Li-Fi with my little eye … a vulnerability Proponents of visible light communications (VLC) like “Li-Fi” love reminding us of the bonkers speeds they can get (200 Gbps last year, for example), but just like its radio-spectrum counterpart, it needs protection against eavesdropping and jamming. In this paper at Arxiv, Warsaw University of Technology researcher Grzegorz …
This kind of exploit is builtin to li-fi, which was over ever a proof of concept for particular installations. Ie. a typical research project which is unlikely ever to be directly usable, but still worth doing.
In practice: where can't you use either wired or wireless ethernet?
It's a great idea to get high speed data everywhere in a room; however this requires an infrastructure to be put in place. Think how long it's taking to replace incandescent lamps with LED ones, and all the trouble with dimmers, for example. Then there is the reliability aspect. We expect hard-wired stuff to last decades. I put in two power sockets with built-in USB chargers. They lasted a year and I had to take them out.
So, it's more likely this initially will find an application in a new building construction which needs this capability, not consumer but industrial.
What problem does it solve? Individual video streaming to portable applicances in a space?
Does it reduce building wiring? It would certainly make shop floor reconfiguration a bit easier, but still power is needed, so network cabling in parallel is not too much more of a burden.
There's a problem waiting to be solved out there, but I've not figured it out yet.
The biggest advantages might come in places like conference centres where the Wi-Fi is often saturated.
Nope, not saturated, just poorly set up. With professional equipment and a little skill you can always setup a wifi network correctly so that contention will not be an issue. But the uplink might: many conference centres have very poor internet connections: 10 MB/s won't cut it for a conference of more than about 20 people.
>There's a problem waiting to be solved out there, but I've not figured it out yet.
Light is easier to contain than parts of the radio spectrum which pass through walls. Easier to contain = more bandwidth (if you have walls). For home use, probably more secure than traditional wireless and more bandwidth as the signal doesn't leak to the neighbours or get disrupted by 2.4Ghz devices, microwaves etc. You have more wires (to each room), but you have a lot more bandwidth.
"how many sysadmins have found rogue Wi-Fi access points in the businesses they work for?"
Not comparable issues in my mind. Bringing in a Wi-Fi thingy and hooking it in to the network is easy enough - they're cheap, rather easy to set up and also easy to keep inconspicuous. If the network admin is not whitelisting connected equipment, then anything goes anyway.
Climbing up to the ceiling to replace a light is not inconspicuous, and the hardware is less easy to get (won't stop a determined attacker, though). Obviously, once in place it'll be a devil to notice, but it still falls under whitelist control (if that is in place).
With a minimum of attention, network access shouldn't be possible, which would reduce the attack to DoS-levels - still a nuisance, but not really a security issue.
"Climbing up to the ceiling to replace a light is not inconspicuous,"
Wouldn't a sensible ne'erdowell just run a desk-based widget, and rely on ceiling bounce?
The scenario where I could see LiFi (maybe using a wifi backchannel) being useful is in a sports stadium, where thousands of data consumers (replays, umpire calls, stats, chat, ads(sorry)) want a fat feed, and can all be relied on to be stationary and pointing towards a centre node - and then only because of the massive bandwidth available (and needed). Possibly better to just wire up seat-backs, though, but then you can't pull the BYOD game so easily.
For what it's worth - I still have the doohickeys somewhere that would link your printer to your PC via IR*. One box plugged in the PC, one in the printer - centronics interface, naturally. The boxes didn't even need to be able to "see" each other; the IR signals would bounce off the ceiling and you'd have a duplex connection.
In theory anyway.
*IR is in the visible spectrum - for some species. Just not for humans.
WiFi and light are handy communications mediums for casual traffic but you'd be wise not to use it for sensitive applications. For that we've got good, old-fashioned, wire......and something called 802.1x. This protocol was originally developed for routers/switches to establish who was connected to a particular network port before allowing general traffic on that port. It was extended to wireless but for some reason lots of SysAdmin people ignore it even though its been built into Microsoft's server offerings since the earliest days of WiFi.
If you're really paranoid then you could also try using Token Ring on that wire. This has the advantage that any attempt to interfere with the network medium - interrupt the traffic, add or remove a node and so on -- is immediately detected. Of course, you'd be playing with Token Ring.....
It seems that we don't do the obvious things but rather take an incomplete technology, use it for a purpose it wasn't designed for and then find flaws in it. We then patch the flaws rather than figuring out how to bulletproof the technology in the first place.
(BTW -- you are aware that the earliest use of 802.11 was interconnecting laptops with infrared?)
I'll grant you that running Token Ring might be a clever idea...but where do you get the hardware these days? Wasn't Madge/Ringdale the last Token Ring company left standing?
A better question might be where you'd get the drivers even if you find some old TR hardware sitting around. The Linux TR project (http://www.linuxtr.net/ , guess I can't embed links) site is still online but it's been years since that site was updated and I'd be shocked to learn that any TR driver code was still present and workable. Drivers for Token Ring hardware in Windows versions after Server 2003 and XP would also be quite surprising.
Maybe NetBSD has something?
(In case you're wondering, yes I do in fact still have a working Token Ring network. Or I would very quickly upon powering up the backbone gear and any nodes. I've got a few IBM 8228 and 8226 MAUs in reserve, while a North Hills LAT3371 MAU is my primary. It's tied into my Ethernet network with an IBM 8229 LAN bridge. For a while I even had separate 16 and 4 megabit networks, with a PS/2 Model 50 bridging between them.)
First of all, there are all the security problems of any wireless connection, except that windowless rooms can contain them better than WLAN.
Second those advertised speeds are bogus, yes you can reach them on fibre easily, but in actual rooms you have lots of reflections. Those reflections will give you your signal delayed by a certain amount... without very expensive DSPs working at speeds unobtainable today, that's very hard to get rid of. Also, yes, we can modulate Lasers to those speeds, but not by turning them on and off. LEDs have further problems as they are, comparatively, wide band when not modulated. It's very much like those early transmitters that use an arc to transmit.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
