Ordinary Copper Telephone Wire Could Carry Gigabit Broadband Speeds (newscientist.com) 129
Fibre-optic cable is being laid across the globe at great expense to speed up people's internet connections, but researchers claim that the copper telephone wire already in use across the country can achieve data rates three times higher than currently seen at a fraction of the price, at least over short distances. New Scientist: Their technique to boost speeds may help to ease the transition to nationwide fibre optic, and may also be of use in countries that use similar twisted-pair copper wire. Ergin Dinc at the University of Cambridge and his colleagues say that twisted pairs of copper wire, of the type used for decades as telephone lines and now repurposed for broadband internet, can support a frequency fives times higher than is currently used, which would dramatically improve data transmission rates. Above that limit, the researchers found that the wire essentially acts as an aerial and transforms any signal sent along it into radio waves that dissipate before reaching their destination. "These cables are actually very old, invented by Alexander Graham Bell, and since then no one has looked into the theoretical limits," says Dinc. He and his colleagues say that their findings may allow houses near fibre-optic cables to achieve higher speeds than they currently enjoy without the expense of running fibre all the way to their home. Fibre-optic cables carry groups of photons to represent data, and huge numbers of these groups can be sent along the line one after another without waiting for the first to arrive. Fibre connections in use today typically operate at 1 gigabit per second, but theoretical speeds could be many thousands of times higher.
One small problem (Score:5, Interesting)
A large percentage of these copper wires are already in a sorry state and need replaced anyway. I don't think the marginal cost difference of replacing last-mile copper with copper vs replacing copper with fiber amounts to much compared to the labor cost.
Re: (Score:2)
Still pretty ordinary... (Score:3)
Yeah, how "ordinary" are copper pairs any more?
AFAIK most new homes still install phone connections to a house, even though few use them anymore... even if that is dying off, there are a LOT of pre-existing houses and apartments that all have old twisted pair wiring going to them.
Last mile has always been the hardest hurdle so it would be great if you could use older wiring and get reasonable speeds.
Re:Still pretty ordinary... (Score:4, Informative)
When are we just going to bite the bullet and wire all homes for fiber?
Twisted pair is dead, dead, dead, and no variant of DSL is going to make it different.
Coax is damn near dead, what with its requirement for amplifiers, splitters, crimps, etc. It's a stunningly poor choice for data.
Just put in fiber.
Re: (Score:2)
You do realize that fiber also needs "amplifiers, splitters, crimps, etc". They just may be further apart on a fiber system compared to a coax system.
Re: (Score:2)
Coax is damn near dead
Not really. The "last mile" with coax was replaced a couple decades ago with fiber-to-the-neighborhood and much shorter coax loops. The bonus is that most of that coax is already installed and can carry 1 gigabit of data.
Fiber-to-the-neighborhood is also used in many modern DSL plants.
Have you noticed those big metal boxes mounted on utility poles? They are where the fiber terminates and the coax or telephone lines make their way to the home.
Re: (Score:2)
Coax is damn near dead, what with its requirement for amplifiers, splitters, crimps, etc. It's a stunningly poor choice for data.
Interesting comment given how coax is used for the overwhelming majority of broadband and cable systems at speeds well and truly exceeding what any family currently needs.
Please don't use the word dead to describe the living.
Also amplifiers, splitters and crimps? I see you've never actually had anything to do with fibre before, not only do all the above exist for fibre they are significantly more difficult to work with.
Re: (Score:2)
My house was built in 2016. No twisted pair at all. In fact, there's not even a dmarc on the side of the house, and I don't think anyone else in my small neighborhood has one either - we're only 8 houses on a street. I would be surprised if the phone company was even contacted by the contractor who built the place; there may not even be wire in the underground utility vault at the neighborhood entrance.
Definitely have cable though.
