Copper: Still a Precious Metal?

For FTTH vendors and zealots, news that Alcatel-Lucent (NYSE: ALU) has lab-trialed an access system that delivered 300-Mbit/s downstream bandwidth over 400 meters on two copper pairs may not be altogether welcome news, suggesting as it does that the age of the all-fiber access network is about to be postponed yet again. Yet from the point of view of consumers, it's potentially very good news – and it might even turn out be good news for those already committed to rolling out FTTH.

The Alcatel-Lucent trial combines a very modern advance known as "vectoring," which is analogous to the technique used in noise-cancellation headphones, with a very old one known as "phantom mode," which in effect turns two bonded pairs into three (the third being the virtual "phantom" pair).

ECI Telecom Ltd. and standardized by the International Telecommunication Union (ITU) as G.vector last month and as profile 17a in the DSL family, is a big improvement even on its own, delivering 100 percent gain over standard VDSL2 in Alcatel-Lucent's trials when it was applied to all pairs in a bundle. Others, including both ECI and Ericsson AB (Nasdaq: ERIC), have demonstrated similar improvements in trials. But vectoring gets a big boost if it's combined with phantom-mode operation. Combined, these techniques should deliver at least 100 Mbit/s on loops up to 1km long.

The experiment used good-quality 0.6mm-gauge copper, though Alcatel-Lucent said it had tried to replicate real-world loop conditions. Field trials planned soon will initially be on Swisscom AG (NYSE: SCM)'s famously well-engineered access loops, and the advertised advances may not be achievable everywhere. But whatever the state of the copper, one thing that shouldn't affect performance is variable cross-talk, the bane of all access networks running DSL (especially VDSL): Cross-talk is almost completely eliminated by vectoring.

There are plenty of caveats. First, carriers would need new CPE, and new linecards with a dedicated chip to handle the heavy-duty signal processing that's associated with real-time noise cancellation. Second, it requires two pairs of wires for each household; Alcatel-Lucent believes that many telcos do have the spare pairs available, but either way it clearly adds to the cost. Third, it's not symmetrical, unlike, say, P2P Ethernet FTTH: Upstream bandwidth will be lower. Fourth, a lot more testing needs to be done on phantom-mode operation to understand its effect on electromagnetic compatibility and performance on a wider variety of cable types. Finally, the network operator has to have control over every pair to gain the full benefits of vectoring – so standard local-loop unbundling, nearly ubiquitous in Europe, wouldn't be possible.

Yet any technology advance that extends the life of the copper network is to be welcomed. Demand for bandwidth will continue to grow inexorably, forcing telcos to upgrade networks and – eventually – replace the copper network altogether. But it takes a long time to build out an all-fiber network, especially in a large country. Experience to date suggests that getting to, say, 50 percent all-fiber households is a 10-year (minimum) project in major incumbent telco networks, and to achieve the kind of reach typical in DSL networks today will likely take longer than that. At our last count, only 2.5 percent of the world's households were connected to any kind of fiber. (See FTTH Review & Five-Year Forecast: The Road to Next-Gen PON.) And with cable MSOs advertising higher-speed offers, telcos need any fill-in technology they can find.

As they come to commercial fruition in 2011 or 2012, such techniques as vectoring and phantom mode will effectively help pave the route to the all-fiber network that will surely replace copper in the long run. By promising a new gold standard of 100-Mbit/s service delivery to many more households, it will buy time to make the transition to fiber – time that the vast majority of wireline access builders will need.

— Graham Finnie, Chief Analyst, Heavy Reading

bitsarebits 12/5/2012 | 4:35:39 PM
re: Copper: Still a Precious Metal?

phantom mode was first proposed in 2003 in a ITU standard by Stanford Professor Cioffi and his student B. Lee, etc.

paolo.franzoi 12/5/2012 | 4:35:37 PM
re: Copper: Still a Precious Metal?


The basic problem is that the whole thing ends up with lots of limits except for very short loop distribution.  That means if you look at U-verse (where I think they are running 2500 foot loops) that you would get better results but they will have to add power to the cabinets to take care of the extra dsp power required.  You would need new DSL line cards and new DSL modems.  If you have problematic copper (which they never tell you the results for in these demos - they use pristine copper in a pristine setting), then you are not going to get better results.

You do all of this to end up with less bandwidth than you can get with fiber.  Do it a couple of times and you end up with costs that exceed the cost of FiOS.  The worst part is that you end up doing all this while your cable company competition has rolled out Docsis 3 in much less time.

Will there be a market for this stuff?  I suppose in very short reach and in-building distribution.  The usual suspects for driving the initial implementations of very short reach high bandwidth DSL have gone fiber (Korea, Japan, and Singapore).  The Europeans?  AT&T?

I know that the technology involved in buggy whips today is greatly improved over those made 100 years ago.  




shygye75 12/5/2012 | 4:35:21 PM
re: Copper: Still a Precious Metal?

If this were a simple upgrade to squeeze another five years out of copper networks, it would be worth a shot. But as you say, copper plant is far from uniform in quality, so any telco that does take this path will have to be very careful about where to deploy. And ultimately, this is a temporary fix. I wonder if there are any valid cost reasons to take this path at all, given the inevitability of having to move to all-fiber at some point.

Sign In