Alcatel Lucent Sets New 10 Gbps Copper Speed Record Wednesday Jul 09 2014 14:21 EDT Alcatel Lucent this week announced that they've achieved yet another data transmission speed record over copper lines. According to the company's announcement, Bell Labs, the research arm of Alcatel-Lucent, has transmitted data at speeds of 10 Gbps over copper using a technology they're calling XG-Fast. XG-Fast won't be made commercially available until sometime in 2015, according to the company. As is usually the case these advancements come with great distance restrictions; 10 Gbps was achieved using bonded pairs over a distance of just thirty meters. |
mixdup join:2003-06-28 Alpharetta, GA |
mixdup
Member
2014-Jul-9 2:32 pm
What's the point...To do the 1Gbps symmetrical, which is the actual selling point of this technology, you have a max distance of 70 meters. That's shorter than the max distance of gigabit ethernet over category 6. You're going to have to put fiber all the way to that 70 meters, then some electronics. Why not just put the fiber the additional 300 feet? The electronics Alcatel are trying to sell can't be cheaper than paying a couple of guys to spend three hours burying 300 feet of fiber.
And if the point is for MDUs, then why not terminate to ethernet instead? | |
| |
1 recommendation |
Re: What's the point...or coax, or CAT, or wireless, or any of the more sane options. | |
| | |
to mixdup
No need for cat6 - 1GBase-T is designed to work over plain cat5 just like 100Base-TX and will work at distances up to 100m on good cat5 as long as there are no strong noise sources nearby to induce differential noise in the pairs.
Cat6 was a requirement for 1GBase-TX but 1GBase-TX is practically extinct.
10GBase-T can work on up to 55m of cat6 or 100m of cat6A. | |
| | | |
Re: What's the point...cat5e can do 1Gb, not cat5. Google Fiber is using 5e from fiber jack to network box. | |
| | | | |
| |
to mixdup
I wanna know how many hundreds of millions they threw at this BS so-called innovation in technology?!? the hell with copper!! Just lay fiber!! This sounds like another black budget project like the flying F-35 dumptruck from Lockheed Martin. $160+ million per plane and the damn thing doesn't even fly that well. In truth it only costs about $20 million a plane and the other $140 million they are pocketing... | |
| | aaronwt Premium Member join:2004-11-07 Woodbridge, VA Asus RT-AX89
|
to mixdup
said by mixdup:To do the 1Gbps symmetrical, which is the actual selling point of this technology, you have a max distance of 70 meters. That's shorter than the max distance of gigabit ethernet over category 6. You're going to have to put fiber all the way to that 70 meters, then some electronics. Why not just put the fiber the additional 300 feet? The electronics Alcatel are trying to sell can't be cheaper than paying a couple of guys to spend three hours burying 300 feet of fiber.
And if the point is for MDUs, then why not terminate to ethernet instead? Cat5e is also rated for 1Gbps at 100 meters. | |
| | |
to mixdup
The telephone companies want to avoid having to enter private property to do upgrades. The costs to do that go up significantly compared to being able to just install technology on the public right of way telephone poles, in buried conduit near public roadways, or in above ground cabinets along public rights of way. So you might get a significant data rate upgrade at a relatively low cost compared to FTTH. Alcatel-Lucnet has stated in the past that they do communicate to their DSL clients that eventually FTTH will be the only technological solution left to apply to an upgrade situations, and that these new ways of using DSL are only temporary solutions. A lot of people I know think Alcatel-Lucent will get out of the DSL business by 2027. Unless some more technological magic is conjured up, DSL will be pointless compared to doing an FTTP deployment. | |
| | | mixdup join:2003-06-28 Alpharetta, GA |
mixdup
Member
2014-Jul-9 6:07 pm
