>And security would be a snap—if you can’t see the light, you can’t access the data.
That doesn't make any sense at all. People don't intercept data transmission with their eyes. If your computer or device can "see" the light, so can mine.
I think the point is with, say, wifi, it is broadcasted in all directions, where LiFi would be pointed directly at the destination, which makes it much harder to intercept.
I already use Ethernet over household wiring to get data from the uverse router to my wife's office. I think it's 200 Mbs, but whatever, it's completely transparent to use. It may be the only completely foolproof piece of hardware in the house.
Is this meant for receiver-only, or bidirectional transfer? (he says bidirectional in the ted talk, but only talks about phones using their cameras to receive).
Seems a little like a technology in search of a problem.
You could always use radio waves for the upload band. That way, it will still increase both download and upload speeds (uploads will can have a wider band now that they don't compete with downloads for the same airspace), it just won't increase upload speeds as much.
This is the right idea for sure. Data usage is almost always asymmetric - people download much more than they upload. So let's widen the download channel using visible light and still use the sufficiently large RF channel for the up stream.
You avoid interference in the same way you avoid it for radio wave communication: multiplexing. It could be time-based (time slots for each sender), frequency-based (different colors) or code-based (look up CDMA).
Plus, you already have 'spatial multiplexing' because light doesn't pass through walls as opposed to RF.
It's also convenient in that it takes advantage of the existing infrastructure if you're performing the communication with general purpose LED lamps, which are being rapidly adopted for their energy efficiency.
I think the coolest part of this would be if someone were to take it and combine it with a "standard" (I really don't know how standard it is) Powerline ethernet setup. Would make for a more reliable than wifi setup in some buildings while being easily compatible with existing hardware.
Saw this in the university research lab where it was developed about a year ago while touring new tech in Germany. Very cool to see that it's still moving forward.
This blew my mind. It makes sense that light will be more controllable (think lamps projecting photos onto a wall), but the light itself being an internet source is trippy.
Does it means every light source is connected to a network cable ? If so, where is the security ? All it takes is to find a network cable and intercept the traffic. And there will be thousands of network cables in every building (at least as many as power cables). For me 'security' sounds like false advertising, in practice.
I really hope they don't use this to autopilot cars. That seems completely reckless. If a terrorist happened to get an LED flashlight that disrupted the transmission of whatever light signals were going to the autopiloted cars, there would be chaos. It would be so easy interrupt these signals.
Thinking more about it, any signal "controlling" an autopiloted car would be insecure and be an easy exploit that any malicious-minded people could take advantage of. But that's a discussion that's besides the point.
39 comments
[ 6.4 ms ] story [ 89.1 ms ] threadThat doesn't make any sense at all. People don't intercept data transmission with their eyes. If your computer or device can "see" the light, so can mine.
http://blog.ted.com/2011/08/02/wireless-data-from-every-ligh...
It seems like a good idea but he does not address how to get the data to each light fixture
Is this meant for receiver-only, or bidirectional transfer? (he says bidirectional in the ted talk, but only talks about phones using their cameras to receive).
Seems a little like a technology in search of a problem.
Plus, you already have 'spatial multiplexing' because light doesn't pass through walls as opposed to RF.
http://news.ycombinator.com/item?id=2850288
LEDs can go to much higher speeds - but even still, this isn't new. Researchers in Japan and Boston University have been producing demos for years:
http://www.bu.edu/smartlighting/research/smart-light-prototy...
http://www.naka-lab.jp/index_e.html
This technique was used in at least the early 90s
Thinking more about it, any signal "controlling" an autopiloted car would be insecure and be an easy exploit that any malicious-minded people could take advantage of. But that's a discussion that's besides the point.
Well, same meaning I guess. Only literal.