AWG26) wire. Less resistance, common size, ability to shove more pairs in a smaller place, rated at a nice round even 20amp. And exactly the same gauge they use internally at an exchange to connect things and shock horror is also the standard gauge network cable is in general today.

      Years ago (back in the Post Offic days) cable was Paper wrapped 0.723mm stuff or the equivalent to AWG21. They had to shove more power down it back then. Ah how technology has progressed.

      INTERNAL cabling which is what you are thinking about is called CW1308 and that does indeed have a 0.5mm diameter. That stuff is not used externally. The pairs are not twisted enough to prevent interference. In fact a BT internal extension AKA Home wiring solution for FTTC wont use this stuff either, that also uses AWG26 twisted pair.

      The rest of your nonsense i am not going to address further as you have now resorted to name calling with the “You’ve only succeeded in getting yourself dismissed as a knee-jerk crank.” Remark.

      I am shocked. No matter which name you use insults are it seems what you rely upon, be it accusing people of being other posters that have previously made you look a complete moron. Or direct insulting. At least grow up and have the nerve to use a singular name on here.

      Anyone else still reading now knows why as i pointed out why im a long time reader but a rare contributor, because of people like you that splatter sites like this with putrid flem.

    • MikeW

      You’re paranoid that I’m someone else? That I need to make it appear that someone agrees with me?

      It is nice to get recognition from others, but it isn’t sought. I’m quite capable of standing by my own opinions. And quite capable of admitting when I’m wrong. Read on – you’ll see one.

      Line Length Document:
      You have this thing about assumptions and theory, like you’re sure they’re absolute rubbish. You discard both theory and assumption, even if it is 99% correct, and that degree of correctness is sufficient for most purposes.

      Ofcom are quite happy with these particular assumptions, both in determining the market for fibre services, and in determining whether price ratios for WLR are correct. No doubt BT is too, otherwise the final document would reflect something different.

      I think it is fair to use the line length document as the basis for initial discussion. If you have a specific problem with it, if it shows something you know to be incorrect, then do come back and let us know what the issue is.

      So, back to the document itself…

      Actually, the document doesn’t use theory, but it does make one assumption.

      It starts with measurements of attenuation, and then turns that back into distances based on the information in BT’s SIN 351 about line characteristics.

      So the distance is based on an assumption.

      Thankfully, however, VDSL2 modem’s don’t care about the distance either. All they care about is the attenuation of the signal.

      So, if the incorrect assumption has been made about how to convert attenuation into distance, it really doesn’t matter. The modem will still act like the modem should for the distance we think, because it is seeing that attenuation value.

      The Analysis Mason document covers this in detail. Look there for more of a critique.

      Wire sizes:

      I’ll ignore the nitpicking of the American Wire Gauge sizes. If we’re going to argue about whether the access network is 0.4mm vs 0.5mm, I think we can safely say that the argument covers AWG 26 vs AWG 24.

      Internal CW1308 cable is most certainly not what I am thinking about. I am distinctly talking about the entirety of the access network, from pressurised E-side, through D-side, to drop wire. For FTTC, obviously only the D-side and drop wires really matter, plus the tie cables. The internal stuff is irrelevent when the filter is placed at the end of the drop wire.

      If I’m wrong on the size, then so is the BT specification for the UK access network, BT SIN 351, specifically quotes this: “The range of local lines can be represented by between 0 km and 9 km of 0.5mm copper cable with nominal characteristics of 168 Ω/km and 50 nF/km (attenuation at 1600 Hz of 1.7 dB/km)”

      Aren’t those the characteristics of AWG24 wire?
      AWG26 would be more like 268Ω/km or so, a considerable difference.

      CW1308? I’m aware of its vagaries for carrying data compared with the likes of Cat 5e, and agree that it is sold as 0.5mm cable.

      But all I can find as cable sold as drop wire is stuff that is 0.5mm or thicker, but never less. The likes of CW1378, CW1411 or CW1417 seem to match what I see in use today.
      http://www.caledonian-cables.co.uk/telephone/Indoor/CW1406.htm

      But as you point out, old standards were different, and 80 year old E-side cable is likely to be made to those standards. It is highly likely that BT’s “model” for our lines is an average, and that our access network is made by connecting wires of different size and impedence. If some of a line is to AWG21, then to make the average, some of it must be thinner than AWG24.

