| | Trapping LightPage 1 of 1 | http://discovermagazine.com/2001/apr/feattrap
The future sounds amazing if we can learn to trap light and use it in computers and other components instead of electricity. The speed of data transmission would be unparalleled and the energy costs should be lower. | |
|
| Trapping Light
Posted: 2/22/2012 12:30:02 PM | Ya, I heard a little bit about this a few years ago. People are trying to creat computer processors that function by passing light through compressed Noble gases. If I remember correctly, If they are successful, these computers are supposed to have something like ten thousand time the processing power of conventional computers. Really cool stuff.
What I am waiting for though, is for us to start mining the moon for helium three. I would so buy a car run off a helium three reactor. | |
|
| Trapping Light
Posted: 2/22/2012 1:28:05 PM | | It's expected to speed up the internet immensely as well as other uses in nanotechnology (with lasers in medicine,etc.). The link I posted mentions using photon crystals, but the article was dated 2001 so I'm sure there has been much progression since then. Still sounds exciting though. | |
|
| Trapping Light
Posted: 2/22/2012 4:13:55 PM | There will be a whole wave of new technology coming. | |
|
| Trapping Light
Posted: 2/22/2012 4:43:33 PM | When it comes to the net I didn't see them mention anything about repeaters in fiber - it automatically came to mind as a possible barrier.
http://en.wikipedia.org/wiki/Optical_communications_repeater
http://en.wikipedia.org/wiki/Repeater
Did I overlook how they expect to factor those in or are they then not needed.
As for using computers with gasses inside holding in light - that would pretty much crap on anyone wanting to put together their own computers. I can't even imagine what happens to heatsinks if they can't be exposed or release the gases if you made a closed system.
This still sounds like infancy planning so I guess those things are secondary stages. | |
|
| Trapping Light
Posted: 2/22/2012 5:31:39 PM | veevee:
I have no idea about the specifics - it was a short article and it was a few years ago. I tryed to dig up some info on it before I wrote that but couldn't find anything, then again, I only spent about ten minutes trying so...
I think it would be cool if someone could dig up the info though. It sounded like really promising stuff. I can't imagin that it simply dropped off the map. | |
|
| Trapping Light
Posted: 2/22/2012 8:01:12 PM | | Quantum computing is about 20 years away and it offers speeds much faster than this. Not much point in pursuing one stream when the other stream is going to leave it in the dust. | |
|
| Trapping Light
Posted: 2/23/2012 3:46:03 AM | | Technology will not advance so much in computers. The internet is simply addictive at this point in time. If you think computers will advance you might want to buy facebook stock, but there is a reason they are finally selling. | |
|
| Trapping Light
Posted: 2/23/2012 7:08:50 AM | veevee:
"As for using computers with gasses inside holding in light - that would pretty much crap on anyone wanting to put together their own computers."
Unless chemistry is a hobby of yours lol. | |
|
| Trapping Light
Posted: 2/25/2012 12:55:45 AM | Really can't see light taking the place of electricity in computers. In fact electricity based computers could be a whole lot faster than they currently are.
- All too often networks are set up for the least common denominator, which currently is 10mbits. More common than 10mbits is 100mbits (almost all networking tech in the past 20 years is 10/100, meaning it can do either or). Networking tech made in the past 5 to 10 years is 10/100/1000. There is a massive problem with router technology in the 1000mbit range. Currently there is only one home router that supports 10/100 AND 1000 at the same time, all the rest instantly downgrade to 10/100 if any devices plugged into the network are not capable of 1000.
- ISPs rarely ever use jumbo frames, even though there is not a single computer in use that can't handle them. Switching from the standard frame size to jumbo frames nets a massive improvement not just in speed of the network, but also reliability, error control, and more efficient bandwidth usage. It's tradeoff is slightly more processing time, but you'd have to be using a computer from the 80s or early 90s to notice it.
- RJ-45 cable is usually ran with the cheapest available, which generally comes from home depot at $45/500ft for riser cable. Riser cable is brittle and easy to crack/break if it's not handled with care, all too often a network tech will bend the stuff at 90 degree angles so that it fits snugly against a wall or rail. For about $15 more you can get 500ft of high grade RJ-45 which has grounded shielding, much more effective especially on a 100 or 1000mbits network, and especially around devices which emit larges amounts of em noise (fluorescent lights, CRT monitors/televisions, microwaves, refrigerators, electric heaters, etc).
