Schneier on Security
A blog covering security and security technology.
« Threats vs. Vulnerabilities |
| Detecting Words and Phrases in Encrypted VoIP Calls »
March 24, 2011
Transmitting Data Through Steel
This is cool:
Tristan Lawry, doctoral candidate in electrical and computer engineering, has developed equipment which can transmit data at high rates through thick, solid steel or other barriers. Significantly, Lawry's kit also transmits power. One obvious application here would be transmission through the steel pressure hull of a submarine: at the moment such hulls must have hundreds of penetrations for power and data cables, each one adding expense, weight and maintenance burden.
What's interesting is that this technology can be used to transmit through TEMPEST shielding.
If you had the through-metal technology now reinvented by Lawry, however, your intruder -- inside mole or cleaner or pizza delivery, whatever -- could stick an unobtrusive device to a suitable bit of structure inside the Faraday cage of shielding where it would be unlikely to be found. A surveillance team outside the cage could stick the other half of the kit to the same piece of metal (perhaps a structural I-beam, for instance, or the hull of a ship) and they would then have an electronic ear inside the opposition's unbreachable Faraday citadel, one which would need no battery changes and could potentially stay in operation for years.
Spooks might use such techniques even where there was no Faraday cage, simply to avoid the need for battery changes and detectable/jammable radio transmissions in ordinary audio or video bugs.
Naturally, if you knew how such equipment worked you might be able to detect or block it -- hence the understandable plea from the British spooks to BAE to keep the details under wraps.
Unfortunately for the spooks, Lawry has now blown the gaff: his equipment works using ultrasound. His piezo-electric transducers send data at no less than 12 megabytes a second, plus 50 watts of power, through 2.5 inches of steel -- and Lawry is confident that this could easily be improved upon. It seems certain that performance could be traded for range, to deal with the circumstances faced by surveillance operatives rather than submarine designers.
Posted on March 24, 2011 at 7:37 AM
• 41 Comments
To receive these entries once a month by e-mail, sign up for the Crypto-Gram Newsletter.
The information transfer is trivial, but 50 watts of power? That's pretty insane. And easy to detect, judging from the noise in the video (although maybe that was due to aliasing).
I find this rather implausible - data rate is proportional to bandwidth so a 12.4Mbytes/s data rate implies an audio frequency possibly approaching 100 MHz, which I can't believe.
Note with sound in steel you can't do tricks with polarization.
"What's interesting is that this technology can be used to transmit through TEMPEST shielding"
Err not sure who came up with that but they might be in for a bit of a surprise.
First of TEMPEST is not just about EM energy being radiated it is also about other forms of energy including mechanical vibration up into the hundreds of MHz (think 3 uM wavelength sound or less).
Now as many "old school" electronic engineers know the inductors (coils) in an oscillator tank suffer from "microphonics" (and some know it works the other way as well). The old school cure for this was box it up in a tin box about three times the dimensions of the coil and fill with bees wax.
The second thing is that "TEMPEST" screaning is very very rarely "single skinned". Also not all the screens are solid plate often they are fine copper wire mesh or pressed metal mesh, neither of which is a particularly effective conductor of sound.
All that aside I suspect the "spooks" did not want the very high levels of mechanical vibration simply as it would muck up their other "listening" activities.
I suspect that a small ultrasound transducer driven by a long sequence PRBS with nonlinear feedback would muck the communications up.
After all it's not as though it's that difficult to test (think ultrasound imagining equipment)...
Couldn't you defeat this simply with a double Faraday cage with an air gap between?
@ physisist & patent attorney,
"I find this rather implausible - data rate i proportional to bandwidth so a 12.4Mbytes/s dat rate implies an audio frequency possibly approaching 100 MHz, which I can't believe"
Err that sounds like a "gut reaction".
First off the data rate and the transmitted symbol rate do not have to be the same and are often not. Some systems can put sixbits of data into a single transmitted symbol.
Secondly you can have data rates as high as +-0.35fc.