...about that DMARC (Score:2)
Given your house is fairly new, I'd imagine this should either be super simple to clarify or is well worth the escalation effort required to define that demarc. If nothing else, access required to reach 911 should be well documented given the technical hurdles discussed.
Re: (Score:2)
Re: (Score:2)
Problem is, can't dial 911 on a fiber line.
You can't dial 911 without fiber. Whether it's a fiber-backed cell tower or a copper landline terminated at a fiber-backed DSLAM.
As far as fiber to the home, you can purchase any telephone service over the fiber you want. Whether that's VoIP or traditional phone service from your fiber modem.
In the US, there is no federal mandate that unserviced landlines are required to connect 911 calls. And even in areas where it's allowed or required, it won't give location information and it can't allow a call back.
Re: (Score:2)
For what it's worth, I just bought a home that was built in 2016. There is no RJ-11 wiring at all in this house. There is, however, some ethernet drops in various rooms as well as a data panel in the wall where those runs terminate, with electrical in it so you can power up a small gigabit ethernet switch in there. No, you aren't going to get a 48-port rackmount in there, but the vast majority of people don't need that, and those that do can figure out how to patch one in via an existing cable from anoth
Re: Still pretty ordinary... (Score:2)
I'm pretty sure that a pair of cat5(e) to each jack instead of "phone" cable has been the norm since at least 2004. The cost difference between a foot of 3-pair cat3 and a foot of 4-pair cat5(e) is basically nil, and has been for years. Cat5e works just fine for POTS within a house, and is a lot more versatile over the long run.
The problem with speculatively pulling fiber in a new house is the 90% likelihood that it will be almost as obsolete by the time you finally go to use it as copper UTP. If using exis
Re: (Score:2)
Even if they can get gigabit out of it, it's just going to keep falling behind.
In major cities in Japan the standard is now 10Gb fibre, with 20Gb being available in some areas. It carries 8k TV as well as internet data.
By the time these guys figure out how to get 1Gb out of copper that speed won't be particularly impressive. It probably won't be universally 1Gb either, it will be like all the other DSL standards where the exact speed depends on the condition and length of the wiring.
It probably won't be sym
Re: (Score:2)
Another small problem (Score:2)
Here in the UK, most of the "copper twisted pair" phone lines are actually aluminium. The incumbent telco, British Telecom, switched away from using copper as a cost-saving/profit boosting exercise after it was split out from the GPO and privatised.
So, while this research is interesting in the abstract, it has little to no practical use in the country where the research is taking place.
As an aside, I'd also say that "These cables are actually very old, invented by Alexander Graham Bell, and since then no
Re: (Score:2)
In (the old) theory, the limit was based on frequencies under 1 GHz. This experiment deals with higher frequencies past a dead island of frequencies that won't work around 1GHz. The signal travels over the wire differently. Not likely to work very well with aluminum, though.
Re: (Score:2)
This. Australia's NBN tried that and it was a disaster. They had an anti-fibre policy, because politics, so they "upgraded" the network, found that it didn't work, and wound up replacing thousands of kms of copper.
no one has looked into the theoretical limits? (Score:5, Insightful)
Right. ATT, Verizon, Bell Labs, Cisco, Ubiquiti, on and on. None of these companies have looked into the theoretical limits of CAT3.
Oh wait. That's right. They did. That's how we got all the variants of DSL. And that's why we moved on to better tech, because CAT3 performance doesn't cut the mustard. They even admit the in TFA that these speeds could only be achieved over distances so short that they'd need to backhaul data closer to the customer.
This whole article is D-U-M-B.
Re:no one has looked into the theoretical limits? (Score:4, Interesting)
Everything the article says makes me think...hmm...useless for DSL but if you had 4 of these pairs it might be a good future Ethernet standard.
Re: (Score:2)
Everything the article says makes me think...hmm...useless for DSL but if you had 4 of these pairs it might be a good future Ethernet standard.
If you're only interested in going 100m, sure.
