Alcatel-Lucent Gives DSL Networks a Gigabit Boost 120
coondoggie writes "Alcatel-Lucent and Telekom Austria have completed the world's first trial of G.fast, new technology enabling gigabit broadband over existing copper networks. The technology is only intended for distances up to 100 meters or 0.06 miles. But at that distance and less it helps copper keep up with fiber." It works, says the linked article, "by continuously analyzing the noise conditions on copper lines, and then creates a new anti-noise signal to cancel it out, much like noise-canceling headphones."
A Cautionary Yay (Score:3)
Nobody's going to roll this out unless Google or another larger player starts rolling it out and making the existing 6 Mb/s connections unpalatable to consumers.
We'll probably see this in about 10 years.
Re: (Score:2)
I hope we never see this. It's a waste of resources - much better to just run fiber (Or ethernet, since we're talking about 100m) those last few meters and allow for easy upgrades in the future without the burden of ancient telephone cables.
Re:A Cautionary Yay (Score:4, Insightful)
Exactly. Why try to get the most out of something that's already bought and paid for when you can get something completely new and force people to either pay nearly usurious rates for slow speeds or go without.
Re: (Score:2)
Because the upgrade itself is unlikely to be cheap?
You have to compare the cost of the Gfast upgrade against the cost of replacing the run with fiber. That includes future maintenance and upgrades.
Re: (Score:2)
Re:A Cautionary Yay (Score:5, Insightful)
Disagree. This is great for apartment buildings/complexes where running fiber to the buildings isn't the problem. It's running fiber within the buildings (to the individual units) that's expensive. Something like a third of Americans live in apartments, so a system that spans the last few hundred feet in apartments without needing to rewire buildings would definitely be a win.
Re: (Score:1)
Agree, the target audience for this are undoubtedly apartment complexes / existing buildings.
I had verizon FIOS in an apartment last year. The fiber came into a utility closet in the basement. The individual apartments were wired up from there using DSL over the existing phone line. Unfortunately, the cap on plans available to apartments in that area was 20mbits, due to the limits of the DSL technology they used. Something like this would allow a fiber provider to offer gigabit internet to an apartment
Re: (Score:2)
My building has a 5Gbs fibre connection, I spoke with the teleco service guy when they were out replacing the equipment a couple months ago. The limiting factor now is the fact the building was wired about 12 years ago with Cat 5. Overall I can't complain too much other than when it's bad weather out on the weekend and everyone seems to be home streaming Netflix. Usually it's throttling at the server I'm connected to that is my bottleneck.
Re: (Score:2)
CAT-5 should get you gigabit ethernet speeds, no problem. If you've got more than 5 people in your building, the 5Gbps fiber connection is, in fact, still the bottleneck. You can oversubscribe a line only so far, before it gets congested.
Re: (Score:2)
Although this is kind of true (for good quality CAT5 cables), it needs to be pointed out that the range will be shorter than with CAT5e. You will never be able to get 1Gbps over a 100m long CAT5 cable. If you're lucky, you can get devices to negotiate gigabit speed at slightly over 60m.
Re: (Score:2)
Yes you will. The 1000BASE-T (802.3ab) specifications specifically say 100m on CAT-5 cable, and plenty of people have done so, successfully.
http://www.eetimes.com/document.asp?doc_id=1205868 [eetimes.com]
You might have range issues if an idiot patched your CAT-5 cables and did too much untwisting, or made similar mistakes. But if a professional installed it, they were tested and certified (by a certifier that cost several thousand dollars) to fully meet CAT-
Re: (Score:2)
Are you kidding me? You think buying a truck-load of brand-new high-tech DSLAMs and DSL modems for each apt is going to be CHEAPER than running CAT-6 or fiber up a wall?
I've lived in apartment complexes big and small, and I can't think of one of them that would be difficult to rewire. With multi-story units, you're talki
Re: (Score:2)
So you've got some kind of magic sauce that doesn't require new equipment on each side of those new wires you're pulling? That would be pretty impressive. If it's not one kind of new equipment, then its some other kind of new equipment, after all.
Re: (Score:2)
A gigabit ethernet switch is DIRT CHEAP compared to a D-SLAM.
And one is dirt cheap, while the other is insanely expensive. Blindly advocating the use of the insanely expensive one instead, just makes you look... insane.
