Home » ISP News »

Alcatel-Lucent ISP Trials Deliver 130Mbps via VDSL2 Broadband Vectoring

Posted Thursday, January 3rd, 2013 (8:55 am) by Mark Jackson (Score 1,578)
fast broadband uk

Global telecoms firm Alcatel-Lucent last year released the results from its recent ISP trials of VDSL2 Vectoring technology, which revealed that the solution has the potential to push BT’s UK superfast FTTC download speeds up to 130Mbps (Megabits) at around 300m to 600m (metres) from a local street cabinet (possibly hitting 200Mbps with line bonding).

Alcatel-Lucent became the first telecoms technology company to launch a commercial vectoring solution in 2011, which is designed to reduce crosstalk (interference) on copper based VDSL2 telephone cables and can thus open the door to faster broadband speeds over greater distances (i.e. this works in a similar way to the noise cancellation technology used in some headphones).

BT are already understood to be examining the solution as a potential upgrade for their existing Fibre-to-the-Cabinet (FTTC) service, which currently delivers download speeds of up to 80Mbps (20Mbps uploads) by running a fibre optic cable to your local street cabinet (profile 17a). However the “last mile” connection into homes is done by using VDSL2 technology over your existing copper telephone line (similar to ADSL2+ broadband but faster over the shorter distances).

The new trial data appears to support earlier performance estimates for Alcatel-Lucent’s Vectoring solution, which has now been tested with a total of 17 ISPs from around the world (11 of which were European providers). This is useful because the EU expects 50% of European homes to have access to a 100Mbps connection by 2020 and Vectoring could conceivably help some countries to deliver that.

alcatel-lucent vdsl2 fttc vectoring trial results

Alcatel-Lucent’s Trial Report

Seventeen very high speed digital subscriber line 2 (VDSL2) vectoring trials, conducted by Alcatel-Lucent with various [ISPs] around the world, have demonstrated that 100Mbps is achievable over copper at 400 m — and even up to 500 m. These trials focused mostly on 300 m to 600 m loop lengths, reflecting the typical fiber to the node (FTTN) topologies that CSPs are considering for vectoring.

Because trial results improved significantly in a short time with technology and product evolution, Figure 1 uses color coding to indicate when specific downstream bit rates were recorded. Some of the early trials, beginning in July 2010, were capped at 100Mbps due to technology and prototype limitations. But the most recent trials in April 2012 achieved bit rates between 100 Mbps and 130 Mbps.

These improvements are partly explained by the switch from prototypes to commercial hardware and software. In addition, G.inp coding now provides a retransmission-based error-handling mechanism for VDSL2 vectoring which results in significantly lower overhead. By replacing the Forward Error Correction (FEC) mechanism, G.inp avoids the FEC-associated fixed overhead of about 12 percent, and G.inp only kicks in when there’s actually an error to correct.

Cable diameter does have an impact on the results and accounts for most of the variation (between trials performed in the same timeframe) in Figure 1. In general, CSPs can expect to see very good vectoring gains whatever cable type they use. The good news is vectoring delivers greater gain on poorer cable that is heavily impacted by crosstalk, because the crosstalk can be canceled by vectoring.

The Vectoring solution was also able to boost upload speeds from the current 20Mbps figure and up to a typical speed of 40Mbps at loop lengths from 200m to 500m (achieved while download speeds sat at around 100Mbps). This was possible because vectoring allows ISPs to “relax their upstream power back off settings” (UPBO), which boosts upstream bit rates.

Customers with deep pockets and two copper pairs available to their home will also be able to use vectoring and line bonding together (e.g. linking 2 phone lines for faster speeds), which in the trials delivered a download speed of 200Mbps (50Mbps uploads) at 400m (with normal UPBO settings – relaxing this could boost upstream performance even further). At 800m, with vectoring disabled (i.e. only line-bonding was used), this fell to just 75Mbps downstream and 17Mbps upstream.

So what can we expect in the future? BT does plan to boost its FTTC (VDSL2) service speeds up to around 100Mbps or more in the not too distant future, although initially this is likely to come from a more cost effective increase to the relevant spectrum allocation (currently set at 17MHz). By contrast vectoring would carry a more noticeable upgrade cost and BT is currently more focused on expanding its raw FTTC’s coverage.