      If current standards are 0.5mm, and older standards varied, then the thing to wonder is how much of the installed base meets which standard. And that, obviously is going to lead to some variation in the results we see.

      I guess the same goes for where aluminium has been placed in the D-side.

      Name Calling and Nonsense:

      I’m glad you won’t address any more of my writing, as it means I won’t have to refute it again.

      The name calling?

      Well, when you reject theory and assumption merely because it is a theory or an assumption, without regard to accuracy, your reaction can indeed be considered to be a knee-jerk reaction: One taken without due consideration or thought. A reaction stated, but without a matching justification.

      When you apply the problems from poor marketing choices of ADSL2+, and automatically apply them to the technical issues of VDSL2, your mis-application can be considered as a knee-jerk reaction. Again, one taken without due regard and consideration of the facts. An opinion stated, but without a matching justification.

      When you espouse a theory about how bad things are going to be without any facts to back it up, you might be considered overly pessimistic. It is another sign of a knee-jerk reaction to previous experiences. An opinion, but without a matching justification.

      I’m afraid all I can see in your reactions here are precisely that – kneejerk reactions to previous experience, particularly with ADSL2+, with little to no justification.

      You largely aren’t responding with any logic, with any reasoning, or with any rebutting arguments. You won’t provide quotes that support your statements, and barely provide links (which don’t, themselves support either). You dismiss the things you don’t like as theory or assumption (your use of capitals gives you away when being so dismissive, as though it instantly proves the issue), without stating why that theory or assumption shouldn’t be used.

      To me, a crank is someone behaving without reason, without logic, and without engaging. Someone who dismisses without reasonable reason. I have many worse words that I could have used, but no nicer ones.

      So yes – you got called the most polite name that I could think of, alongside an adjective that fundamentally describes why.

      At least you reacted to that, rather than the way you just dismissed all the other discussions. That is an improvement.

      And I’ll confirm. I have no other names on other posts across this site. I have no alter-egos to stroke myself with, and Mark will be able to confirm the consistency of that (as far as anything can be confirmed by IP logging).

      Any other posts accusing you of something, to cause you to stick a label of “complete moron” on, are down to others. The “putrid phlegm” that you talk about is from someone else.

      I’m not here to call names (though I fall into that trap if I get exasperated by someone who won’t engage properly). I *am* here to inform and improve the debate on a serious subject, and learn from other’s experience. I don’t write often, and usually reserve it for the more complex of technical issues.

      It is interesting that many posters don’t have a photo attached. While mine *could* be someone else, it isn’t.

  14. zemadeiran

    IMHO,

    Copper in the last mile is on it’s last legs and even BT themselves have acknowledged this by moving forward with FTTPoD.

    It will take a little time but as more and more people opt for a fibre connection due to it’s future proofing and lack of distance based issue’s, BT will be able to reduce install costs if there is someone already connected to the Aggregation node surely?

    • MikeW

      Agreed – so long as “on its last legs” we can agree that copper probably has another decade or two at carrying data. Maybe 5 years for business, 5-10 years for heavy residential data users and 10-20 for light residential users.

      FTTPoD is a brilliant strategy for making FTTP into a user-pays entity, while leveraging the backbone installed for FTTC – which itself gives BT the 10-20 years to make money from FTTPoD and to perform the build gradually (which is to say, at the same speed as Australia’s NBN anyway).

      But I don’t see them reducing the install cost significantly. The costs incurred by BT for the first user will be much higher than actually paid by that user – in essence the first user(s) will be getting a subsidised install. As more users come onto that segment, their fees will go to offset those initial costs.

      Some point later, then yes – the install cost will go down. But I think BT will want to milk the early adopters.

      Eventually, BT will start converting whole sections to fibre, and abandoning the copper. The FOX exchange trial shows this – and is probably tied to the idea that the smaller exchange buildings will no longer be needed to hold either telephony equipment or broadband equipment. It can all be somewhere 20-30km away.