Re: (Score:2)
Who said this needs the same kind of DSLAM that existing DSL uses? Isn't the reason DSLAMs are expensive because they sit at the central office and can serve hundreds/thousands of customers? You have no idea how expensive one that sits in an apartment building basement serving a hundred customers would cost.
Blindly believing new technologies cost the same as old technologies is... insane.
Re: (Score:2)
It won't... It'll need one that's far, far more expensive. Ditto for the modems... What does a new DSL modem cost today? At least $50 or so? Now you've got to upgrade it to this expensive new gigabit DSL technology, and give it a gigabit network interface, and put one in every apartment. I guarantee a new run of CAT-6 cabling in the most difficult case, will cost less than just the price of the modems for each apartment...
Re: (Score:2)
Near as I can tell a 48 port DSLAM is about a couple thousand more than 48 port Gigabit ethernet. If you can get 48 apartments rewired for Cat5 for just a couple thousand dollars, you'll have to let me know how, cause that's a pretty impressively cheap cost. Until you know how much this new technology costs, you're just spouting BS, so why don't you just stop embarrassing yourself.
Re: (Score:2)
No, you're the one with the insane schemes, despite knowing shit about the subject until I came along and called you out for the blatant stupidity.
Re: (Score:2)
Nope. Everyone is free to plug their WiFi APs/Routers into the building ethernet, just as they would into a DSL modem. And "managed" switches allow you to isolate every switch port onto their own VLAN, just as if they were on physically separate switches, so they can't communicate directly with each other. You need a VLAN compatible router upstream, but any $40 home router tha
Re:A Cautionary Yay (Score:4, Insightful)
I doubt you'll ever hear an argument that you shouldn't run modern cabling no matter what decade we live in - the problem is that the money just isn't there most of the time to do so to replace infrastructure that is existing.
Re: (Score:2)
I don't believe the can make it work without significantly different hardware than what is currently available. That means that for most situations, running a new cable would be cheaper than adding costly hardware on both ends.
neat idea (Score:4, Interesting)
The noise-cancelling scheme sounds interesting. The hardest part though is figuring out what exactly is noise - so it sounds like they would have to either invert the intended signal to cancel it out on the path to the noise measurement, or they would have to periodically turn the signal off so they can get a clean measure.
Re:neat idea (Score:4, Insightful)
My VDSL link from AT&T already does noise measurement. Buried inside of the web page for the magic box on my end, I can see a graph of what portions of its spectrum it is actually using.
Seems to work OK: There is a very neat notch which corresponds with a nearby AM broadcast tower.
Meanwhile, it doesn't have to turn off the entire signal all at once. Just parts of it. One end says "Hey, George, we're going to turn off 1.6MHz and look for noise there," and the other end says "OK Bob. Let's do it."
Or at least, that's one way. *shrug*
Re: (Score:2)
Mmm true - or even just use a consistent "rotation" to do it, in that the rotation is a waveform itself that the other end can sync to - then there's no negotiation to be worried about!
Or you can send bursts of a signal (along with such a scheme or just periodically) and the other end can know what was sent (prearranged?) and with some simple subtraction it has a noise profile
Sorry. I enjoy thinking about this kind of thing. Can you see the sidebands around the AM tower as well? Or is the resolution not goo
Re: (Score:2)
It is interesting stuff to me also. I made a screenshot:
http://s23.postimg.org/rw44oefaz/1330am.png [postimg.org]
It's a graphical representation of the frequencies that are deemed most useful for my particular circuit (which is an edge-case at around 3000 feet.)
No idea about sidebands since I don't work with AM, or anything else in that part of the spectrum. And the other notches are unexplained as well.
Re: (Score:2)
Well basically when you modulate a carrier (AM is modulating the amplitude of the carrier with the data signal) half the energy ends up in the carrier signal, and the other half gets split up between to sidebands that sit to either side of the carrier on the spectrum. The sidebands are mirror-images of each other.
There's techniques to suppress them (meaning you don't waste half your transmission energy on redundant signal) but broadcast AM/FM radio doesn't utilize it. Other advantages mean a more efficient
Re: (Score:2)
But I don't want to go down the rabbit hole. I don't get paid for that. And I don't know how that's useful to me.
Things I care about:
Troubleshooting a digital Astro or P.25 system on 700/800MHz? Check.