- Majority of networks out there are using switches which don't support very many concurrent connections. This means you get a couple high bandwidth services running and the entire network bogs down instead of just the segment those services are running on. A typical cable modem can support upwards to a thousand (sometimes more) concurrent connections, yet the rest of the gear can barely support 20 before packet loss becomes an issue.
- All too often networking equipment which requires somewhat heavy processing uses a processor in the 200 to 400mhz range, which was designed at the around time the SNES was released. Despite 1ghz and higher processors being relatively cheap these days. Even my expensive home router only uses an 800mhz processor, and it currently yields the highest benchmarks out of any plastic box router you can find. This further compounds the problem of switches with low amount of concurrent connections, because it causes more frames to be en-queued than transmitted.
- The ultimate limiter is the infrastructure. We're still using landlines which lost 50% or more of the electricity they carry to heat, let alone em interference. It's been over a decade since the last alternative was offered in this regard, which is closer to 10% loss of electricity and it allows the electrons to travel through the wire instead of on the surface of the wire. It doesn't take a genius to figure out there is more volume in a cylinder shape than there is circumference, in fact it takes middle school level math. Even more that same alternative can not only conduct data signals but also electricity for power at the same time due to how neat and ordered the electron flow becomes inside of it.
That last bit is ultimately why most thing fiber optic is better than electrical cabling. With fiber optic your transmitting through the cable instead of on the surface of the cable. But fiber optic has one big flaw in this regard that a larger cable doesn't increase throughput, something that alternative cabling for data/power does not suffer from.
These are just a few reasons why electrical based networks are perceptibly slow. It's not because of the technology, it's because we're so damn stubborn about upgrading antiquated technology, even in cases where the new technology would yield more profit in the long run. | |
|
| Trapping Light
Posted: 2/25/2012 2:44:32 AM | "Really can't see light taking the place of electricity in computers. In fact electricity based computers could be a whole lot faster than they currently are."
No I cant either, but that's only because the idea that light could carrry electrical current/voltage is counter-intuitive, and it's hard to wrap ones head around..we are simply limited at the moment by the confines of the human imagination (massively on the decrease I might add) and that "how to do it" quantum leap. There has however, been ongoing research into this, 15 years at least if I remember correctly. From what I understand of light and it's characteristics, I very much doubt it will be a possibility though, light simply isnt a suitable medium. I think diamond technology however is a possibility.
"We're still using landlines which lost 50% or more of the electricity they carry to heat, let alone em interference. It's been over a decade since the last alternative was offered in this regard, which is closer to 10% loss of electricity"
Cabling isnt the real issue, it's light years ahead of the technology. However, if the incumbents were to change (even just) cabling everytime a new one came out, just to provide the fastest, cleanest service, then we'd be back in the 50's, with one phone in the village. It simply wouldnt be cost effective for the provider or the customers. My knowledge of losses is somewhat old but ifmemory serves it's far less than 50%. Losses on cabling isnt the hardest thing to rectify either...shielding/insulation can take care of that. It's the losses and leaks elsewhere that can prove to be a nightmare. Switch rooms are one hell of a lot cooler now than they were 30 odd years ago, not to mention quieter. And yes, once it becomes cost effective Im sure the providers will make the changes..And what about obsolescense, a disaster waiting to happen in a lot of technologies, move into the wrong technology too fast and you wont have a business in 10 years. So yeh, maybe there is the technology out there, but availing of it might not always be the best recourse. the same goes for computers too...changing up just for the sake of it might not always be wise, unless it's a necessity (Im referring to industry, not the domestic market).
Also, in industry, there might not always be the personell that can work with the newest of the new technology, it's growing so fast (Moores Law I think) tis hard enough to get to grips with what they already have, some employees would be eternally on training, never do a days work in their lives.
"and it allows the electrons to travel through the wire instead of on the surface of the wire. It doesn't take a genius to figure out there is more volume in a cylinder shape than there is circumference, in fact it takes middle school level math".
Chrono, thats RF you're talking about, it's called the skin effect. I dont know much about computers, but I doubt they operate on RF, it would blow the SMT to smithereens. In fact, I'd even say RF is unwanted on SMT, hell even the slightest leak would cause havoc. And yes agreed that there is more volume in a cylinder than it's circumference, but RF doesnt care, by it's very nature it will travel along the outer circumference. In fact, a lot of RF cables are hollow for this very reason. oh LOL...would love to see the teachers face if she was asked to teach skin depth maths to primary kids, her response would certainly be electric. | |
|
|
Home
login 