  15. Robert Waller

    Self install is feasible for properties that already have a NTE5 master, if they don’t then an Engineer install must be an option as microfilters are not really good enough for VDSL, given that they could reduce the speed. CPs could provide the neceessary iPlate to replace the normal customer access on the NTE5 if needed.

    I am already using my own router/modem equipment and find it much better than the supplied HG612 modem + HH3 provided by BT, I have never lost a synch since it was installed, but am wary of the Vectoring, as to whether this will affect my equipment, as I am not sure if it is compatible.

    Bob…..Infinity 2 w Fritz! 7390 (79.9/20 Mbits)

  16. Dean

    @Robert

    If your equipment has a manual you may be able to check if its compatible with VDSL2 vectored.

    Current FTTC (VDSL2) is i believe referred to as G.993.2

    FTTC (VDSL2) with vectoring i believe is referred to as G.993.5

    If the documentation for your equipment is good it should mention standards like that somewhere :)

  17. DTMark

    I can just see it now. FTTC arrives, sign up, estimate 20Meg or whatever.

    Plug the kit into the socket and it manages say 6Meg with almost no upstream at all.

    Ring up the ISP to get what you ordered: “is it plugged into the master socket?”

    “Yes, it’s plugged into one of the two master sockets”.

    “There will only be one – the one with the stripe across it”

    “No, there are two of them and neither of them have a stripe on them, and only one of them works at all for DSL anyway”

    “Have you checked outside?”

    “No. I know where the wiring goes. I shall describe it to you if you like”

    “Well, we can’t help you as your line doesn’t have a master socket.”

    “It’s not my line, it’s your line. What you seem to be saying is that you’ve installed a new service with obsolete equipment that was never going to perform, and you have wasted your time and mine. I suppose you’d like me to let you off the estimated speed, and not to cancel this outright now and keep on paying you for this poor service for another 11 months for nothing”

    At which point does the ISP have the option – a bit like SFI, call it “OEI” (Obsolete Equipment Investigations) to pay BT to swap out the socketry because they’ll lose even more money if I cancel instead? If so, very clever.

    • Dean

      Careful DTMARK you will have a life story chucked at you by a singular individual with multiple names and accusations you are the poster Deduction as you have not taken a 100% positive BT line with your posting.

    • FibreFred

      No, we know who DTMark is… (and you)

      Plus I agree with what he is saying, makes perfect sense. I can see a lot of issues coming from self installs.

    • DTMark

      I believe I am right in saying that I have always posted under this name, that I think that my opinions have been pretty well consistent throughout my time on here, and I can certainly confirm that I do not have a dual identity except that my forum name is Mark2 and the only reason for that is that I forgot my password and I couldn’t read the CAPCHA code to retrieve it. I have no connection to any ISP except that my partner works for Virgin Media supply chain ten miles away despite the fact that we can’t get cable here :(

      FibreFred strikes me as one of the “old school” BT engineers who still wants to see a good job done and a quality service provided. Whether that is factually true or not, I have no idea. We can all guess ;)

    • MikeW

      @DTMark

      Agree with you totally here (horror though this might be to a certain other poster), except I couldn’t work out the reference to the socket with the stripe.

      BT will, I’m sure, have their extra “Please install my FTTC modem like I really should have ordered in the first place” product lined up, slightly different from a traditional fault visit.

      Who pays? It won’t be BT, and it’ll probably be the end-user. The ISPs aren’t going to want to cover it, even if doing so would make them look nice.

      I expect there to be some adjustments to the ISP-level fault wizards that push people into choosing this as a paid option.

      Will there be more errors than the current engineer installs?

      There certainly will be more problems than, say, the installs we were seeing 6 months ago. And different ISP or broadband self-help forums will have to take up the slack.

      Since then we seem to be getting a lot of reports of botched installs, where the installation hasn’t even included a faceplate, or used the master, or some other simple fault. Whether these are due to new Openreach staff, or ill-trained contract staff, doesn’t really matter. It acts as a warning for the kind of problems we’re going to see.