Re: (Score:2)
That's not much different than the limits they were getting in their tests anyway.
Re: (Score:2)
I have a cable modem sitting on my desk next to a router. The router plugs into my wall. It then travels to different rooms in the house, where I have televisions hooked up to the wall, because 4K over wifi is daft. Downstairs, next to the TV, I have a NAS containing digital movies.
All of that is easily within 100m and would benefit from that technology.
Re: (Score:2)
I have a cable modem sitting on my desk next to a router. The router plugs into my wall. It then travels to different rooms in the house, where I have televisions hooked up to the wall, because 4K over wifi is daft. Downstairs, next to the TV, I have a NAS containing digital movies.
All of that is easily within 100m and would benefit from that technology.
Why? Existing Ethernet and even Powerline ethernet would be far better alternatives. This is a solution looking for a problem.
Re: (Score:2)
Because maybe someone doesn't want to spend a lot of time / effort / money to run ethernet everywhere in a home they may or may not actually own? And powerline networking has great numbers, but never actually measures up, and sometimes doesn't work at all if there's surge suppression or noise filters between the endpoints.
I don't know why everyone is being so curmudgeony towards another solution to higher bandwidth connections which happens to reuse wire in millions of structures that is not being used for
Re: (Score:2)
Note that when my house was built, they just used cat5e for everything, including the RJ11 phone jacks. The subcontractor said it wasn't worth the effort to ever run less than cat5e, even for boring old telephony. This was nearly 20 years ago.
So it's quite possible construction in the recent past already has cat5e, and just have to change out the jacks.
Conversely, I don't know if this research shows anything practical to do for older construction that wouldn't be met by, say, MoCA, which uses commonly ran
Re: no one has looked into the theoretical limits? (Score:2)
Re: (Score:2)
This thread was spawned by someone saying essentially 'ok, so maybe it wouldn't work for long haul, but I could use my phone wire as home networking even though my house isn't wired, so maybe this research could be released as a home networking solution if not a long-haul technology. To quote:
"Downstairs, next to the TV, I have a NAS containing digital movies.
All of that is easily within 100m and would benefit from that technology."
So before accusing someone of being a dumbass, try verifying the context of
Re: (Score:2)
That's in the house, dumbass. They're talking about the other side of the demarc and the telco sure as fuck didn't run a million cat5e cables four lines at a time from the CO...
Did you actually follow this thread or just start commenting? You're completely out of context here.
Re:no one has looked into the theoretical limits? (Score:5, Informative)
DSL is a tradeoff between distance and speed. This is a slightly different tradeoff with more speed at a shorter distance, that was not of interest to the telcos when DSL was developed. However, this could well be of great interest to condo buildings, apartment complexes and the like. If you only need to pull fiber to one spot in a development that already has mostly-unused copper going from that spot to each individual apartment, it's a lot faster and cheaper than pulling new fiber to every unit.
Re: (Score:2)
DSL is a tradeoff between distance and speed. This is a slightly different tradeoff with more speed at a shorter distance, that was not of interest to the telcos when DSL was developed. However, this could well be of great interest to condo buildings, apartment complexes and the like. If you only need to pull fiber to one spot in a development that already has mostly-unused copper going from that spot to each individual apartment, it's a lot faster and cheaper than pulling new fiber to every unit.
Even then, it's just a stop gap for 5 or less years.
Re: (Score:2)
I would bet that the people in those buildings would be happy to have better service for those 5 years.
Re: (Score:2)
That's already being done in many, if not most, places. In the US, cable companies started doing fiber-to-the-neighborhood over two decades ago and DSL started a few years after that.
Some phone companies, like Verizon, opted to go for broke on fiber-to-the-premises, but they're not doing that anymore and sold off a few of their FiOS plants to Frontier many years ago. The only places where FiOS is still being expanded are DC and NYC because they sued Verizon to keep building FiOS.