All of this is before you use data compression techniques to increase the average data rate.
certainly a gut reaction, but his demo is done with audio which would likely be nearer 128k BITS per sec not 12.4M BYTES per sec.
I'm not doubting that he can transmit data but suspect that the rate claim is inflated by several orders of magnitude.
"Couldn't you defeat this simply with a double Faraday cage with an air gap between?"
Why another Faraday cage? Why not wood? Or clear plastic? Or something non-conductive?
Say it with me: arms race.
Also, how about through insulation foam or drywall?
Interesting reference to beeswax, I have repaired transistor radios and tuned transmitter coils with hot wax :)
Ultrasonic transducers are used to study fractures and defects in solids,(especially steels and other alloys) in environments from materials science labs to massive structures that are designed to work under tremendous loads. think dock gantries, submarine hulls ..
Low power is used in such transducers which rely on the mapping the echoes from the cracks to plot out in 3D the location and size of a crack or fracture
Old school PCM\QAM was able to achieve high transfer rates over power lines, technology certainly would have improved over the past decades
Only thing I see as an issue is the gradual weakening of the structure using this technology
Data transfer patterns would lead to localized heating, change the temper of the structure, weakening any stressed members, but then again
@physisist & patent attorney
Don't know for sure since I am far from an ultrasound expert, but some quick googling learns that 20Mhz is not out of the ordinary. I even came accross 1Ghz. All for medical applications.
If this technique additionally uses several channels, as claimed, the claimed data speed appears easily attainable?
@ physisist & patent attorney,
Clive Robinson is correct. The carrier frequency has nothing to do with the data rate. With AM they could use a very high sample rate and have different amplitutes represent words 2 or more bits long. For example lets say the wave sample below represents one half of a wave form after the audio wave is converted to an AC wave:
Amplitude 1 = 0 mV +- .20 mV = bits 00
Amplitude 2 = 1 mV +- .20 mV = bits 01
Amplitude 3 = 2 mV +- .20 mV = bits 10
Amplitude 4 = 3 mV +- .20 mV = bits 11
If the above audio signal is 10kHz and a sample rate of 40 kHz, 40 kHz * 2 bits = 80Kb/s.
You might want to brush up on your Shannon-Hartley theorem, before speaking quite so authoritatively.
My guess is that steel hulls could support at least 30Mhz ultrasound signals, assuming a wide band coupling transducer (say 20Mhz BW) and say 1000 channels QFDM with QAM256 modulation you should be able to get about 300Mbps through the steel hull before Shannon channel limits are reached.
Now the big problem with a Steel hull is that the impedance mismatch air-steel-air this would result in very long reverberation times so signals will need some serious guard-banding and packet shaping. But since we are talking theoretical limits the claim of: 12Mbytes / sec = approx 120Mbps (with overhead) is a long way below the channel limit.
I'm wondering about the Watts value, myself.
I don't have any memory of ever learning about translating sonic energy into Watts used by the receiver.
How many watts are in ultrasonic sound waves? I'm assuming it's fairly material-independent, but that recapturing wattage from a solid is much easier than from a fluid.
Could 40 Watts be the energy strength of one of the transmitters?
Are all 40 Watts used on the other end?
Is the power-distribution thing theoretical, or proof-of-concept?
RE: undetectable listening. Couldn't someone poll the room for ultrasound vibrations? Would it be detectable in the air?
Anyone else find the idea of a submarine's hull screaming at frequencies not seen in nature? That and I wonder how well it works once you start damping the steel.
Well, given this is pretty immature technology, one has to think it will be improved upon in time. In any case, it is interesting, and illustrative of the innovative thrust of the human mind - no problem goes unsolved forever!
As for the Tempest aspect of it, I know of one such terminal (originally designed by a friend for Zenith in the late 70's or early 80's) that didn't use electrical signals between the keyboard and processor unit, but passive light pipes. The keys would simply break a light path, so processor box could be suitably shielded, and suitable shielding on the connecting pieces and keyboard would eliminate the hole in the processor box for the optical cables. However, if the processor box itself were to be compromised...