    • Dean

      “FibreFred strikes me as one of the “old school” BT engineers who still wants to see a good job done and a quality service provided. Whether that is factually true or not, I have no idea. We can all guess”

      Strikes me more like an immature person of low IQ probably in or around his 20s stuck at home all day claiming benefit given the amount of hours in a day he appears on here. One of scourges of society we have to deal with today.

      He still thinks i am a poster named deduction even though im a BT user and have been for years (another speed test this evening for him to show it was not a one off)
      http://www.speedtest.net/result/2663421614.png

      Looking back at his arguments with Deduction, i doubt would ever touch BT with a barge pole, they really do seem to hate them. Also if you look back you will see a suspicious trend of names that always agree with fred, while everyone else just seems to think he is talking nonsense. Stupid often agrees with stupid ;) New_londoner as an example has from what i can see never ever commented in any news item by himself, or had a different opinion to that of Fibrefred.

      In the past month just looking back at a few news items that have a lot of comments (over 10 per news item) Fibrefred he has accused 4 people just from what i have noticed of being deduction. Clearly he has a serious issue with that user.

      An older, mature person would have got over very quickly. A young, immature and i believe the kids call it butt hurt person of low IQ though would just would not be able to get let it go though. It is pretty clear which of the two he is ;) It is funny, well as funny as sad, lonely, bored and jobless gets.

    • Dean

      To further illustrate how he persistently tries show he is correct but ultimately shows himself to be a fool, you only have to see the last exchange of posts here between him and myself.

      I explain external cable and to re-quote myself…
      “INTERNAL cabling which is what you are thinking about is called CW1308 and that does indeed have a 0.5mm diameter. That stuff is not used externally.”

      He then insists external cable is still 0.5mm with this statement…..
      “But all I can find as cable sold as drop wire is stuff that is 0.5mm or thicker, but never less. The likes of CW1378, CW1411 or CW1417 seem to match what I see in use today.
      http://www.caledonian-cables.co.uk/telephone/Indoor/CW1406.htm

      No i know he is a little dimwitted (no matter which name it uses) but surely anyone would notice the word “INDOOR” in his so called link evidence was a bit of a give away that is not for external use.

      Unfortunately for him unlike deduction who he thinks everyone is i will obviously not be addressing him further. Unlike the many faces of new_londoner/fibrefred i realise it is a fruitless task trying to educate an individual when they do not even know what the word indoor means.

      The cable he linked to and the insulator used in some of the types on that page actually breaks UK fire regs, so i would hope BT or anyone with any sense would not use it in this country at all.

    • DTMark

      “Dean” – this is just a suggestion:

      I suspect that a key reason that people may suspect you are another poster is that you have the same (in my opinion) slightly puerile but recognisible habit of attacking the person as opposed to the issue.

      So whether or not you are the same person – and brevity is probably not my strongest suit – it may perhaps pay to focus just a little less on the person and more on the topic.

      And, separately but equally, I continue to believe it to be somewhat disingenous for some people obviously employed by certain companies not to disclose this fact publicly and state their position and capacity. However I also realise the politics involved and that in particular, some ISPs playing their subservient part in the vertical monopoly may not wish to publicly bite at the proverbial hand which feeds.

    • keith

      Nah Dean i would say is 100% right, Fred has also on more than one occasion accused me of being Deduction.

      Why he thinks many individuals are the same person only he knows. I have previously demonstrated to him who my ISP is and Dean has also demonstrated who his ISP is, yet the cretin thinks we are the same individual.

      Deduction thought BT and those with any ties to them were scumbags, if Fred is a BT engineer or employee of any type i would have to admit that point of view holds some water. Even more bizarre is if Deduction and Dean were the same individual then it boggles even the greatest minds in history why anyone would think he has BT as his ISP, if he hates them so much.

      Oh and no i am not Deduction or Dean, before the troll of many names says i am. I am far too old to be up at 3.30am in the morning as Dean was and still be awake at 10am.

    • FibreFred

      “Why he thinks many individuals are the same person only he knows.”

      Probably because he uses so many multiple id’s? Anyway, I know better than to feed.

      I also find it amusing that if you post any comments in support of BT you automatically become an employee and/or have shares.

    • keith

      Take that up with DTMark as it was him that suggested you work/worked for BT.