Re: What's the point...said by davidhoffman:The telephone companies want to avoid having to enter private property to do upgrades. The costs to do that go up significantly compared to being able to just install technology on the public right of way telephone poles, in buried conduit near public roadways, or in above ground cabinets along public rights of way. So you might get a significant data rate upgrade at a relatively low cost compared to FTTH. Alcatel-Lucnet has stated in the past that they do communicate to their DSL clients that eventually FTTH will be the only technological solution left to apply to an upgrade situations, and that these new ways of using DSL are only temporary solutions. A lot of people I know think Alcatel-Lucent will get out of the DSL business by 2027. Unless some more technological magic is conjured up, DSL will be pointless compared to doing an FTTP deployment. Why does burying cable on someone's private property cost any more than burying it in right of way? When AT&T installed phone at my grandfather's house and charter installed cable at my house, both buried, they didn't pay me to do it. I understand these are technologies to squeeze what you can out of copper and eventually they will hit a wall, but it seems like they've hit that wall. If you're going to go up to 100-300 feet away from the house and then still put in some expensive electronics--essentially a one user DSLAM--you might as well put the money you were going to spend on the fiber-to-copper electronics into paying some laborers to bury the fiber in the yard. | |
| | | | BiggA Premium Member join:2005-11-23 Central CT |
BiggA
Premium Member
2014-Jul-9 6:53 pm
Re: What's the point...Exactly. It seems rather pointless. AT&T has already tapped out VDSL. They need to go to GPON and do it the right way! | |
|
| |
Flyonthewall to mixdup
Anon
2014-Jul-9 6:04 pm
to mixdup
Seems pointless to me too, like when Edison wanted to build power plants every block to use DC power. | |
| | |
to mixdup
said by mixdup:To do the 1Gbps symmetrical, which is the actual selling point of this technology, you have a max distance of 70 meters. That's shorter than the max distance of gigabit ethernet over category 6. You're going to have to put fiber all the way to that 70 meters, then some electronics. Why not just put the fiber the additional 300 feet? The electronics Alcatel are trying to sell can't be cheaper than paying a couple of guys to spend three hours burying 300 feet of fiber.
And if the point is for MDUs, then why not terminate to ethernet instead? This is over ONE pair. Ethernet uses 4 pairs. | |
| | | |
Re: What's the point...When you are down to sub-100m fiber-to-the-curb/pole distances, it is probably cheaper to re-wire the neighborhood with cat6 and Ethernet switches than use modems and DSLAM with DSPs powerful enough to handle 1-10Gbps: a 1GbE switch costs less than $10/port while a 1/1Gbps modem would likely cost over $250 and the DSLAM/whatever at the other end should cost at least as much per port as well.
This makes little sense beyond cases where running new cables is practically impossible. | |
| | | | mixdup join:2003-06-28 Alpharetta, GA |
mixdup
Member
2014-Jul-10 12:09 am
Re: What's the point...said by InvalidError:When you are down to sub-100m fiber-to-the-curb/pole distances, it is probably cheaper to re-wire the neighborhood with cat6 and Ethernet switches than use modems and DSLAM with DSPs powerful enough to handle 1-10Gbps: a 1GbE switch costs less than $10/port while a 1/1Gbps modem would likely cost over $250 and the DSLAM/whatever at the other end should cost at least as much per port as well.
This makes little sense beyond cases where running new cables is practically impossible. This is exactly my point. And in the case of an MDU, if you have two pair to bond then you could probably get good enough ethernet performance out of those two pair if you were too cheap to re-wire with Cat5+. | |
| | | | | |
Re: What's the point...The main problem with using whatever-is-there with Ethernet is that Ethernet has no mechanisms to cope with "whatever" if that 'whatever' is too far from the expected wiring performance.
The only "modern" Ethernet standard that works over only a pair of cat5-grade pairs is 100Base-TX, which would still be a welcome upgrade for many. 1GBase-TX on the other hand requires four pairs of cat6/6a while 1GBase-T requires four pairs of cat5.