Re: (Score:2)
They even admit the in TFA that these speeds could only be achieved over distances so short that they'd need to backhaul data closer to the customer.
Just what do you think FTTC and FTTN are? The whole point of this is to extend the last very short distance from fibre to the end user. Nothing dumb about it, a technical solution to a fucked up fibre rollout that never bothered to do that last little step needed in the name of saving costs which unfortunately is very relevant in a few countries.
Re: (Score:2)
Fiber to the curb allows different techniques than fiber to the node.
The DSL techniques have been invented for fiber to the node.
Re: (Score:2)
Which isn't to say I see much advantage of trying to dig up phone cable at the curb. They can already tunnel fiber horizontally to homes underneath yards quite efficiently, so there's little point.
Re: (Score:2)
There's a market you're not thinking of though - this could be used to great effect by running fiber into older multi-tenant building that has pre-existing telephone wiring, and using this to give everyone in that building a somewhat decent connection without having to run drops into each and every unit. Put in a single fiber drop in a utility closet, bridge it to the building telephone wiring, and light it up.
Apartment and small office building owners might love something like this where they aren't going
Re: (Score:2)
distances so short that they'd need to backhaul data closer to the customer
...which is what modern coax cable internet already started doing with "fiber-to-the-neighborhood" decades ago.
Plenty of work has gone into UTP limits (Score:5, Interesting)
We can push a lot more over UTP than we do today. However, it's not easy to do over significant distance (which would be needed) and even at short distance, it's a power hog to try to push it that hard. To try to use copper, it would require digging up a bunch of it anyway, get power to more places along the line to actively repeat, and then probably fail due to the general decrepit state of the wiring.
Re: (Score:2)
significant distance (which would be needed)
Why? The point of this is not to run internet over your phone line back to your phone exchange ala DSL. The point of this is to do the last hop from FTTC or FTTN. The distances are not significant. In some cases we're talking about 10m.
PopSci explanations (Score:5, Informative)
It's gotten to be an art trying to understand technically what these popsci articles are describing in layspeak.
Anyways, here's a direct link to the paper: https://www.nature.com/article... [nature.com]
Basically, they're talking about matching a standard telco twisted pair with a balun, and find they can keep it behaving as a transmission line until 5GHz, and higher frequencies by increasing the twists.
Re: (Score:2)
Basically, they're talking about matching a standard telco twisted pair with a balun, and find they can keep it behaving as a transmission line until 5GHz, and higher frequencies by increasing the twists.
None of what the authors conclude is particularly surprising. Twisted pair is basically a variant on the standard two wire transmission line and they demonstrate that if you can make a decent broadband balun then you can get energy in and out of it.
However, it seems to me that the paper has demonstrated a theoretical possibility whilst overlooking a range of practical problems.
2nd link (Score:2)
Re: (Score:2)
Dedicated data use (Score:5, Interesting)
Unless I'm mistaken, one of the main technical limitations capping data rates over twisted pair telephone lines is to preserve the quality of legacy analog voice communication (IE what the lines were originally meant for 140 years ago). That is preserving analog phone on the pair the data is on, as well as preventing it from bleeding over to the adjoining pairs of copper that are bundled together.
It seems to me if that was not a concern then data rates could use that bandwidth as well, thus increasing data bandwidth.
Re: (Score:3)
Maybe it's time in some regions to do what they did for TV, switch to digital.
Give people a few years to buy a converter box or replace their POTs equipment and then re-cast the infrastructure as digital only
Re: (Score:2)
While we're putting digital transceivers in the field for anyone looking for PSTN service, let's just convert the whole thing lock-stock-and-barrel into IP connectivity and give them a VoIP transceiver instead. I mean, that's only been around for like 20 years now, and packets are packets.
Let's stop pretending that analog voice that has had the upper and lower frequencies hacked off is anything more than it is, and start treating it as real-time streamed data like any other real-time streamed data.