      I personally think the other posters assumptions in your latest little escapade to turn the thread into your own personal Special Olympics are far more accurate. Maybe that is half the trouble with internet in this country its so cheap even the layabouts like you can even afford it.

      I have to go now. Some of us only have a 30 minute lunch break and not all day like yourself to talk crap.

    • MikeW

      @DTMark – Sorry Dean seems to have stolen your thread here.

      @Dean

      Ah, Dean, you do make me laugh.

      A company makes a mistake with their URL, and lists an outdoor cable with the word “indoor” within the URL? And you base a whole rant on it?

      You missed an opportunity there – the URL only indicates CW1406 wire, but the content has the temerity to cover 4 different wires. You could have ranted about that too.

      And once you have decided it *must* be indoor cable, you decide it is a fire hazard.

      Take a look at the content of the page rather than the URL, and what do you see?

      Well a big clue will be the high tensile steel strengtheners. Come across that much in indoor cable, do you?

      Did you read the “Application” text? As you appear incapable of following links and investigating issues fully for yourself, I’ll copy it here. Note the use of the words “pole” and “building”. They’re a bit of a clue.

      “Applications
      The drop wires are designed for use in local area network to link between pole and building or two buildings. The high tensile steel strength member wires are incorporated inside the cable to provide support. The cables can be used for crossing high voltage cables up to 11kV.”

      Here, have some more links to 0.5mm drop wires, that match the same specs.
      http://www.fscables.com/Signal++Control/Telephone+Cables/CW1411++CW1417+BT+Drop+Wire+Cable/list.htm
      http://www.addison-cables.com/product/Telephone%20Cables/Drop%20Wires%20No.%2012,%2010B,%20CAD55M%20&%2010%20to%20CW%201406,%20CW%201411,%20CW%201417%20&%20CW%201378.htm

      I think, on balance, it is a relief to know you won’t be talking to me any more. My day got just a little brighter.

    • keith

      “A company makes a mistake with their URL, and lists an outdoor cable with the word “indoor” within the URL? And you base a whole rant on it?

      You missed an opportunity there – the URL only indicates CW1406 wire, but the content has the temerity to cover 4 different wires. You could have ranted about that too.

      And once you have decided it *must* be indoor cable, you decide it is a fire hazard.

      Take a look at the content of the page rather than the URL, and what do you see?”

      From what i see he is correct. I see right at the top of the page it is listed as INDOOR CABLE in the INDOOR CABLE section.

      Is that what you mean by “content of the page”?

      My god it does not matter what name you use it appears you are not only dim as dishwater but now also blind.

      Im not going to comment on the Fire thing i can not be bothered to go and see if you are wrong on that also.

    • MikeW

      et tu keith?

      “From what i see he is correct. I see right at the top of the page it is listed as INDOOR CABLE in the INDOOR CABLE section.

      The only mention of “Indoor” on the content of the page is within the breadcrumbs at the top – highlighting the structure of the website. It reflects directly what is in the URL.

      Care to comment on the part of the page that reflects the function of the cable, rather than the structure of the website? This is a site for telecomms news, not for web design critique.

      Care to comment on the cable in the other examples?

      Care to use google, and come up with an example of BT’s drop wire 10b (aka CW1411) that isn’t 0.5mm?

      “My god it does not matter what name you use it appears you are not only dim as dishwater but now also blind.”

      LOL. My apparent dimness is but a reflection of the mental prowess and genius arrayed against me.

      I’ve changed my mind. I’m enjoying this!

    • keith

      “The only mention of “Indoor” on the content of the page is within the breadcrumbs at the top – highlighting the structure of the website. It reflects directly what is in the URL.”

      Which is web page content, which says indoor does it not? Which makes you an idiot which did not look before you commented again does it not?

      “Care to comment on the part of the page that reflects the function of the cable, rather than the structure of the website? This is a site for telecomms news, not for web design critique.”

      Sure, it says its for pole to home yet is listed in an INDOOR cabling section, which makes relying on their information being correct just like information you provide a fools belief.

      “Care to use google, and come up with an example of BT’s drop wire 10b (aka CW1411) that isn’t 0.5mm?”