Meanwhile the ITU-T’s new VDSL2 G.vector (G.993.5) specification has only just completed its first interoperability testing and other firms are already developing their own solutions for boosting DSL speeds (e.g. DSL Rings). In other words it’s still too early to say precisely what solution BT will go for but FTTC performance is clearly destined to increase.

Last March 2012 BT told ISPreview.co.uk, “We believe other technologies such as vectoring could see some FTTC lines deliver 100Mbps or above, though these new technologies will not be applicable on all lines” (here).

NOTE: I actually drafted this article a couple of months back but a date bug meant it was never posted and we all failed to notice because of a busy news week. I’ve updated it and re-posted today.

Delicious
Add to Diigo
Add to Slashdot
Tags:
Leave a Comment
19 Responses
  1. sam

    130mbps would be great, that will last us several years.

    • Toonshorty

      The current broadband average is around 7Mb/s. 130Mb/s is ~20x faster.

      By that logic when everyone was on 56Kb/s then 1Mb/s would last us several years. It didn’t.

  2. DTMark

    A few questions as ever ;)

    1. With longer lines using ADSL, the upgrade to ADSL2+ can actually make them unstable, it’s just a “push too far”. This may be solved by the router renegotiating and connecting via ADSL as before. With FTTC, as the profile changes to increase speed for the shortest lines, what happens to the longer ones – do they “drop back to the previous profile”? Do they just slow down? Or is there no change?

    2. The article mentions cable diameter as having “an impact”. Cable diamater and metal (copper, alu) has a massive bearing on ADSL performance, VDSL even more so. Given that I don’t think there’s a database of metals, cable diameters and so on, how can this be quantified in the “real world” other than by implementing it and seeing what it does?

    3. Am I right in thinking that vectoring is only possible if BT continue to have the monopoly on the end delivery via the street cabinets, so for instance, the monment vectoring is in place, SLU is made impossible?

  3. 1. If you’re talking about vectoring then AL’s suggestion is that would help speeds on longer lines by reducing interference, though I’d like to see how it works within BT’s own environment first (AL don’t offer ISP specific results from their trials).

    2. AL must have some idea about this as they claim that vectoring does help on such lines but didn’t publish any clear data to support that finding.

    3. I wouldn’t say anything is impossible but it usually comes down to a matter of money and or desire. I’m not going to make any predictions until BT confirms precisely what it has in mind for after the next profile change.

    • DTMark

      On point 1, I mean, without vectoring. For instance when the profile went from (17?) to (30?) and any future similar changes in the absence of vectoring.

      Could make a significant impact as if we ended up with FTTC here, most D-sides are “long” on one of the two cabinets, considering only length not quality.

      So if you have user A getting 76Meg just across the road from the cab and user B getting 8 Meg down the road, and the profile changes again, user A gets more, what does user B get?

    • Anoyed tax payer

      User B might get an improvement if the speed loss down to 8mb is due to crosstalk

    • DTMark

      What if it has nothing to do with crosstalk, but is because user B’s D-side is say 1500m long and partly or wholly made of aluminium so delivers a stable (insert number here) meg at the moment.

      If the profile changes so the headline speed improves, what is the real effect on user B’s speed..

      Does it go up, down or stay the same?

    • Anoyed tax payer

      If after having a line length of 1500 meters of aluminium and there is no crosstalk, then that would be a miracle and also would defy the laws of physics.

      Even with a length that long of copper you will get crosstalk.

    • DTMark

      … and without the implemenation of vectoring, which is still only theoretical?

      The point is: ADSL2+ shortened the reach of the solution. But, those beyond its reach could retain the same performance by “dropping back” to ADSL1, still supported.

      So while ADSL2+ wasn’t even pursued in some other countries and dismissed as a waste of time, it did improve speeds for some, and for those it did not, performance did not degrade.

      So if we pursue headline speeds and the profile changes again to the benefit of user A, what happens to user B?

    • Anoyed tax payer

      For user B, the only hope would be FTTP on demand. (That’s if they have the money for it)

    • DTMark

      That’s the conclusion that I reached – no matter what you do at the cab, if there’s much more than 1000m of copper, fibre is needed.