Yup, same way us guitar players fix up old pickups that have gone 'microphonic' ;)
sounds like a great way to wirelessly charge your cell phone.
50 watts is enough to power some fairly serious equipment, e.g. video cameras with lights or sonar transmitters/receivers. With proper shaping, you should be able to get almost all of the signal to the receiver and very little going to the rest of the hull.
For spook applications remember that sound sensors do just fine with a few nanowatts coming in. (You'd need a few microwatts if you wanted to power the outside sensor as well.)
What I would like to do is to build a system for transmitting data using scalar waves (such as shown in eg books and kits by Konstantin Meyl). Unless perhaps that has already been done by someone?
I'm thinking their might be a relativly simple passive defence to this.
As you mention the air metal interface will be baddly mismatched, thus what would happen if you mad the inside a "grating" this would cause most of the energy to get reflected back from that surface and effectivly bounce around inside the metal or back out the input side.
Mind you I must admit I'm very interested in the power transfer aspect I can see the likes of a rod or wire being an effective "wave guide" but through a metal plate... I'm not so sure, then of course there is the question of the pickup transducer....
Hmm there's a lot of interesting thoughts.
Oh I've mentioned TEMPEST and "sound" a few times on this blog and I've been thinking back to the first academicaly published asspect of it.
As far as I can remember it was to do with the first experimental setup for Quantum Crypto where he mentioned that he knew it was working because he could hear the poleriser changing on the oposit side of the room.
er... submariners tend not to like radiating noise to the environment. they actually make considerable effort to stay silent, because radiated noise is the main way (only way?) to find them and thus, to kill them. not sure how well ultrasonic is transmitted by seawater, but i suspect that there is zero chance of this scheme being implemented on a submarine. small thru-hull fittings (power, data) are not a difficult engineering problem; subs all have to have several large thru-hull fittings anyway e.g. "the egress", ventilation, weapons systems, motive power, etc etc.
What's that buzzing I hear? Sounds like "a solution in search of a problem".
20MHz sounds a bit high to me, by about a factor of ten.
The limiting frequency is set by the crystal size in the material; if the wavelength is smaller than the crystal size, attenuation goes 'way up.
(And so that's perhaps why such a crazy amount of power is needed, and even more if you need to transmit over larger distances)
Re: acoustic impedance mismatch: PZT transducers will match reasonable well, so not so much trouble getting signal in and out, the attenuation (see above) means echos aren't an issue, and at >1Mhz will not propagate in air or water.
For the baud rate: just look at what modems do with a voice phone line (~5kHz bandwidth).
this is true for military submarines, but it could be useful for research submersibles, especially as it avoids adding additional holes in the pressure vessel.
Air - Steel impedance mismatch
Could this be the reason why they are talking 50 watts of power?
Don't forget that PETA will be all over you when it turns out that your ultrasound emissions cause the whales to cry...
@ Clive: Where could I get some of that "ultrasound imagining equipment"? -- You know, where you apply the ultrasound beam to someone's head, and you can tell what they're thinking (imagining)? :)
@ Sam: Sure, make fun of the fact that we've been on the verge of extincting yet a few more species, and these being rather magnificent ones at that. Actual studies have shown that dolphins are more intelligent than kill-'em-all-off blog commenters.
@ Clive R
Actually the Air-steel-Air impedance mismatch would be an impossible issue to solve, but fortunately submarines are usually underwater so hopefully there is no air in the comms path.... that said, at 50W for an OFDM 20Mhz BW they are well above the water cavitation power limit, so this implies a distributed transducer. (i.e. parallel data paths or possibly MiMo style Rx/Tx pairs)
The system will need a wide band ultrasound transducer plus a suitable coupling fluid, so the datapath is
electrical =>mechanical=>acoustic => fluid =>steel hull
At each of these points you must correctly match the impedance if you want to efficiently transfer power.