      Not particularly as obviously that is not the same cable as what is used in cabinets. Unless you are saying there are hundreds of only few pair cable running into a cabinet? Rather than only a couple of cables of hundreds of pairs? Not the same wire in the ground back to cabinets either, given that is longer that will have far more baring on FTTC speed (of any variant) than any drop wire which is only a couple of metres in length. As already explained to you some underground stuff is so old its paper wrapped. What do you think every homes drop wire is continuous back to cabinet/exchange and at no point feeds into a bigger bundle of pairs or connects to another cable? Yeah BT lay hundreds of SEPARATE cables from each road underground back to a cabinet rather than it being fed from the pole into a bigger singular cable of hundreds of pairs. Clueless does not do you justice.

      “LOL. My apparent dimness is but a reflection of the mental prowess and genius arrayed against me.”

      Is that your “though” process at work AGAIN?

      Anyway another day another dollar, work over and home time for me. Good to see you still stuck at home all day still.

    • MikeW

      Keith,

      Pausing for breath is a good idea, as noted by New_Londoner. A very good recommendation.

      But please remember that way back in this conversation I acknowledged that, whatever standards are *now*, they didn’t always apply. That cable placed underground years ago met whatever standard applied when it was installed. And that we couldn’t therefore know everything about the current set of cables.

      What I said, back wherever, was that BT currently state a model of 0.5mm, and that if there are some wires larger than this, then for 0.5mm to be an average, there must be some wires smaller too.

      So having said that…

      The ongoing discussion over drop wires, therefore, isn’t about its size, or how far back to the cabinet it goes (which could be anywhere between 99%). The argument left was about whether the British Telecom specs CWxxxx (that happened to be for drop wires 10, 10b and 12) was for indoor or outdoor locations.

      Continue to ridicule me for believing drop wire is for outdoor work if you must.

      As for the aspect I am apparently clueless on – the architecture of the D-side. I’m not sure what I’ve said that makes you think I believe BT run drop wire all the way to the cabinet? Examples I’ve used are merely that – examples.

      As for the paper-insulated cables. Yes I’m aware of these. Didn’t Dean mention them (and mention they could be AWG21)?

      I know they exist extensively in the network, and that they’re old. I’ve heard of one that dates back to before 1900.

      I’m also aware that they need very good protection from water intrusion (rather obviously), and that BT use compressed air pumped into them from the exchange building to help that protection, with manifolds at the PCP.

      For that reason, I’ve always tended to believe that they only exist (or now remain) on the E-side, as the larger cables there, up to 4800 pairs.

      The D-side, to my understanding, is mostly wired in cables with 10′s to low-100′s of pairs. However, the D-side ends up with a lot of underground joints to DP’s, where the drop wires lead from.

      The likelihood of water ingress is therefore going to be much higher in the D-side, and the opportunity to protect the cable with compressed air is much lower. I can’t say it is impossible that there are paper cables out in the D-side, but if there is, it is likely to be at considerably more risk of damage and total failure. I doubt there is a significant amount now, but I am prepared to be surprised.

      If, however, there is a significant proportion of that older, fatter, paper-covered copper in the D-side, we should probably rejoice. That means lower attenuation, better speeds, and better vectoring for those customers.

      Do you know of any paper-insulated cable in the D-side?

    • keith

      Oh you still here after being wrong about the vectored products speeds and now having no clue about cabling infrastructure either? What was you saying about some users being some type of crank?

      Please reply in essay form again, im going to bed, though as is obvious you have probably been waiting for a response all day.

  18. FibreFred

    It will be interesting to see how this goes. I expect some ISP’s will start to offer the self service to see how it fairs, if its more trouble than is worth (costs more in follow up visits due to complaints regarding speed stability) I expect they’ll pull it.