Re:Dedicated data use (Score:5, Informative)
I don't think that's actually a problem. Here's how the spectrum is allocated:
https://en.wikipedia.org/wiki/... [wikipedia.org]
Annex A and M are the ones you'll usually see when it's combined with POTS. Annex B allows it to coexist with ISDN.
Even with the generous guard band up to 25kHz - high enough to prevent crosstalk from being a problem - analog voice is less than 2% of the available spectrum.
The only time it would matter is if you're on a very long line (living in the sticks with traditional DSL to the CO, instead of neighborhood DSLAMs that most areas have). In that case the upper spectrum is progressively lost, and at some point you might want to reclaim the bottom end (I or J), or boost the power of lower frequencies while cutting high frequencies to stay in the required power envelope (L). Very few places deploy these though because it increases crosstalk and it only helps if you're way out in the country - the bottom end is a rounding error unless your line is already terribly attenuated.
Queue the hams (Score:3)
Not to mention pissing off a lot of ham radio operators.
Sigh (Score:2)
Re: (Score:2)
also reduces the impact of power outages..If you and the CO have power, you have connectivity...If they have copper that needs repeaters every 500 feet, that's a lot of active failure points and power outages along the way (we KNOW how battery maintenance goes) can kill everything.
Also much of the existing copper plant (at least in the USA) is 50-70 years old, well beyond its lifespan, between weather, rodents, electrolysis, and other stresses, many areas have very fe
Most copper plant needs attention anyway (Score:3)
Most of the copper pair infrastructure is well past its lifetime, at least in the USA, technicians keep running out of good pairs in some segments...weather, rodents, etc. have done their damage over the last 70 years.... replacing it with fiber is cheaper than replacing the copper
Also, reading through the article, the longest distance I see mentioned is 300m (about 1000 ft in the USA), the need isn't in the 1000 foot drop to the premise (heck, if nothing else, one could AirFiber that), but rather the 1-10 miles to the CO where the real bottle neck is. /P
Re: (Score:2)
Ah, the old There's Gold in Them Thar Copper (Score:2)
https://www.blog.adva.com/en/t... [adva.com]
I have tried many of the point-to-point magic black boxes that hook to the end of copper pairs and guess what?
FIBER IS BETTER.
Anything over a meter's length worth of cable and you are back to 100 mb/s.
The copper is not twisted, suffers from crosstalk, and capacitance over long runs. Also, getting dry pairs in a city point-to-point is difficult.
Re: (Score:2)
The copper is not twisted, suffers from crosstalk, and capacitance over long runs.
Telephone wires have been twisted since some times in the 19th century. It keeps out magnetic interference at audio frequencies, and minimises crosstalk. Capacitance over long runs is dealt with by appropriate impedance matching, which is around 110 ohms for twisted pair. In the old days of wired telephony, the nominal impedance was 600 ohms, so loading inductors were need to make the cable look like 600 ohms impedance over long distances, but only for audio frequencies.
The trouble is, there is no way that
Have we forgotten? (Score:3, Interesting)
In the local box outside the house we've got aluminium wires mixed with copper simply twisted together.
These effectively act like terminators at high frequencies.
These researchers are idiots. (Score:2)
No one disputes that wire on it's own can do great speed. The connection points and constant splices from repairs are the problem.
Re: (Score:2)
fiber to the node (Score:2)
But it doesn't work long distance (Score:4, Insightful)
You don't need to be a research scientist to figure out you can get some really fast speeds over unshielded-twisted pair using ridiculously high frequencies. But the higher the frequency, the shorter its propagation distance.
DSL is already on borrowed time, and the last thing we need are absentee-landlord ISPs like CenturyLink to eek out a slightly higher performance for urban residents while rural residents suffer from a lack of sufficient broadband for another decade. Out here in rural Minnesota, there are many households that are still stuck on DSL on speeds of 1 Mbps or lower because they're too far away from the PBX. What we need is a full investment in fiber infrastructure nationwide.
copy is rotting in the ground (Score:2)
Copper is to be completely sunset by August of 2022. Trying to use copper now is attempting to use garbage that has been ignored and not maintained for almost 7 years. This is too little, too late and really stupid.