2.4GHz Wifi? You betcha.
Unlicensed 5.4/5.7GHz? Sure. Lots of spectrum there, but gotta avoid weather Doppler radar for parts of it, depending on orientation.
~60GHz microwave links? Ugh, because the alignment is contingent (literally) on time of day and phase of moon, but, sure: A solid full-duplex 1Gbp
Re: (Score:2)
Well it might interest you to know that some of the encoding methods used at these lower frequencies can be used on the higher ones, too.
For example, Olivia MFSK [wikipedia.org], which can be copied even when buried 10dB under the noise floor!
Re: (Score:2)
There's unpredictable (random) noise and there's predictable noise. You can't do much about random noise except for trying to determine how much of the noise there is in particular frequency bands. But you can work around predictable noise. The general idea is for the telco equipment to run a bundle of connections in sync. Then they can correlate noise going from connection to connection in a bundle. Then they modify signals transmitted to a particular connection to include anti-noise component, that is a n
Noise canceling is NOT the key. (Score:2)
The noise-cancelling scheme sounds interesting.
If you'll read TFA a little more closely than the OP did, you'll find that the noise-canceling thing is NOT how they got the 1G-ish single-pair link to work.
What the noise-canceling thing is about is when you have TWO OR MORE pairs bundled into a single logical link. Then it figures out what the cross-talk between the individual pairs looks like and cancels THAT out. This lets the individual signal pairs run as fast as a lone pair and the total bandwidth of N
Re: (Score:2)
If you'll read my post a little more clearly, it's the noise-cancelling aspect that I personally find interesting.
Good information, but keep the snark turned off.
great announcement (Score:4, Interesting)
Re: (Score:2)
Most likely this will be used for the last 100 metres from the cabinet to the home, in a Fibre To The Cabinet (or Curb) scheme.
Yes, the article mentions this at the end.
Re: (Score:1)
Dude, this isn't going to reach from the cabinet, it'll have to be in a small box at the curb and at that point why bother, just run a piece of glass or plastic and have a network that will work for 25+ years instead of a 5 year stop-gap measure. To put this in perspective we run into 100m limitations within a single building, it often requires carefully planning where to place the IDF(s) to make sure that all drops are within the 100m length limit for ethernet, using this for a last mile solution is stupid
Re:great announcement (Score:4, Informative)
Dude, this isn't going to reach from the cabinet, it'll have to be in a small box at the curb and at that point why bother, just run a piece of glass or plastic and have a network that will work for 25+ years instead of a 5 year stop-gap measure. To put this in perspective we run into 100m limitations within a single building, it often requires carefully planning where to place the IDF(s) to make sure that all drops are within the 100m length limit for ethernet, using this for a last mile solution is stupid.
well do you want to rewire an entire apartment block with ethernet or use this? that's what it pretty much boils down to.
Re: (Score:3)
Why do so many comments in this thread assume everyone lives in a single-family home? In an apartment setting, it's huge cost savings to avoid having to rewire that last few hundred feet to the individual units.
Re: (Score:2)
In an apartment setting, it's huge cost savings to avoid having to rewire that last few hundred feet to the individual units.
But the ongoing maintenance costs of maintaining proprietary network gear would be more expensive, wouldn't it?
Re: (Score:2)
Do you think that just because you used the word "proprietary"? I'm not really sure what maintenance cost you're talking about. There's replacing failed gear and... what am I missing? And why do we believe it will be any more maintenance intensive than fiber gear?
The cost of running fiber through an existing apartment building can be huge. In some cases, costs can be high enough that it almost doesn't matter what the ongoing maintenance costs are.
Re: (Score:2)
You're still going to run into problems unless it's a fairly small apartment building, 10 stories up is half your limit and then you still have two horizontal runs that max out at ~60-70'.
Re: (Score:2)
Including horizontal runs, 100 meters probably gets you coverage for most buildings up to 15 stories. Really wide buildings usually have multiple telephone risers and you'd put a converter at the base of every riser. Taller than that really depends on how the existing telephone lines lay out within the building. You could probably put an additional one of these on the 15th floor, within the riser, for many 30 story buildings and be covered up to 30 stories (getting a single fiber line up the riser is no
Re: (Score:1)
Re: (Score:2)
Over a distance like that, you can just use ethernet and repeaters. There's no need for anything more expensive or exotic than that.