    • cyclope

      @FibreFred: Not being funny here, but why is there going to be a load of issues from self installs? necessary work @ end users home, isn’t hard,

      A filtered faceplate ,and a modem, oh and if the openreach bod actually brings the data extension kit (if pre ordered by isp) & the isp could simply ask the customer if they felt comfortable about installing their selves or they could supply step by step detailed instructions ,

      I don’t see there being any massive increase in faults relating directly to self installs, Especially having seen and heard how even some openreach staff go about the managed install,Are they being paid by piecework? and why are they not provided with the right equipment by openreach to do their job properly instead of cutting corners instead

    • FibreFred

      I didn’t say loads of issues I said it will be interesting to see how it goes. I could have done what the engineer did for me no probs, I wonder how many would bother to wall mount the modem though

  19. TheFacts

    Vectoring trial to start in Barnet and Briantree.

  20. MikeW

    Just seen a quote from the DSL Conference, attributed to BT – presumably Kevin Foster who was due to speak:

    “VDSL Vectoring is seen as a speed enabler rather than a speed booster”

    I take from that, that they mean that it won’t increase the headline speed much, but it will get more people (who have slower speeds than the headline) up nearer (or to) that headline.

    That could indicate that BT intend to use it to turn down the power for shorter lines rather than boost their speeds, with the aim of allowing more “noise-free” spectrum for remote users.

    Perhaps we won’t have a headline speed change at all.

  21. New_Londoner

    Polite Request to All

    Could we all agree that there may in fact be more than one person on the planet that disagrees with our respective views, resist accusing all those that disagree of being the same other person. For the record I have never posted under any of the (many other) names that have been suggested as part of an extensive multi-id ploy – I have neither the time nor inclination, suspect others don’t either.

    Also, it would be really good if we could resist referring to others in derogatory, sometimes downright hostile and insulting terms. Whilst banter is fair enough, I’ve had the misfortune in the past to have posts directed towards me that were borderline criminal and certainly more than a little unpleasant. Whilst MarkJ removed the offending material very promptly (thank you!), people using this site should be sufficiently articulate to be able to argue a point without resorting to such puerile, offensive language.

    Back on topic!
    The debate around vectoring should be resolved soon enough anyway, with the planned UK technical trials. So we will find out who is right, can probably avoid another 100 postings arguing the point now. Phew!

    • Ignitionnet

      You are aware that ‘some’ people apparently think you’re Liv Garfield, right N_L?

    • New_Londoner

      @Ignitionnet
      Fantastic, news to me!

      For the record, wrong sex and wrong company, apart from that …. :)

    • FibreFred

      That would also mean I am Liv as well I presume? Seen as I account for most posters on here according to one ;)

  22. Bob

    There is a lot of misunderstanding of vectoring and the benefits from it. The reality is that for most it will make little difference. There is also a bandwidth overhead with vectoring. Copper4 lines are already being used pretty much at there limit. I suspect that vectoring may not even be deployed. The cost benefit will be to small some users will also be adversely affected by Vectoring

    At the moment take up of HS broadband is quite low but once usage becomes high the problems & limitation will become apparent. Already in areas with with take up reports of speed dropping significantly are common

  23. Telecom Engineer

    Vectoring will allow those lines with significant drops to regain their initial speeds. For example my line dropped max attainable 50 meg due to cross talk. Luckily at under 400m of copper that still leaves me 90 meg.
    It appears BT are looking at vectoring more as a stability measure than an upgrade purely for headline speeds, combined with experiance from digital regions offerings I suspect we could see a simple lifting of profiles to allow the 100meg product without vectoring being a prerequisit. There are significant technical challanges introducing vectoring into non vector ready dslams so it would be nice to know what vendor and model the trail dslam is. Seems to me we are years off seeing nationally this if at all. Shame BT didnt opt for Alcatel dslams with vectoring a simple software upgrade from the start, as we see with the ongoin adsl 2 rollout they never rush to replace working kit despite its technical inferiority to competition.
    I have seen 80meg on the 12mhz profile so there must be plenty of headroom on short lines under 17mhz. However without vectoring to increase coverage launching such a product may cause too many complaints…. perhaps a remapping of current products i.e. a 50m and 100m to replace 40 – 80?. At least with a free uplift they can market and get publicity without a huge backlash? Just a thought..

    • MikeW

      Someone mentioned (not sure which forum) that the rollout of vectoring (presumably if it goes ahead at all) was due to complete in 2017. That does suggest some amount of hardware switching.

