Slashdot Mirror
Plastic Optical Fibre: Cheap and Bendy
Motivator_Bob writes: "The Sydney Morning Herald has an article on making optical fibres from plastic rather than the traditional glass."Advances in optical-fibre making at the Australian Photonics research centre could bring communications at the speed of light into Australian homes and businesses in the next few years. The advance - microstructured polymer optical fibres (MPOF) - allows the manufacture of optical fibres that are much smaller, cheaper, more rugged and easier to make than glass fibres..."
Because the cable company wants to sell you kable modem service, the telcos dont want to give up $1000/mnt T-1 contracts, and your city wants ISPs and telcos to pay exorbitant franchise fees.
The biggest trick the devil pulled was letting lawyers become politicians so they can write the laws.
So let me get this straight; all the dark fibre we have in the states is now obsolete and therefore useless? Great, thanks, just checking.
"Don't bother lighting it up now, boys, just chop it up good when we start laying the new stuff."
-- "Government is the great fiction through which everybody endeavors to live at the expense of everybody else."
Interesting quote:
Maybe I'm just naive (probably), but the limiting factor today for broadband Internet access is the cost of the bandwidth, possibly due to the stranglehold a few key companies have on access to their backbones. The cable that comes into my house can be used for speeds in excess of 30Mbps, if I recall correctly, yet I have a mere 1.5-2Mbps (at $39.95/mo). Admittedly, DSL has technical limitations on speed, but even so, the large limiting factor seems to be the cost of an OC-12/48/96 connection to the 'Net, right?
When is that gonna change?!? What is needed to bring about that change? Regulation?
Exec 1: Gee guys, with this cheaper fiber, we could roll out much better speeds than what we get on the copper we use now!
Execs #2 & #3: Woo-hoo, that'll really help us get a leg up on the competition!
Exec #1: Oh, wait... We don't have any competition. We don't have to share our lines with anyone, so no one else can get their foot in the door here. I guess we'll have to bonus our expansion money out to ourselves, instead.
Exec #2 (holding plastic fiber up to his eye) : Hey, Dick, I think I can see you through this thing. Neato. Somebody get me a martini.
This tagline is umop apisdn.
Single-mode fiber has a mixture of materials that have varying indices of reflection, so that the cable is a light guide instead of a light tunnel. This allows for a shorter path because the light is kept closer to the middle. Btw, the latency of copper is much, much greater than typical single or multimode fiber because of capacitive and inductive coupling. Fiber mainly has the advantage of higher bandwidth and noise immunity, but it wont ever reduce you ping beyond (distance / c).
The biggest trick the devil pulled was letting lawyers become politicians so they can write the laws.
Unlike photons, electrons have mass. Nothing with mass moves at anything close to the speed of light.
What is the speed of electrons down a copper wire?
I'll admit that I didn't do the math to re-check this, but I seem to remember the velocity of electrons in copper wire being on the order of a few cm/second - much less than the speed of light. The confusion my be coming from the fact that when you stick an extra electron in the end of an otherwise neutrally charged wire, the spare charge sets up an electrical field that pushes a different electron out the other end (assuming it's grounded or generally has another place to go to). It's the electrical field that travels at the speed of light, not the electrons themselves.
>In fact, as light in fibre optic cabling bounces off the insides of the plastic tubing, it takes a less direct route and thus technically has a _higher_ latency than copper wire.
Ummm.. not quite. It's true that the light bounces around inside the fiber, but due to the low index difference between the core of the fiber and it's surrounding cladding the angle of the bounce is pretty small and wouldn't really increase the distance the light needs to travel The distance increase is proportional to 1/cos[angle] so when the angle is near zero, cos[angle] is near 1 and 1/cos[angle] is pretty near 1 meaning no big change in the distance traveled by Joe Photon. For electrical wires, speed is limited by the capacitance/inductance ratio of the cable and is typically around 2-3 times slower than free-space light.
All in all, it's a good thing that electrons don't go the speed of light in our house wiring - I used to work with a synchrotron, which is a device that gets the electrons moving at relativistic speeds, and whenever the beam of elecrons went around a corner it produced enough X-rays due to the angular acceleration to flash-fry a horse. Be glad that copper wire electrons are slow, since if they were fast we'd get cooked every time a bit of house wiring was anything less than perfectly straight.
I just read that telecoms have an excess of long haul bandwidth, which means that the issues are the cost of the last mile and consumer uptake.
I wouldn't be surprised if the main issue is the latter. I have spent some time trying to convince one of my coworkers at a major computer hardware company to get broadband, but he doesn't think he needs it. The uses of broadband are not necessarily obvious if you don't have it.
I also read that several telecoms will try to address this issues by selling capped broadband at a lower price.
Until we get the internet it's self running in gigabit (1000megabit or better) and have buttloads of bandwidth out there ala the full scale release and switch to Internet2 and ipv6, and replace all the routing and switching gear to the new gigibit or higher stuff it means nothing.. too many people are spoiled with T1 or better speeds into their homes while many many businesses have a 256-512 Kbit connection MAX due to the huge costs with a real net connection. (that is approx $1000.00 a month.. with yout fractional T1 and ISP access costs...)
... but I doubt that it will impact the costs of the fiber it's self, termination or maintaince.
fiber into the home.... WHY? is it needed? no.. will it be needed ? not for at least 10-15 years. and it wont be useful to anyone for a lot longer than that. The cost of laying fiber is not the cost of the fiber... it's the cost of directional boring or the manual labor to install it... regular old "expensive" glass fiber is dirt cheap. and most places lie down 24 or 48 count fiber when they only need 1 or 2 of them.. as the cost difference is minimal. (plus you can make gobs of cash selling the dark to other companies)
Plastic fiber is a neat idea
Do not look at laser with remaining good eye.