Re: (Score:1)
Re: (Score:2)
Their secret pricing is super-annoying and theoretically illegal -- how can you sign a contract without knowing what the rate will be when the promotion ends??? Ridiculous. It's just fodder for a class action lawsuit that will yield $5 coupons off an HBO subscription... but I digress.
I ditched TimeWarner back when Code Red was making the rounds on Windows boxes and they shut off inbound port 80. I complained and several support people told me they considered web hosting "an enterprise service" and if I want
Re: (Score:1)
You can't compare DSL to Wifi because wired connections are typically more reliable than wireless (not even discussing security concerns here) and less prone to noise and generate less radio waves
Re: (Score:2)
OK, so, 100 Mbps at the scale of a block is the real point here. Well, 100 Mbps would be a huge upgrade from my current cable speeds.
To me, 1 Gb at 100 yards sounds like the answer to a backwards question: "how close
What an improvement over gigabit ethernet! (Score:3)
With gigabit ethernet, you can go 100 meters with cat6 wiring.
So, all this provides is the ability to use a single pair instead of two pair...at the expense of having equipment to terminate it at each end.
This has zero applications for delivering broadband. Nobody is within 100 meters of a DSLAM.
Re: (Score:3, Informative)
Lots of people are within 100 metres of the cabinet however, in a FTTC scheme.
And perhaps this method also works to improve bitrate at longer distances too.
Re: (Score:2)
At that point you might as well just run the last drop and pay much less for the needed equipment. A fiber modem plus a router with good ol' Cat 6/1000 BaseT will definitely be cheaper than said fiber modem plus DSLAM plus a DSL modem per customer.
Re: (Score:2)
100 metres as the crow files, maybe... but within 100m of cabling, which often doesn't follow the most efficient path? Heck, my house is 10m from a VDSL-enabled telecoms cabinet but I happen to know that my house is actually connected to one that's 80m away (as the crow flies) and that there's about 150m of cable between it and my phone socket.
Basically no-one would get to see that 1Gb since, as TFA states, they managed to get 800Mb/s over 100m over a single brand-new cable and got 500Mb/s over an older sin
Re: (Score:2)
Which you have to run to each house, sounds not very practical.
Phone lines are already going to each home, so this is great for fiber to the neighborhood then use the POTS line since it is already there.
Re: (Score:2)
Funny, true story:
My uncle was classified as a "First Responder", so according to some or other Homeland Security policy he's federally mandated to have always-on high speed internet access in his home.
There's nothing but state parks and corn fields around his property for probably five miles in any direction. Cable TV isn't even available in the nearest town, which has a population of a few hundred people.
So Verizon bought a 5m x 5x plot of cornfield directly across the street from his house and built a ti
The point is to use existing wiring (Score:1)
The point is to make the best use of the likely-decades-old "telephone wire" going from the "pole," "telephone box" (for underground wires), or in some cases, "neighborhood fiber box" to the customer's "internet box" (e.g. DSL modem).
This wire is typically no better than "CAT-3" and frequently far worse, electrically speaking. If it's older than a few decades, it's probably chosen for low cost and voice-grade capability. It may also be run close enough to other wiring that it will pick up noise that is to
Re: (Score:2)
1Gb @100m doesn't require Cat6, only Cat5, 10Gb @100m requires Cat6A (Cat6 isn't used in any official standard, there's an annex that calls it out for distances to 55m but it was never adopted and real world testing shows a drop off in reliability around 37m)
Re: (Score:2)
Apartment buildings.
If the walls are already up, and the phone wiring already exists, and none of it is Cat-anything: Put the DSLAM (or whatever) in a closet, just like you would a gigabit switch.
And done.
Sure, it costs money. But so does any other chunk of infrastructure that people are willing to pay for.
(Also: Cat6 is unnecessary for the gigabit standard that we commonly use today, 802.3ab, which says you can go 100 meters on Cat5. Even Cat5e is overkill. Cat6? Meh.)
*SIGH* now if only the teclos would deploy it... (Score:1)
I'm running on 10/1 cable modem in Central Florida. I can't get DSL at the house - it's not available for any price.