If settling for 100Mbps, might as well go with VDSL2: the 30a profile enables up to 100M/100M at 150-200m using a single pair of cat3/4 with proper support for all manners of common line imperfections. | |
| | | | | | mixdup join:2003-06-28 Alpharetta, GA |
mixdup
Member
2014-Jul-10 7:29 am
Re: What's the point...said by InvalidError:The main problem with using whatever-is-there with Ethernet is that Ethernet has no mechanisms to cope with "whatever" if that 'whatever' is too far from the expected wiring performance.
The only "modern" Ethernet standard that works over only a pair of cat5-grade pairs is 100Base-TX, which would still be a welcome upgrade for many. 1GBase-TX on the other hand requires four pairs of cat6/6a while 1GBase-T requires four pairs of cat5.
If settling for 100Mbps, might as well go with VDSL2: the 30a profile enables up to 100M/100M at 150-200m using a single pair of cat3/4 with proper support for all manners of common line imperfections. But can you do a single port VDSL2 device that converts from fiber to VDSL for less than the cost of either running fiber all the way to the customer's premise or for less than the cost of an ethernet media converter? | |
| | | | | | | |
Re: What's the point...Why would you want a single-port VDSL2 DSLAM?
The ~200m range at 100Mbps is enough to reach 20-150 subscribers from a single pedestal in (sub-)urban neighborhoods. Alcatel makes sealed 24-ports remotes that can be hung on aerial cables, strapped to utility poles, stuffed in curb-side boxes or in underground bunkers for buried wiring, stuffed in MDUs' comms closets, bolted outside buildings, etc. for these sorts of deep-VDSL2 deployments. | |
| | | | | | | BiggA Premium Member join:2005-11-23 Central CT ·Frontier FiberOp.. Asus RT-AC68
|
to mixdup
said by mixdup:But can you do a single port VDSL2 device that converts from fiber to VDSL for less than the cost of either running fiber all the way to the customer's premise or for less than the cost of an ethernet media converter? It only really makes sense in a large MDU where you have noncompliant landlords who won't allow a full FTTH build to happen. Verizon has used it for this in NYC at 50/10 speeds, but they prefer to go full FTTH when they have cooperative landlords. And in that case, it's not a single port conversion, it's many ports at once. said by tmc8080:BTW, gigabit Ethernet would be silly to deploy when, in 2015- 2016(by the time the build would near actually hooking someone up) ATT could just as easily deploy 10, 40, and 100 gigabit fiber ONT/Nodes for a negligible premium over any dsl builds. It's a foregone conclusion that AT&T will face enormous re-build costs when they have to convert fiber dslams to nodes which will no longer conform to the distance requirements from the home in many cases since a ton of dsl lines go thousands of feet (and get crappy bandwidth as a result). I'm pretty sure fiber splicing labor won't be any cheaper in the future. And AT&T is making things even more complicated by running fiber out of their VRADs. They should skip over the VRADs, and do it right by running the GPON right out of the CO, so that they can eventually get rid of the VRADs entirely. But what are you saying about distance? Fiber can go something like 8 miles, whereas VDSL craps out just shy of a mile with pair bonding, 3kft without. So running fiber out of VRADs, although a crummy solution for AT&T, will allow them to connect way more people. said by mixdup:That fiber from amazon isn't suitable for outdoor/burying use, but your point is a valid one. However, a lot of the cost isn't in the fiber itself but in the costs of installation (digging up someone's yard, and more to the point the cost of splicing). If it was digging, most of CT would have GPON for U-Verse, as we have mostly overhead lines, where AT&T can just slap wires up with little effort. But we don't. Because AT&T is too cheap to even wire the low-hanging fruit for fiber. If they did 100% GPON where there are overhead lines, and then GPON where there are extra conduits, and VDSL for some older underground and MDU installations, they'd probably have be able to offer 100mbps service to 95%+ of CT's population... said by InvalidError:The ~200m range at 100Mbps is enough to reach 20-150 subscribers from a single pedestal in (sub-)urban neighborhoods. Alcatel makes sealed 