      Which is the argument I’ve made all along since the idea of only going as far as the cabinet was mooted as a solution.

      And which would be why I can’t see how it’s cost effective to pay for cabs to be enabled without analysing them at a street/estate level to see whether or not this will work at that specific location, and then select the correct technology for the goals at each.

      Just as: the fact you’re connected to an ADSL enabled exchange doesn’t mean you can get the service and if you can, it may be too slow to even bother having, like here, I’d hate to think we were ploughing money into another stopgap solution like ADSL at this point with the same resultant patchy style of semi-random availability of usable speeds.

  4. Kyle

    Rather than pushing the limits of those already receiving relatively good speeds for what they signed up for, how about looking at those who are ‘passed’ by FTTC and achieve slower speeds than ADSL2+?!

    The ISPr article of the other day, discussing how to identify a threshold fault has made me want to push BT to do something. How they could ever reach a figure of 26.2/5.6 and my achieving 16.8/0.4 is beyond me. I’m considered to be passed by a superfast service and to be receiving a superfast service. Unlikely.

    • DTMark

      One possibility, which I alluded to above (no database of metals and diameters) is that your D-side is made of aluminium, so the performance might be 40%+ less than if it were copper. But the estimator wouldn’t have known about that. In my post (third one down) point 2 is addressing this.

      The answer, apparently, is for you to pay a grand or two to have the old D-side swapped for some fibre.

  5. Phil

    BT FTTC 130/20 mean that Virgin Media will be a loser soon. Good move by BT.

    • DTMark

      To get those sorts of speeds over FTTC will most likely see the majority with an installation fee of between £1k and £2k.

      What does VM cost to install these days, about £30?

    • Alex Atkin

      How do you figure that out, considering Virgin could do 200Mbit on their current configuration, 400Mbit if they added more channel and have field tested 1.5Gbit for businesses in London.

      Then there is DOCSIS 3.1 which will enable 10Gbit down and 1Gbit up within the next couple of years. There is plenty of potential left for Virgin, as long as they get sort out their core network capacity.

  6. Darren

    My 300M mixed metal d-side flips between 80Mb sync and 70Mb sync with interleaving every few months. If vectoring could give it a bit more head room to stay at 80M sync with interleaving off I would be happy. The small speed drop I can live with but the heavy interleving is annoying. Page loading is noticably slower especially stuff like google maps.

    I’d be pleasantly supprised if BT did anything further to better FTTC speeds, with the availability of FTTP-OD it’s in ther interests not to. At least not untill it had been available for a while and they have a feel for demand. Why improve FTTC if people are willing to spend on FTTP-OD!

    Is the curent hardware capable of supporting profile 30a and vectoring? I’m guesing yes for 30a although modems only have a 100Mb ethernet port, not sure about vectoring though?

    • Anoyed tax payer

      The street cabs need major hardware upgrades in order for vectoring to work, also 30a is not compatible with vectoring. So it’s either 17a with vectoring or just 30a.

      I think what BT will do is upgrade to 30a and the rest will be left to FTTP on demand.

  7. Alex Atkin

    It is my understanding that vectoring has to apply to the entire cable bundle equally, as the whole principle is that it predicts the crosstalk from one pair to another and compensates.

    So this would seem to be a none-starter except for exchanges 100% covered by FTTC and on the condition that all LLU providers will switch to using BTs FTTC cabinets. That means switching off ADSL at the exchange.

    What are the chances of all that happening?

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 Superfast ISPs
  • PlusNet £3.75 (*14.99)
    * Speed: 38Mbps - Unlimited
    * Gift: None
  • BT £5.00 (*15.00)
    * Speed: 38Mbps - 20GB
    * Gift: £80 Sainsburys Voucher
  • Sky Broadband £10.00 (*20.00)
    * Speed: 38Mbps - Unlimited
    * Gift: None
  • Virgin Media £10.00 (*15.50)
    * Speed: 50Mbps - Unlimited (FUP)
    * Gift: None
  • TalkTalk £11.75 (*13.50)
    * Speed: 38Mbps - Unlimited
    * Gift: £120 Love2shop Voucher
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)