Slashdot Mirror
Flexible Optic Fiber Promises Cheaper Last Mile
bn0p writes "Ars Technica has an article on a Korean company that has developed a low-cost, flexible, plastic optical fiber that could bring cheaper 2.5 Gbps connections to homes and apartments. While not as fast as glass fiber, it is significantly faster than copper. In related news, Corning recently announced a flexible glass fiber that can be bent repeatedly without losing signal strength. The Corning fiber incorporates nanostructures in the cladding of the fiber that act as 'light guardrails' to keep the light in the fiber. The glass fiber could be as much as four times faster than plastic fiber. Neither fiber is available commercially yet, but both should help with the last mile problem when they are deployed."
If you don't have other reasons to dig trenches etc, then wireless is typically far cheaper because the installation costs are zero.
Engineering is the art of compromise.
Not in the United States anyway, our "last mile problem" has a lot more to do with entrenched telecom and cable companies with regional monopolies than any sort of fiber bendiness.
I don't think the cost of the actual cable will change the equation very much. I've been out of it for more than ten years, but even then you could get fiber for less than $1/foot - I assume it's even cheaper now. I have to believe most of the cost lies in planning, getting permits, and digging trenches.
What I don't get is why we seemingly refuse to invest for the long-term in the United States. Sure, some companies do, generally the smarter ones. But when it comes to public infrastructure, politicians haven't found a way to inform the public that by spending 2x as much now, we're saving 20x as much over the next n years.
I know that technology evolves at a rapid rate, but if we invest more money now and use the same amount of energy* now (compared to doing investing less money and the same amount of energy), then we can use the energy that's left over from not having to double our efforts next year for other causes.
*energy here is refering to human capital.
The abysmal state of what passes for "broadband" in the US has nothing to do with a lack of network technology for connecting the last mile.
This is a layer 1 solution to a layer 8 and 9 problem.
It's not the cable itself that's expensive, it's laying it.
If you mod me down, I will become more powerful than you can imagine....
I'd rather just run a couple copper wires with the fiber if you want a combined power/information source. PoE is something of a hack. You'd be able to push more power with a couple of dedicated wires, for lower cost - simpler construction because you only need a couple wires even if you increase the gauge, don't need to twist them, etc...
As for the other poster, I've seen those connectors - complex and expensive looking even(especially?) compared with fiber connectors. After a certain point fiber IS cheaper than copper.
Though to be honest, wiring new houses with cat6 today would be cheap, you could have a little switching center in a closet somewhere to decide which ones are phone lines and which ones are data lines. Stick the networking hardware in there as well. Running it through conduit would increase cost, but allow relatively inexpensive upgrades(like if fiber to the desktop becomes common enough to be in houses).
Still, I'd estimate that cat6 won't be unusably obsolete to the desktop for ~15 years. Gigabit ethernet is fast enough to stream at least a compressed HDTV stream. Of course, you'd technically be able to run HDTV over a 10mbit connection by my calcs at least.
I don't read AC A human right
The article writers and poster have no idea what they are talking about. Unfortunately most of us fiber scientists & engineers got laid off during the tech crash.
Plastic fiber has been around for decades. It is cheap. The problem with plastic fiber is that your signal won't go as far as with a glass fiber. However, for "last-mile" use, you don't need to worry about signal loss since you aren't going very far. The big cost in "last-mile" is digging up the ground and putting in the cable/conduit/fiber. The cost of the fiber is negligible compared to getting right-of-way and the cost of labor.
The cost of the fiber is so low, that normally when you dig up the ground to put in fiber, you put in lots & lots of fiber (since it is so cheap), just in case you need it in the future. This is called dark fiber, and there are millions of miles of dark fiber all over North America (from the tech boom) that used to belong to dotcom upstarts & their venture capitalists.
And as for bending fiber, you can always bend fiber. When you make very thin glass or plastic fibers for optical purposes, they are flexible. Has everything been running in straight lines?! Idiots.
Now, there is minimum bend radius for fiber, and if you bend your fiber beyond that, then you start to get some loss. Normally this isn't a problem, and you can't bend the fiber that far anyway - fiber has a cladding & outer sheath (which varies depending on the application indoor/buried/underwater), which limits the amount of bend, preventing bend loss.
I think you're correct - cable itself isn't the greatest expense - even the custom ends on the cables are fairly cheap, though more expensive than with copper as they are a bit finicky. But the installation is the expensive part. Civil utilities are installed in new subdivisions by government contractors by the local city or county in most places, but television cable and phone lines? I'm not so sure who foots the bill for that infrastructure. This doesn't even mention that installing new fiber in already existing subdivisions of single family homes has got to be expen$ive for sure. Telco and cable monopolies have little incentive to upgrade existing infrastructure in the last mile until current infrastructure has inadequate bandwidth for content. They're already running into some bandwidth contention during peak hours, which is why they are increasingly using throttling techniques and traffic shaping (like the forged RTS packets my own provider, Comcast, has recently gained so much noteriety for). That has turned out to be problematic for public relations and will rapidly provide diminishing returns, but I don't expect cable providers and telcos to do much until they absolutely are forced to do so. Adding customers adds to monthly cashflow; redeploying infrastructure doesn't. I suggest an actuarial decision will drive the deployment: When paid HD content outgrows the existing pipe in the last mile, the providers will build the infrastructure they need to reap those new revenues. So I expect content, not Internet traffic, will be the driver for the deployment.
Thats why when Verison gets over to your house and wires it up with fiber your cable company is not only going to drop your rates but also triple your download and upload.
How do you think progress is made? At any given point in time there will be one bottleneck in a system. Things progress by removing the bottlenecks one by one. You fix the slowest part and then move on to the next slowest part. Over time, the system as a whole evolves to become faster as its parts do.
If it exposes problems upstream then great! It means we have removed a bottleneck and the next worst one will be fixed. Otherwise companies will just say "well there's no demand from consumers so we don't need to improve our infrastructure."
Besides, fiber requires a transmitter for every line - it's a star topology by default, not a bus like coax can be. That increases expense.
I don't think that's right. You can use passive splitters to connect multiple devices to the same fiber line. Verizon does this for their FiOS service: it's how they connect 32 houses on a single fiber line.
Here be a wiki on the subject:
http://en.wikipedia.org/wiki/Passive_optical_network