Twistd pair? Wanna bet? (Score:2)
One big problem is that telcos, trying to support voice, didn't use twisted pair. They used pairs that are unshielded and not usually twisted. The transition to the use of TP for the local loop didn't start until the late 1980s, so newer neighborhoods might have it, but older telco wiring was plain balanced pairs. Yes, a bit of cross-talk was common, but people were used to it.
Getting DSL to move it's specified max of 24 Mb/s seemed pretty amazing, but was seldom (if ever) achieved and that came only in re
Twisted pair from the telco? (Score:2)
Bell invented cables? Who knew? (Score:2)
56k and worse (Score:2)
So we had to struggle with 56k (and worse!) analog modems for YEARS, and now they want to try improving the speeds over old-ass tech?
Fuck right off. There's better technology out there and we've already moved on.
Re:Could, but it's shit (Score:5, Insightful)
For many people there still isn't a better alternative, unless it's purely wireless / satellite based. The wires are there, it makes sense to make optimal use of them.
Re: (Score:2)
In many cases the copper wire comm will be unidirectional and you'd get a very limited performance in the other direction.
And in addition to this many copper lines are now already torn down so this tech is 15 years too late.
Re: (Score:2)
Depends what part of the world you're in and your specific ISP (or the infrastructure they're leasing). The last mile of telephone wire vs. coax is only half the story - here in my part of Germany, the cable network back-end is pretty flaky (ping & jitter issues, especially when you're stuck on a 6-to-4 connection and things like gaming servers are stlil IPv4 only). The (V)DSL / jank-ass-phone-wire network on the other hand, while slower in terms of overall bandwidth, is rock solid because the infrastru
Re: (Score:2)
Cable is a shared loop, so you are more susceptible to what your neighbors are doing. Some shoddy equipment that creates noise can mess it up for everyone in the area.
Giving dedicated IPv4 to users is not cost effective anymore, even if the ISP has enough address space from an old legacy allocation it would be more profitable to sell the address space to the likes of amazon or microsoft. You will end up sharing a single IPv4 address with any number of other customers, but at least the ISP also provides IPv6
Re:Could, but it's shit (Score:5, Interesting)
You don't seem to understand how DOCSIS actually works. That fat coax cable the cable co. brings to your house is the least part of the system.
I would think that having a single wire, rather than a bundle of wires that exhibit crosstalk with one other randomly, would be a rather big part of the system. Cable just has to have termination to prevent reflection. Phone would likely have a huge multipath problem, which would be a much bigger nightmare.
But let's go and see if copper really can be made to work. Remember when DSL wasn't capable of more than 56k baud?
No. I don't remember DSL ever being that slow. Analog modems were that slow. But ISDN was faster than 56k, and predates DSL. :-)
The problem with high speeds on phone lines is the range. Being able to send a high-speed signal 10 meters is fine if you have a single pair of wires (no crosstalk) and you're only 33 feet or so from the pole, but in practice, it isn't useful. The average local loop length for areas that don't have FTTN is a couple of miles. Even if you could provide gigabit data for half a mile, that's still a *lot* of fiber to light up every neighborhood, and you're still pulling FTTN to all the neighborhoods that don't have it yet.
And remember that most wealthier areas already have FTTN, so the only places still using DSL en masse are typically areas where the neighborhood isn't wealthy enough for the phone company to be willing to run fiber and shorten the local loop. This doesn't help with that problem at all.
IMO, it doesn't seem like a very good use of effort trying to squeeze extra performance out of legacy infrastructure that in many cases is barely even functional at DSL speeds. We would be far better off making a commitment to bring fiber to every home so that we have infrastructure that is modern and reliable.