Lightning strikes blew up a bunch of landline phone connections in the neighborhood recently, including mine. I saw the tech working on the repair at my box, which is two doors down from me, about 150 feet from my house. I took him a bottle of cold water because I wanted to get a look at the inside of the box. Chatted with him for a few minutes while he was swapping cards and drinking the
Try some political pressure (Score:1)
Depending on local and state politics, political pressure and "public shame" can overrule bean-counters.
Re: (Score:1)
...and then Johnny woke up in the real world and realized he'd been dreaming of imaginary places again...
I mean, "public shame"? Are you for fucking real?
Re: (Score:1)
Probably doesn't matter (Score:3)
The phone companies have long since proven they aren't going to make any further substantial investment in their copper networks, and are simply determined to milk them for as long as possible. They are in fact actively trying to shed their copper networks and go wireless, which has less regulation and higher profit margins.
The odds of AT&T/Verizon making a huge investment in technology that will be lucky to last a couple of years (fiber scales to 10 Gb fairly simply, and cable can probably get close with future revisions of DOCSIS), in a domain they are actively withdrawing from, is pretty much zero.
Re: (Score:2)
I read somewhere that British Telecom has a stock market value that is lower than the scrap value of the copper cables it owns. I suspect many other companies are in a similar position, and it shows why they aren't that keen to install new copper cables.
Re: (Score:2)
Yes and the analysis was completely wrong. The amount of copper BT owns is a lot less than they claimed as they miscounted it, and a good deal of it is copper clad steel as copper does not have sufficient strength to be strung from a pole to the house.
Re: (Score:2)
FIOS is 15/5 in NYC for the cheapest tier. why would i want something faster? netflix needs 3-5 mbps
Re: (Score:2)
The last Olympics needed ~6-7Mbps for 720p, ~8-10Mbps for 1080p so if you had two people in your house who wanted to watch different sports you'd very tight at 720p and SOL at 1080p. Those streams were also pretty blocky, the realtime compressors probably could have used 2-3x the bandwidth to make things smoother.
Re: (Score:2)
so upgrade to the highest tier service that month
what do you need it for every day?
Re: (Score:2)
I'm on 2Mb/s cable you insensitive clod!
Re: (Score:2)
So just a more expensive Gigabit Ethernet then (Score:1)
As someone with an ISP in the DSL game. WTF? So Gigabit at 100 meters? Isn't this really just Gigabit Ethernet over a single pair? And really, who is within 100 Meters of their DSLAM. That would cover maybe four homes.
Re: (Score:3)
Re: (Score:2)
You run fiber to the curb, then use this to deliver that to the home.
Re: (Score:1)
If your running fiber to the curb, just run PON and go fiber to the home. less latency, cheaper, better overall.
Re: (Score:2)
How is it cheaper to run new fiber than to use the existing copper?
A considerable amount of cost is tech time, which has to mesh with time when users are home. That means doing a single block can take a week or more.
Re: (Score:2)
Yeah, but what about the encryption and all other BS you will need for PON?
Re: (Score:1)
PON is just an example, but there are other fiber options. We handle it by controlling the endpoints and not allowing user supplied endpoints. Or official hand off is the Ethernet port, not the fiber.
Re: (Score:2)
Yeah, because no one could ever put in their own endpoint. I hope that is not your actual answer.
The handoff can be fairy dust, if the fiber is on my property you had better not just trust the endpoints like that.
Re: (Score:2)
Considering DSL's promises, 100 Mb/s at 100m would be very lucky (and a lot more expensive than Gigabit ethernet, which is dirt cheap).
The only practical application this has is in fiber to the Curb/Building scenarios, where Fiber is run to a local, central location and then distributed via some other interface. Using DSL would theoretically allow for existing wiring to be used, but that sounds like major trouble in the making. It's much easier to just run fiber to the customer's premises and give them a fi
My love of copper and the Problem with FIOS... (Score:1)
When FIOS came to my neighborhood, most people cheered. I didn't. Not because I didn't want the higher speeds, but because I was concerned about the bigger picture. The baby bells (mine being Verizon) were ordered by the courts to share their copper with competing companies. My understanding is that they were under no obligation to share their fiber. As a result, they've been ripping out copper everywhere they can to replace it with fiber.