The relevant quantity in fibers is the critical angle, beyond which all light is reflected inside the higher-index core. (Actually, the whole ray-optics picture is not completely accurate for fibers with features, like the core size, comparable to the wavelength...but it's qualitatively the right idea.) (Which, by the way, has nothing to do with the reflection disappearing from the puddle, since that is a reflection into the lower-index medium, air. The puddle effect has more to do with your shadow blocking the light.)
Note also, by the way, that it's not so much that the index of the polymer fiber core has been increased, its that the effective index of the cladding is decreased (by adding lots of thin holes/veins, hence the name microstructured fiber). And you can do the same thing with glass fibers. (Because of the higher effective contrast, you can confine light more tightly and e.g. enhance nonlinear effects.
(You were on the right track that it's the bending light loss, and the advantage therein of higher index contrast, that the article was referring to.)
Microstructuring can also go in the other direction to photonic crystal fibers and guiding light in air.)
If a thing is not diminished by being shared, it is not rightly owned if it is only owned & not shared. S. Augustine
Actually, electrical signals in the neighborhood of 10-100MHz propagate through copper at about 0.1C or (18,600 mi/sec or 30,000 km/sec).
That depends on the geometry. Use thicker copper and/or space it farther apart and the signal goes faster. A less lossy dilectric helps, too.
At low frequencies - i.e. where the stray resistance of the line's copper is small compared to the characteristic impedence - the speed is dominated by the dilectric constant of the space between the conductor - and you get your approximate 70% of lightspeed. At higher frequencies (or longer wire) the line acts progressively more like a series-resistors-parallel-capacitors delay line cum low-pass filter. This slows and attenuates the signal, the higher frequencies more than the lower ones.
Selective slowing (phase shift) of the higher frequencies smears out pulses, while selective attenuation weakens them compared to noise.
This can be compensated for to some extent (by amplifying and phase shifting the higher frequencies before transmission and after reception). But there's a limit to how much of that can be done: Too much at the transmitter and you exceed the allowable signal level for the wire (causing cross-talk into the weaker signals going the other way nearby). Go far enough out and the high-frequency signals get down near the noise level, so amplification at the far end just jacks up the noise, too, and they're lost. That's why DSL will only go so far (without a repeater/regenerator).
Telephone wiring was designed for audio of only a few kilohertz, distances of a medium-sized town (rural wiring is a special case), and MANY wires in a bundle. So it uses very thin copper. Central offices were spaced in urban areas so that everybody they feed would be close enough to get a good audio signal. But DSL uses higher frequencies which peter out closer to the source.
Within the distances and frequencies where a copper structure will act as a transmission line rather than being ruined by this effect you're still talking about 70% of c.
But when the poster said "propagate through copper" he MIGHT have been talking about the "skin effect". Eddy currents in the copper due to changes in magnetic fields produce a compensating field, and the result is the field doesn't enter the copper until the eddy currents die due to the copper's resistance. (That's why magnetic fields won't enter a superconductor - to a first approximation.)
But that confuses "propagating through copper" with "propagating along a copper transmission line". In a transmission line (or any other waveguide) the signal and energy don't propagate
through the conductor(s). They propagate through the SPACE BETWEEN the conductor(s). Raise the resistance of the conductors and you increase the speed of penetration of signals into the conductor, but slow its propagation along the line.
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
I just read that telecoms have an excess of long haul bandwidth, which means that the issues are the cost of the last mile and consumer uptake.
Yes it's the lack of the "last mile" - and content worth paying for its instalation.
Lots of spare fiber (and empty conduit) was laid when the trenches were open, so most of it is dark. Boxes were bought to light up a few fibers, and even when they're all lit we can bump the speed to get a few more powers of two before stringing more long-haul.
But the network speeds and capacities of the first boxes were calculated using what turned out to be Netcom's overstatement of the rate of growth of the internet's bandwidth. For the last 5 or so years it was only doubling, rather than multiplying by 10.
Doubling every year is no slouch for a growth rate, but it's only about 1/3,125 the traffic the designers of the equipment and networks were planning for at this point. (It was 1/125 at the time of the dotcom bubble burst. Maybe some of those dotcoms WOULD have been profitable if the customer base they'd been told to expect actually existed?)
So there's a bandwidth price war at the wholesale level, telecoms folding up as debts come due without revenue to pay them, and equipment suppliers having a REALLY hard time selling any more stuff.
But with the CLECs pretty much all dead, the ILECs and cable companies (with the pre-installed base) have a virtual duopoly on the last mile. So there's no incentive to push cheap fat pipes into your hands. (Markets need THREE suppliers before competition starts driving costs toward price of production. With only two they'd be cutting their own throats to try to cut each others'.)
So there's no cheap last mile bandwidth. But there's virtually no high-bandwith content available to make it worth peoples' while to buy expensive last-mile bandwidth:
- CARP killed "internet radio".
- The RIAA killed Napster, is killing its clones, and finally going after individuals.
- The RIAA and MPAA are scared spitless of allowing any of their members' digital content on the net, for fear of piracy.
So what does that leave Joe Sixpack that will convince him to pay enough extra for high-speed internet that it's profitable to dig up his street and give him a fiber? Better animated popup ads? Most of the rest of the net is more than adequate at moderate speeds.
High-speed internet will be here as soon as there's a "killer app" requiring high-bandwidth that's popular enough to fund a new last-mile deployment, or a cheap-enough last-mile solution is found to be price-competitive with cable and ILEC-based DSL.
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way