24-ports remotes that can be hung on aerial cables, strapped to utility poles, stuffed in curb-side boxes or in underground bunkers for buried wiring, stuffed in MDUs' comms closets, bolted outside buildings, etc. for these sorts of deep-VDSL2 deployments. No. In a typical "suburban" area, you might a dozen houses in that range, and I know of many more exurban areas where you'd be lucky to get half that. Some urban areas would get you more. Also, the inside wiring distances could cut that down significantly, especially with larger houses, and when you don't really know how far the inside wiring goes when planning. Deep-VDSL2 just sounds really stupid unless there is absolutely no other way to do things. | |
|
| | quetwoThat VoIP Guy Premium Member join:2004-09-04 East Lansing, MI |
to DataRiker
Actually, what they are proposing is using bonded DSL pairs (two pair). Ethernet actually only uses two pairs (1 pair Tx, 1 pair Rx), and an additional, optional pair for PoE. Cut out the PoE, and the wire usage is about the same. | |
| | | | |
Re: What's the point...said by quetwo:Actually, what they are proposing is using bonded DSL pairs (two pair). Ethernet actually only uses two pairs (1 pair Tx, 1 pair Rx), and an additional, optional pair for PoE. Cut out the PoE, and the wire usage is about the same. Gigabit Ethernet actually uses all 4 pairs . | |
|
| |
to mixdup
exactly.. here's a spool on amazon to give you an idea of how irrelevant dsl technologies are in today's world. » www.amazon.com/300FT-Fib ··· 083V134UBTW, gigabit Ethernet would be silly to deploy when, in 2015- 2016(by the time the build would near actually hooking someone up) ATT could just as easily deploy 10, 40, and 100 gigabit fiber ONT/Nodes for a negligible premium over any dsl builds. It's a foregone conclusion that AT&T will face enormous re-build costs when they have to convert fiber dslams to nodes which will no longer conform to the distance requirements from the home in many cases since a ton of dsl lines go thousands of feet (and get crappy bandwidth as a result). I'm pretty sure fiber splicing labor won't be any cheaper in the future. | |
| | | mixdup join:2003-06-28 Alpharetta, GA |
mixdup
Member
2014-Jul-10 9:32 am
Re: What's the point...said by tmc8080:exactly.. here's a spool on amazon to give you an idea of how irrelevant dsl technologies are in today's world.
»www.amazon.com/300FT-Fib ··· 083V134U
BTW, gigabit Ethernet would be silly to deploy when, in 2015- 2016(by the time the build would near actually hooking someone up) ATT could just as easily deploy 10, 40, and 100 gigabit fiber ONT/Nodes for a negligible premium over any dsl builds. It's a foregone conclusion that AT&T will face enormous re-build costs when they have to convert fiber dslams to nodes which will no longer conform to the distance requirements from the home in many cases since a ton of dsl lines go thousands of feet (and get crappy bandwidth as a result). I'm pretty sure fiber splicing labor won't be any cheaper in the future. That fiber from amazon isn't suitable for outdoor/burying use, but your point is a valid one. However, a lot of the cost isn't in the fiber itself but in the costs of installation (digging up someone's yard, and more to the point the cost of splicing). AT&T worked around some of this in Austin, by ordering lengths of fiber pre-made with connectors on the end. Someone goes out, measures how long a run will be down the street or from the street to a house, and then they order from Corning with the ends already on it. Then, they don't have to pay someone to splice it in the field. A lower paid guy can quickly install it and just plug it in. I don't know why this isn't expanded out with more carriers, just start making pre-made fibers in different lengths, and if you need a custom length order it. Seems taking the splicing out of the field would cut down on a lot of the cost. | |
|
|
Keep tryingI mean it's 30m in LAB conditions, not real world.
Maybe this is good for cross-connecting servers or most likely within multi tenant environments (MDU). This is pretty much useless for anything outside of that.
It's good they keep innovating, however stomping on a dead corpse, means you are still dead.
TBH, this is pretty much useless in the US, how we have deployed copper is sketch compared to Europe and the distances are greater.