I understand that the copper infrastructure was aged and probab
Re: (Score:1)
I'm assuming Verizon was your ILEC. Even if you had 8 DSL providers, they all used Verizon copper and either leased Verizon lines, or resold Verizon services. So they all had to deal with Verizon. In my case, Qwest, now Centurylink is the ILEC. They hate to share, they make it nearly impossible charge exorbitant fees, and their local techs flat out blackmail and lie to us on repair issues. So we are doing more and more fiber. We still have to fight the power company for pole space, but once on we are
pointless (Score:5, Informative)
I work with this kind of equipment. The problem isn't the last 100 feet... we've got tech that will do 100mb @ 30,000 feet relighably. If we could get that to people they'd be thrilled. The problem is the trunks leading to the DSAs. They cost upwards of a million dollars a pop to install, which is barely cost effective in city centers... but get out in rural areas where cable companies don't even bother to serve and you have as few as 12 people off a remote. Sorry, but that's only going to get 2 T1s feeding it if they're lucky. Gigabit speeds to and from equipment fed by a 3mb trunk is useless.
The real problem with broadband is the link between the CO and the remote. This goes for DSL and Cable. solve that problem and rural broadband will explode. Cable doesn't even have facilities in those areas so it would have to be over phone copper. Get gigabit speeds on 10+ miles of unshielded copper pairs... that's the goal. Good luck.
Re: (Score:2)
I have to echo this.
Over 100m, Gigabit over copper is already trivially possible. My computer has it built-in, so has anything you've bought in the last few years.
The problem of local connections such as 100m is solved. We're there. It's not a problem. Even a community project that threw Cat6a round the village could do a half-decent job of it at Gigabit speeds over even longer distances.
The problem is the link back to some Internet-connected point, as always. As you say, it's pointless (and quite easy
Re: (Score:2)
This is just stupid. Yes, gigabit ethernet works over CAT-6 data cabling. But guess what? Phone lines AREN'T CAT-6 data. Phone lines are commonly CAT-1 voice. Unshielded, untwisted, etc. If you're super-lucky, some of your phone lines MIGHT be CAT-3 voice, but that's also NOT THE SAME THING AS
Re: (Score:2)
No. I'm not. My point is that Gigabit Ethernet over copper is not anywhere close to "state of the art", and hasn't been for a long time - given the fact that even the cheapest of netbooks comes with a Gigabit Ethernet port capable of similar things. Hell, 10G is buyable today if you know where to look, and 40G is on the way. Yes, it's over twisted pairs, but if you can do 40G over twisted pairs, it's because of the technique and the error correction methods - NOT THE CABLE. The cable helps - obviously
Re: (Score:2)
This is massively, totally, completely, and utterly untrue. It seems to be at the heart of why everything you've said is wrong, and as I said before: "stupid".
You can go down to your nearest tool store and buy phone line by the foot (commonly "CAT-3 voice"). Get about 100m of the stuff, and crimp some RJ45 connectors on the thing, Hook it up between two computers with GbE NICs, and try sen
Re: (Score:2)
They might be thrilled, TODAY, but two+ years from now they'll be moaning about their slow DSL service, and eying cable and FIOS.
Re: (Score:2)
The problem is the trunks leading to the DSAs
By the way, DSA means "Digital Service Area", the area of customer premises served by a DSLAM (remote/outside plant or inside plant).
Re: (Score:2)
Is that why my dial-up sucks due to line noises and only connecting 24000-31200 speeds (mostly 28.8k and 3 kB/sec)? :(
How Far for 50Mb? (Score:2)
There's a ton of people on VDSL2 that get between 5-7Mb. Usually in the 2500-3000 meter range. It would be interesting to see how much of a boost these technologies could lend. Getting 30-50 Mb with a simple DSLAM and Modem swap would be game changer.
Re:Fiber to the curb (Score:5, Funny)
Those 17 people within 100 meters of the hub are going to be thrilled!
Re: (Score:3, Informative)
Great, now all the ads for DSL will be for "Up to 1 Gb/s download speed [tiny print]Actual speed will vary"
Re: (Score:2)
What you're suggesting is compression.
As the data gets larger and the possibilities grow, the size of the information to reference that information gets as big as, if not bigger, than the information itself.
And, at best, you'd get something roughly equivalent to gzip'ing HTML pages, zipping up PDF's, or compressing images (all of which already happen as part of the standard protocols).