If you're interested I'd suggest you look at the ideal material characteristics for 1/4 wavelength matching layers for PZT medical diagnostic ultrasound transducers and their associated Tx and Rx electrical impedance matching circuits. If you use a little imagination you can probably figure out the rest for yourself.
@Snarki, child of Loki
Typical diagnostic medical imaging is done with the pulse burst at 3Mhz to 15Mhz. Your right that the transducer size shrinks as you increase the frequency and this would create a major problem for the power density (hence the cavitation reference)
However, if you used a 2D array transducer with the transducer spacing controlled so that spatial and temporal orthogonality is possible, than you should be easily able to get this data throughput even at lower BW and lower carrier frequencies.
"er... submariners tend not to like radiating noise to the environment. they actually make considerable effort to stay silent"
Yea but sound at Mhz does not travel long distances in water, due to multipath propagation issues, which make direction finding of HF signals difficult. Also the BW of ASW systems is a LONG way below MHz.
"ultrasound imagining equipment"
When I find some you will be the second to know...
Blaim the "predictive spelling" on this smart phone, I probably hit the wrong prediction in haste...
"@ Clive: Where could I get some of that "ultrasound imagining equipment"? -- You know, where you apply the ultrasound beam to someone's head, and you can tell what they're thinking (imagining)? :) "
All you have to do is apply a porn sniffer to their computer that automatically displays what they are looking at on your computer. In this use case, the imagining equipment vendor claims eavesdroppers will know what they are imagining about 90% of the time. (The remaining 10% were deemed so creepy they decided against improving the technology.)
This is a very cool project, but in their efforts to make a bigger story of it the journos have added some nonsensical histrionics.
FIRST POINT: this has nothing to do with TEMPEST.
TEMPEST always and only dealt with UNINTENTIONAL emanations. It has never claimed to provide one jot of defence against a local intruder, e.g. a bug. This is abundantly clear from the testing standards themselves: they make no effort to see if the equipment can be modified to cause it to fail.
For example, TEMPEST equipment may be re-tested from time to time but usually is not continuously monitored. A low tech intruder could defeat many forms of unmonitored shielding with a pair of tin-snips; that doesn't cause them to fail TEMPEST certification. TEMPEST shielding is nothing like 2.5 inch thick steel plate. It is more often a woven mesh, or even ordinary glass coated with an ultra-thin metallic film. There are already many non-RF means to signal through such a shield: IRDA, for example.
There are other standards dealing with a deliberate attack, but protection against an intruder is generally provided by physical controls, i.e. there are no pizza guys or cleaners getting anywhere even close to the shielded area. If an intruder does breach the physical barriers, then the equipment will not be used for classified processing until a sweep team determines exactly what she did.
The real danger with TEMPEST is not someone picking up a signal from a metre away. It is someone sitting out in the car park with a directional antenna and getting a picture of whatever is on your computer screen from 300 m away. Well, you aren't going to do that with 100 MHz ultrasound. The attenuation of ultrasound in air is already severe at 1 MHz and increases rapidly with frequency. At 1 MHz, it is around 1.5 dB per centimetre; at 2 MHz, is around 6 dB per cm. In air this is going to be a very short range signal.
SECOND POINT: there already exist standards and techniques for detecting and shielding against acoustic emanations.
British intelligence services used unintentional acoustic emanations to attack a classified communication system as long ago as 1956, so the basic idea is well known. Within the SIGINT community it even has its own specialist discipline, called ACOUSTINT. And there have been several recent papers on acoustic attacks, ranging from decoding key-presses by their unique acoustic signatures (i.e. space bar makes a different "click" compared to shift key), through to doing side channel cryptanalysis of RSA operations by using the variation in a chip's acosutic hum to assess instantaneous power consumption.
I don't know if these standards extend to micron wavelength ultrasonics, but they probably do.