I did read at one time they could connect them in loops (like token ring), but then now your node is a ring topology, not P2P and that means no gig for you. | |
| fg8578 join:2009-04-26 San Antonio, TX |
fg8578
Member
2014-Jul-9 2:52 pm
It's a startThe distances are very short, that's true. But with continued development, the distances hopefully will increase. | |
| | |
Re: It's a startNot under current understanding, but possibly. Go to the release and watch the video on the Shannon Limit. | |
| | mixdup join:2003-06-28 Alpharetta, GA |
to fg8578
said by fg8578:The distances are very short, that's true. But with continued development, the distances hopefully will increase. You would think so, but given that they're not able to increase distances at lower bandwidths, I don't think this will ever be the case. The new chart for VDSL2 and G.fast have the same speeds at the same distances before. 1000+ feet speeds are not increasing. It's just time to declare copper dead. The industry should be putting their research into making fiber cheaper through improved installation methods. | |
| | | fg8578 join:2009-04-26 San Antonio, TX |
fg8578
Member
2014-Jul-9 3:28 pm
Re: It's a startsaid by mixdup:It's just time to declare copper dead There was a time back in the mid 90's when everyone thought copper couldn't support data rates above 56 kbps. And while complaining how Moore's Law didn't apply to the outside plant, we have gone from 56kbps to a demonstration of 10Gbps (albeit over only 30m of copper). Who knows what the next demonstration will bring? I'm willing to bet that speeds and distances will both increase. | |
| | | | mixdup join:2003-06-28 Alpharetta, GA |
mixdup
Member
2014-Jul-9 3:39 pm
Re: It's a startsaid by fg8578:said by mixdup:It's just time to declare copper dead There was a time back in the mid 90's when everyone thought copper couldn't support data rates above 56 kbps. And while complaining how Moore's Law didn't apply to the outside plant, we have gone from 56kbps to a demonstration of 10Gbps (albeit over only 30m of copper). Who knows what the next demonstration will bring? I'm willing to bet that speeds and distances will both increase. they're already below what can be accomplished with ethernet over two twisted pair. If you have two twisted pair to bond, why not just use ethernet and get more out of them? If distances were going to increase, or if they were working on it, shouldn't that be happening soon? VDSL2 is still stuck at the same distances. They keep eeking out more and more, closer and closer, but what does that matter? What problem does it solve? | |
| | | | | |
Re: It's a start1GBase-T, which is the prevalent form of gigabit Ethernet today, uses all four pairs simultaneously in both directions to achieve full-duplex gigabit performance over 100m of cat5 wiring.
For 1GBase-TX, which has been practically extinct for over a decade already, uses two pairs dedicated to each direction and requires cat6a to achieve 1Gbps at 100m.
1GBase-T was initially more expensive than 1GBase-TX due to its more complex electronics and signal processing (full-duplex across all four pairs instead of half-duplex on a pair of pairs in each direction) but won anyway because most companies valued not having to replace their existing cat5/5e with more expensive cat6/6a far more. The fast demand and mass-production growth for 1GBase-T quickly drove prices below 1GBase-TX and 1GBase-TX vanished. | |
|
| | | djrobx Premium Member join:2000-05-31 Reno, NV |
to fg8578
There was a time back in the mid 90's when everyone thought copper couldn't support data rates above 56 kbps. In the mid 90's everyone thought that the POTS data limit would be 56k. Today, in 2014, the POTS data limit is still 56k. Once we reached the limit of what dialup could do we had to move to a different technologies (ADSL, Cable, Fiber) to make further progress. -- Rob | |
| | | | | fg8578 join:2009-04-26 San Antonio, TX |
fg8578
Member
2014-Jul-9 4:47 pm
Re: It's a startsaid by djrobx:In the mid 90's everyone thought that the POTS data limit would be 56k. Today, in 2014, the POTS data limit is still 56k. Once we reached the limit of what dialup could do we had to move to a different technologies (ADSL, Cable, Fiber) to make further progress. Understood. ADSL was specifically developed to get around the POTS limitations. Who knows what will be developed to get around this 10G, 30m limitation? Are you willing to bet that such technology will never be developed? I'm not. | |
| | | | | | ••••••
| | | | |
to fg8578
In the 90s, T1s and T3s generally were for the primo wealthy businesses. Few people knew what to do with such bandwidth, since it was a "geeky" endeavor to utilize such bandwidth at the time.