      The 12MHz profile is, in theory, capable of around 100Mbps max, so if you have seen 80, that’s good – it does suggest a (short) 17MHz system would reach 100 easily.

      I can’t help wonder if, with ASA taking an “up to 80″ system down to “up to 76″, Openreach would rather aim at, say, a 105Mbps sync, so it could reasonably market at “up to 100″.

      Having too many speed levels adds confusion to the market – witness the set of FTTP speeds being discontinued at the moment. You might be right that a small uplift is the best, PR-friendly, option.

  24. Telecom Engineer

    Mike W
    Yes there are very small amounts of paper insulated cable in the dside network. Round me I’d say less than 1%. Its a right bugger to work on as it is lead sheathed ( good for keeping out rfi tho). Most was replaced decades ago as it only takes one failed joint and the water quickly tracks down the cable causing massive failure. Engineers no longer have the time and few remain with the skills to read and plumb such joints, so one rushed or bodged joint and its bye bye cable. The vast majority of cable is indeed cat 3 0.5mm copper polly sheathed and greace filled (so wet in joints doesnt track down the cable). The amount of aluminium varies on area but again is normally 0.5mm grease filled polly, although amounts of 0.9 copper and ali exist on long routes. The majority of any ancient cables or thin guage will be exchange to cabinet i.e. eside cables. They are indeed pressurised rather than grease filled as we dont enter the joints unless there is a major breakdown. For fttc it is only the dside we are concerned with. Dside typically is a 100pr cable leaving a cab falling in number from 50 to 10 pair depending on breakout points for DPs. Drop cables typically are 5 to 2 pair for ug residential or single or double for overhead. Business can have any size cable. Talk of how old a cable is is irrelevant. Any dsl line must pass a modern pair quality test which on some respects is up to 5 times the sin 349 standard but engineers by rule tend to do ten times ( i.e. resistance to earth per leg look for 1 meg ohm rather than 100kohm ). Many cables get replaced on a daily basis, there are dedicated teams on each patch purely for this ) and Hundreds of millions are spent each year on high fault nodes without any live faults driving the replacement.

  25. Telecom Engineer

    Mike,
    That was a really interesting article, i’d love to see how long it takes for the fios install. Reading those comments made me lol, very much like what we get here, people saw a paper label (put for conveniance not as the only record) and so an arguement about failure to keep records and the open joints ( where the engineers are obviously drying out joints ) and people assume that they are normally in such a state. In truth there is nothing stopping Verizon replacing the copper, BT do this work when our exchanges have flooded, the difference in that network is they have total monopoly and thus can make the call to rip out the copper and fit fibre. Infact it will be cheaper as no additional duct build is required if you accept major outage. Much easier to splice a few fibres than joint thousands of pairs per cable. Over here LLU operators would go crazy if BT tried that.

  26. stephen Adams

    Just a couple of non-technical points.
    1. The reason that BT want to do customer installalation is that it easy to program in an engineer you own to visit a cabinet. Getting a customer to meet an appointment date is very problematical.
    2. The solution to the distance problem is simple and has been avaialble since System-X was introduced (digital exchange working). You provide an extension to the exchange in the cabinet, supplied by Fibre, which reduces the distance to the customer to less than 100m as the line card is now in the cabinet and not in the BT Exchange.

  27. Nice post. I learn something totally new and challenging on
    blogs I stumbleupon everyday. It will always be helpful to
    read articles from other writers and practice a little something from other websites.

    My homepage: family photo cards

IMPORTANT: Javascript must be enabled to post (most browsers do this automatically). On mobile devices you may need to load the page in 'Desktop' mode to comment.


Comments RSS Feed

* Your comment might NOT appear immediately (the site cache re-syncs periodically) *
* Comments that break site rules, SPAM, TROLL or post via fake IP/anon proxy servers may be blocked *
Promotion
Cheapest Broadband ISPs
Poll
* Javascript must be ON to vote *
The Top 20 Category Tags
New Forum Topics
Promotion

Copyright © 1999 to Present - ISPreview.co.uk - All Rights Reserved (Terms, Privacy and Cookie Policy, Links (.), Website Rules)