THIRD POINT: the ability to send a signal through solid steel barriers is unremarkable. In and of itself is not even worth a patent, clearly failing the "obviousness" test. If you asked me to design such a system, straight off the top of my head I would come up with at least 3 quite practical alternatives. In descending order of practicality, they are:
* hard x-rays
All of these can be generated by compact electronic sources, can have fairly good directionality, and can be modulated for high data rates. I would guess the modulation is probably fastest for the x-rays, although a pulsed neutron generator can also be modulated extremely fast -- certainly orders of magnitude faster than the acoustic solution. And no, they aren't going to give anyone cancer; to penetrate a few inches of steel is so easy for these phenomena that they could be operated quite effectively at extremely low power outputs.
(There are many other types of signals that can pass through steel, although most seem to be less practical with current technology. Indeed I can think of at least 3 more straight away, which cannot be used with current technology because either the transmitter or receiver would be impracticably large or expensive for this application.)
Not only is it easy to come up with such techniques, but it has actually been done. Indeed for some industrial instruments, it is routine. Right on this blog, a few years back, we discussed an existing product which acoustically sends a signal -- a one-time passcode -- through solid barriers (including safes and vault doors) to an electronic lock.
FOURTH POINT: This is actually a pretty interesting invention, and worth its patent.
The good stuff is not that he sends an ultrasonic signal through steel. Some commenters above have danced around the point that the claimed bandwidth is very high, that it will require extremely high frequencies, and in metals these very short wavelengths suffer many forms of distortion. If you bother to read Lawry's press release, or actually watch his video, *THAT* is the problem he claims to have solved, and is patenting. His invention is not an ultrasonic transducer, it is a modulation method that provides high signal rate to bandwidth ratio, and low distortion, when transmitting through thick steel sections. Cool.
Meyl is either a crackpot or a scammer. Scalar waves in his sense do not exist.
In real physics, a scalar wave is an oscillation in any parameter that, at any given point, has just a single value. There are many examples but the most obvious is ordinary sound in air. The scalar that oscillates in a sound wave is just air pressure: it is a scalar because it has only one particular value at any given point, and that value oscillates as the sound passes by. Ripples on water are also a scalar wave (the oscillating thing is height of water surface.)
Scalar waves are contrasted to vector waves, which are an oscillation in a parameter that has multiple values, or a value and a direction, at any given point. A radio wave is a vector wave, because at any given point it has oscillations in both magnetic and electric fields, and both these forces have a direction as well as a strength. Wiggling a rope is another commonplace example of a vector wave: you can choose to flick the rope in various different directions.
One difference this means is that a vector wave can be *polarised*, if all the field directions point in some particular direction. A scalar wave cannot be polarised. For example, tie a rope to a post and pass the end of the rope through a fence made of vertical bars. If you flick the rope up and down, the wave passes straight through the fence. Flick it side to side, and nearly all the energy is stopped by the fence. The fence is acting as a polarizing filter on your vector wave.
And all those complicated terms seem much less mysterious when you can actually see them demonstrated with a rope!
So, what does this have to do with Meyl's scalar waves? Nothing. He is peddling nonsense and so he stole a real, complicated sounding phrase from physics in the hope of confusing people.
"The attenuation of ultrasound in air is already severe at 1 MHz and increases rapidly with frequency. At 1 MHz, it is around 1.5 dB per centimetre"
I'm not sure where you got this attenuation but I think it is far too low, attenuation is probably more like 10db/Cm/Mhz.
In diagnostic ultrasound the attenuation through the body (mostly water) is about 1dB/cm/Mhz. Air attenuation is at least one or two orders of magnitude higher, plus the impedance mismatch will make it very inefficient to couple ultrasound to air.
"His invention is not an ultrasonic transducer, it is a modulation method that provides high signal rate to bandwidth ratio, and low distortion, when transmitting through thick steel sections. Cool."
Do you have details of the patent?
I looked but could not find any patent by Tristan Lawry.
My guess is that he uses OFDM with raised cosine packet shaping and guard-bands that are longer than the sound reverberation time. (this would be a function of the Impedance mismatch). Not sure there is anything not obvious about this, but than again I think the same statement applies to most modern patents...