Re: Could, but it's shit (Score:2)
My first ISDN line was a BRI. When DSL first became available in my location it wasn't as fast.
Cable blew them away quickly. Then the DSL innovators gained.
DOCSIS uses subscriber loops in the coax plant, sharing bandwidth. Fiber no doubt uses switches, etc, it's just so much faster to premises, or can be.
Re: (Score:2)
The first iteration of DSL was capable of 8mbps on a good line.
Wether your telco configured it that way, or wether your line was short and decent enough to handle it is another matter. A lot of early deployment were intentionally limited to low speeds such as 128k as it still represented an improvement from dialup, didn't cut into sales of higher priced dedicated lines and at that speed was still stable even on quite poor lines.
Re: (Score:2)
Usually FTTH is run over GPON. The most common standard is still 2.5G down / 1.25G up shareable by up to 32 nodes. There is also 10G GPON, 10G down, 2.5G up, but I have no idea how common it is. My local FTTH carrier is on a 2.5/1.25 system selling symmetric 1G service. Still plenty fast for residential, but definitely a shared medium. Have to control costs somehow. Residential customers don't want to pay traditional metro-e costs.
Re: (Score:2)
From my experience, the major costs for FTTH are the actual fiber runs, the drop, and the interface, with the shared infrastructure being big numbers shared among subscribers, and of course interconnect.
Most all the FTTH I've seen uses a network interface that provides the subscriber with 100MB Ethernet, with relatively recent updates to GbE and better. Your interface caps the transfer rates, advertising 10GB to a residential user implies their NI is 10Gb. YMMV.
GPON sure seems similar to the old metro ether
Re: (Score:2)
Next up, data over power lines.. Yet.. Fucking.. Again
Re: (Score:2)
The phone companies are not ripping out the last hundred feet of copper to your house, unless you are buying something like VOIP from them to replace it.
Re:Could, but it's shit (Score:5, Insightful)
It's not necessarily the wealthy areas that telcos target.
A wealthy area might contain tens large homes, each with a sizeable garden. But an area the same size in a poor area is likely to contain hundreds of small apartments. Given that Internet access is now pretty much a necessity, even if you assume that the wealthy subscribe to the most expensive plans and the poor buy the cheapest (or government subsidised), you still make up for it in volume.
Re: (Score:2)
My local phone company has fiber in the main street, buried copper from there about 500 feet into my house. (actually two lines, but we only use one) When the cable company's signal failed, they said it was on my side, and they would gladly extend a new cable to my
Re:Could, but it's shit (Score:4, Insightful)
Re: (Score:2)
Remember when DSL wasn't capable of more than 56k baud?
No, and despite your claims to the contrary below, neither do you. The oldest and slowest ADSL standards were capable of 8mbit downstream at 2km. SDSL was 1.544mbit (that line rate sound familiar?) Even the original DSL patent [google.com] had a floor rate of 160kbit at 18,000 feet.
Re: Could, but it's shit (Score:2)
The standard might have allowed for it but for years there was literally nobody in the US deploying it with more than 1.5 Mbps. The copper is too shit. As it turned out most telcos couldn't even provide that reliably.
Re: (Score:2)
All of the above is true, but doesn't change the fact that "DSL was once only capable of 56k" is a falsehood.
Re: (Score:2)
I wonder if they weren't getting confused with ISDN somehow. While each of the two ISDN channels was supposed to be 64 kbps, in some cases it could only support around 56 per. I forget the details, I hardly did any ISDN.
Re: (Score:2)
Eh, he's claiming he had a BRI and that when DSL was first offered in his area, it wasn't as fast so presumably knows the difference. The LEC may have offered a 128kbit version (many did) but even really shitty copper that hasn't been groomed in decades should support at least 256kbit so at best he's simply conflating "we offered this speed" with "that's the most it was capable of."