Nowadays you have 'babies' with cell phones, making (almost) everybody in a metro area the geek, if you will.
If such innovations ever arrive, it would be too late. Maybe someone will invent a car that drives 500 miles on 1 gallon of gas too. By that time(in addition to the price rising to make it unaffordable anyway), electric plug-in and hydrogen cars will go 300 - 2000 miles on the same "gallon's" worth (slightly exaggerated for effect). Fiber has this kind of scale of innovation and beyond. It's a shame the providers themselves are actively involved in discouraging deployment of fiber after getting so much money as a down payment only to have it squandered. | |
|
ev @74.140.91.x |
ev
Anon
2014-Jul-9 3:11 pm
ZzzMeanwhile existing coaxial can handle more mhz over far greater distances. | |
| | ••• | (Software) Sophos UTM Home Edition Ruckus R310
|
That's GreatTo bad it has the same problem that EVERY OTHER ADVANCE in copper transmission has. It's only good for absurdly short distances.
30 meters.... that's great... but what about people that are 1000 meters or more from the CO? This is where copper is a failure. It's good for voice, but that it.
To get any real progress we need to start getting the fiber closer to the houses. | |
| | tshirt Premium Member join:2004-07-11 Snohomish, WA |
tshirt
Premium Member
2014-Jul-9 7:01 pm
Re: That's Greatsaid by buzz_4_20:To bad it has the same problem that EVERY OTHER ADVANCE in copper transmission has. It's only good for absurdly short distances.
30 meters.... that's great... but what about people that are 1000 meters or more from the CO? This is where copper is a failure. It's good for voice, but that it. Because in many cases it can avoid the most expensive part of the build, that exclusive access from the curb to the house. It won't fit every need, just one tool for the last 100 feet. | |
|
|
15444104 (banned)
Member
2014-Jul-9 4:17 pm
Randall Stephenson ....Just build out the FIBER infrastructure will you !I agree that innovating is a good thing and this method can be used in SOME limited situations but when it comes to servicing homes and businesses they MUST start laying more and more FTTH and FTTB.
That is all.
What I would tell the shareholders regarding the expense to build fiber out entirely is .....those shareholders today should think of their children, and their grandchildrens' future as shareholders.
TO MAKE MONEY YOU HAVE TO SPEND MONEY! | |
| | ••• | |
very usefulif you plan to interconnect your home data center from your garage to your basement. I'm kidding. Still Waaaaaay too short to be useful | |
| Matt7 join:2001-01-02 Columbus, OH |
Matt7
Member
2014-Jul-9 10:19 pm
More copper milking. Just another hope for folks like AT&T who want to milk all they can out of the copper... I can see them using this go do '500/500' or something over copper and say "no one wants 1gig"
| |
| |
big_e
Member
2014-Jul-10 9:49 am
Cost and power consumption could be an issue.A 10GBase-T network card still costs around $400 dollars and this is for a network standard that has been established for years and has been mass produced. The key cost of this the sheer complexity and power consumption needed to do signal processing to push 10 Gigabit down 4 pairs of copper wire rated to carry 250 Mhz. Originally it took over 13 watts per port. Now years later they have been able to get it down to the low single digits.
Now think of the amount of signal processing to push 10 gigabit down a single pair of telephone wire designed originally for voice only. I think it would be an order of magnitude higher than 10GBase-T. If the cost is going to be initially over 1,000 for each transceiver, and it will need 20 watts per port there isn't going to be a very large cost advantage compared to going fiber all the way. | |
| |
wkm001
Member
2014-Jul-10 3:22 pm
Aww man...Only 15 years late... | |
|
| |
|
|