""ultrasound imagining equipment"
When I find some you will be the second to know...
Blaim the "predictive spelling" on this smart phone, I probably hit the wrong prediction in haste..."
Hey, we all do it, and in this case, it was merely good for a chuckle. But there was a fatal airline crash many years ago, in South America, in which "predictive spelling", a/k/a "auto-complete", was the direct cause of providing the opportunity for pilot error.
The USA-native pilot was going to set his nav radio to the next beacon in the flight plan. (Yes, before GPS and inertial guidance became primary sources.) He knew the assigned name of the beacon, and it began with "R". So when he typed "R" into this early generation of puter-assisted navaid GUIs, all of the ones of that initial within X distance were displayed. Normally, in the US, according to a pilot friend of mine, the closest one is at the top of the list, as that's most likely to be the desired one. (Distance-ordered, *not* alphabetic.) So when the list dropped down, he hit "Enter" without further thought or glance.
Unfortunately, that was not only the wrong one, but it was on the opposite side of a mountain ridge that was higher than their current (approach-to-land) altitude. The auto-pilot dutifully homed in on it, and after the usual chain of missed chances to escape disaster (including ignoring the ground-proximity alarm), the plane struck the mountainside.
I asked my pilot friend how reliable is it that "the one I want is going to be at the top of the displayed list". He said, in the US, about 90%. Not good enough for me, thanks, I'll *read* before hitting "enter". But what about in Latin America? No offense to our Southern neighbors, but this US-native pilot said that he wouldn't trust *any* such auto-thing there, and would always - what's the word -- oh, yeah, *read* and make sure he had the right one, or continue scrolling down the list.
I wish I could find the story, but it's a decade or two old, maybe more. Other than the above clues for keywords, I seem to remember (with a low degree of confidence) that the nav beacon in question was named "rojo" ("red" in Spanish). It *might* be this one, in which "the aircraft steered significantly off the airway for unknown reasons..." I might have known the unknown reasons from contacts in the industry, but not sure. (When you get older, your memory is the second thing to go.)
~~ Now, if you'll just tell us which app to "blaim" for the misspelling of "blame", which most likely was *not* predictive spelling..... Cheers! :-)
@ Nick P.: But I already know what they are thinking, with a much greater degree of confidence than the vendor. Another case of the marketing people trying to sell a solution that no one needs to a problem no one has ....
My spelling is truly appalling at the best of times and my use of quote marks not much better either (ask the old hands on this blog). It was suggested by many kindly souls that I should use a spelling checker... Well I use a smart phone for Internet activities such as bloging and the phone I had back then (Motorola Sidekick) had no such thing available.
So now I have an LG Android (1.7) phone... It's slide out keypad unfortunatly has a crap driver which hangs and the key caps so thin some have cracked in half. As for the predictive spell checker it only shows one ore two words across the bottom of the screen in a pop up banner and not always in a sensible order (like your example). Also very anoyingly it can be shown over what you are actually typing...
So I don't always use it so it's not alwayst to blame for some of the spelling mistakes, and on the odd occasion I even pull up the wrong spelling (don't get me going on to/too where/were etc etc or American/English spellings with S/Z 8)
As Bruce once noted it would be pointless me trying to be a sock puppet as "my style" (or lack there of) is almost instantly recognisable ;)
@ MODERATOR AND ALL: The post in German has a spam-link sig, while the translation mixes spam for their product with comments about Yemen's ruler and protests. Is this blog being used for steganographic communications among the protesters? How cool! ... or is it just spam?
@ CLIVE: I don't trust spell-checkers, either, because as you pointed out, they don't distinguish homophones like where/wear or to/too/two. I was just ribbing you a bit. And like Bruce, this writer too pointed out that your style was inimitable, in the discussion of sock puppets, nym forgers, etc. a few days back.
I'd think your apps, like most browsers and OSs, could be localized (localised ;) for en-BR or en-US. Ignore anyone who fusses because someone used one or the other.