Re:Could, but it's shit (Score:5, Informative)
The thing I'd worry about the most is when you try to use wire for high speed digital that wasn't DESIGNED for it. Specifically, the shielding. This is where technologies like "broadband over power lines" generate incredible levels of radio interference, across the entire town. Then after practically rioting to GET bbopl, the people complain about things like their cell phone coverage dropping. (the ham radio operators are usually the hardest hit, and the first ones to sound the alarm, that goes mostly ignored)
Re: (Score:2)
Re: (Score:3)
"Good enough" is the enemy of progress.
If the most you could get from copper cable was 56k, then lots of people would be loudly complaining and the telcos would be pressured to upgrade their infrastructure to fibre.
However the technology for faster data over copper is not perfect, it only works for users with short lines in good condition. So what it does is decrease the number of people who are complaining, which directly hurts those who don't have a short good condition line because they are now part of a
Re: (Score:2)
Bullshit.
If the most you could get from copper cable was 56k, then lots of people would be loudly complaining and the telcos would be pressured to upgrade their infrastructure to fibre.
You just disproved your own points. There was a time when we could only get 56k out of phone lines. The telcos didn't lay fiber down, which disproves your point. Instead they went to digital and got broadband speeds out of the same wires. Now that they may be able to utilize the same wires for even higher speed, you're back, complaining. This is not about "good enough". Nobody is saying 40mb/s is good enough. Instead they're saying it's NOT good enough, and the easiest way may be to utilize
Re: (Score:2)
So instead, people should just wait for telco providers to get around to deploying those "better alternatives" when they get around to it?
This sounds a lot like "I've got mine, everyone else can get fucked" - why do you give a shit if there's another way to get better access to people than they have now, using shit that already exists?
Re: (Score:3)
The telco copper infrastructure is at least 50 years old in most neighborhoods and has not been maintained. It's full of splices, ad-hoc taps, water damage and corrosion. The only answer is to replace it with fiber. I know, I've suffered through years of crappy DSL. I finally have fiber and there is no comparison. This technology is a bandage on top of a huge scab.
Re: (Score:2)
Why? They want to save easement and right-of-way expense but still compete with fiber, shielded coax, 5G, Starlink, and whatever comes after that.
As you cite, the research is based on optimal circumstances, experimentally. In the real world, leaky RF is a HUGE problem and the state of in situ cabling is often horrific, made worse by decades of neglect and Shareholder Profits.
I'm glad there was research. I'm glad it proves that the basic theories about twisted pair technology still holds water.
And it's not t
Lol (Score:2)
"Ordinary Copper Telephone Wire Could Carry Gigabit Broadband Speeds"... almost a whole meter.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
There was a place in Germany doing 230Mbps on DSL in developments. The DSLAM was fed by fiber.
In rural US with similar distances and brand new wire they wouldn't give me more than 7Mbps, hard capped to 7.00000 (even though the modem synched at 12).
Guess how much competition they had!
Re: (Score:3)
It is actually rather expensive to replace copper wire inside the walls of a house or apartment building with fiber optic cables. The only real way to "wire" a house with fiber is to run a wire to each room that needs a connection. If the room has no other way to get the cable there you need to rip into the walls to run the cable. Those same houses more likely than not already have phone cables running to each room that would need a fiber run.
So, if you are comparing running new cable to using existing cabl
Re: who cares about short distances ? (Score:2)
This. I have fiber to one location, and then Ethernet to where I need it. If the house could be wired for high speed data using the phone lines in my walls, that would be great.
Re: (Score:2)
Which is why when you're running wires behind walls (and above inaccessible ceilings) you should always use conduits with a little spare capacity so you can pull new wiring later without ripping
Re: (Score:2)
Typical distance is over several kilometers for phone lines. They talk about 100m and 300m in the linked article. It may still be fast enough above that, but no mention.
Re: (Score:2)