And if you ever wish to have a professional presentation proofread (how alliterative!), I have some experience in the field, although professional proofreaders have obviously disappeared from the scene, with spell-check as a poor substitute, as is evident all over the Web, books, magazines, etc. In consideration of your long history of contributions here, some of which have been quite educational to this reader, I'd volunteer to help, if the task isn't too ominous. (If it's 150 pages, rates are negotiable... ;)
Here are some fun, easy reference guides to usage, in song parody form:
"There, Their, They're" (plus "your/you're/yore"): http://www.amiright.com/parody/60s/...
"Who and Whom" (plus "who's/whose"): http://www.amiright.com/parody/60s/...
Punctuation, specifically colons : semicolons ; and (parentheses): http://www.amiright.com/parody/60s/...
If you want to contact me, there's a comment form at the bottom of the list of my 400+ parodies, linked from the sig of *this post*.
The parody topics themselves range from scientific/educational, to politics and current events, all the way to the downright silly, e. g., sheep-shagging. Reader discretion advised on a few of the double-entendre ones, though no actual pr0n-type stuff. The tech-related ones include several MS-bashes and a celebration of binary, hex, and other base systems.
There is also a reCaptcha to reveal direct email address (vs. the contact form cited above) right after the copyright notice at the bottom of all posted since July 2008, at the date-sorted (and any-header-sortable) listing at the actual host site of the songs, which is
Hope you enjoy them as much as I enjoy your comments here, and the offer to proof your published work prepublication still stands. :)
@ Clive: Tried to respond extensively, including many sources of help, but the comment apparently had too many links:
"...your message was spam filtered and will be held for review by a human. We apologize for the inconvenience."
I hope the human approves it -- *none* of the links were commercial, or tried to sell you anything -- but if it doesn't show in a day or so, stay tuned to this thread -- I'll see if I can point you there in a non-offending fashion.
Ironically, the filtered comment also pointed out the link-spam in the German post a bit above, and that the contents, though appearing to be spam, were possibly steganographic communications among Yemeni protesters. Translate it and see if you agree.
I guess I should have posted in Russian or German, since that spam gets through!
Zero pressure hull penetration electronic periscopes are a reality with plans to spin propeller shafts sans hull penetration electromagnerically to reduce not eliminate pressure hull penetration. Weapons are next followed by crew mitigation. Ed
@ Mr Robinson,
Assuming the post doesn't show, ultra-short version: I offered to proofread your research papers or other stuff you planned to publish. As evidence of qualifications, pointed to where I had made some grammar rules into songs, which are easier to remember. There-their-they're, who-whom, fewer-less, and how to punctuate. Hit the signature on this message to see the list, then search the above terms. Those are old songs. Started adding a captcha to my songs right after the copyright notice, in July 08, so hit any song after that time to get contact info. Cheers.
@ physicist & patent attorney,
The relation of bits/second to necessary bandwidth depends on the modulation technique, and is limited by the SNR. In 802.11n wireless, OFDM has been used to transmit 288 Mbps through a 20 MHz channel. It is not clear what the usable ultrasound bandwidth of a steel plate might be, but transmitting on the order of 100 Mbps through it dosn't sound entirely implausible.
I think the interesting aspects of this article are not so much how feasible the technology is right at this moment, but more, the perceptions invovled.
Faraday cages were deployed in hopes of preventing accidental transmission of data, and they're thought of as a means to do that effectively. Here is an example that defeats, not so much the Faraday cage, as the Idea of the Faraday "Cone of Silence".
A little thinking outside the box will help here.
Using sound to transmit data through ships hulls was in vogue 30 years ago. US Navy research produced a system for data transmission anywhere throughout the hull. One purpose was to download maintenance manuals to a laptop, rather than carry books. Limited size hard drives and no data CDs back then.
Using powerful ultrasound on a submarine hull might affect whales...
Schneier.com is a personal website. Opinions expressed are not necessarily those of BT.