6

I am trying to prevent a potential attack, by first learning how to perform it.

Pretend there is a cheap £20 - £50 computer, which an attacker with physical access to wishes to crash. This computer will remain plugged into the power (electric going through it), but must be crashed (as in not working while the attacker is attacking it).

This attack will be short to medium term, so highest 2 months at a time. It must not cause permanent damage, as much as possible.

The attacker is limited as it is protected by a plastic case, good locks, and various anti tamper protection. However, the gap between the circuits and plastic is only a few centimeters. We have no way of preventing anyone getting physical access.

Must be as easy, cheap, less technical, and safe as possible. No use recommending something which has a long distance so exposes humans or other computers which are nearby. But they will very rarely be within a meter of the device, the device could be pointed in the direction of the computer attached to the front of the plastic case, and put in a lead box to stop leakage?

With the above in mind - what would be the best way to achieve this? EMP, but would that not damage computer? Xrays? Any ideas are welcome.

Please note: we only want physical attacks at the moment, such as using waves. No software attacks, nor hacking WiFi etc... this is to crash the computer, think of this like crashing a calculator (except more powerful).

EDIT: Think of the device like an electric "smart" meter. That shape, cost, design, hardware, almost similar security.

k1308517
  • 1,272
  • 14
  • 27
  • Comments are not for extended discussion; this conversation has been [moved to chat](http://chat.stackexchange.com/rooms/40332/discussion-on-question-by-k1308517-crash-a-computer-using-waves-or-similar-techn). – Rory Alsop May 26 '16 at 16:02
  • Sir or madam, whatever contraption you've come up I congratulate you for having a more creative mind than I. Cause I've racked my brain but can't envision any plausible scenario in my head where a device would have all characteristics you've listed.. :) And add in a security model where you're seriously concerned about attackers using EM attacks against it... . – mostlyinformed May 26 '16 at 20:57
  • Less facetiously, FWIW I fully agree with Jozef Woods's answer below: no matter what method you could use it would be very difficult--if not impossible--to create an EM attack that could reliably crash the software/firmware running on a device without doing damage to any hardware. Perhaps very, very high end attackers (eg. major nation-state intelligence agencies) might have the people with great expertise and the specialized equipment one would probably need to be so precise in EM attack effects, but if you were facing that level of threat you likely wouldn't be asking about it here. – mostlyinformed May 26 '16 at 21:14
  • @halfinformed No. We have seen it done by a regular electrician not secret agents LOL, it might have damaged the hardware but not to the extent where the device breaks and becomes unusable (or that anyone except for forensic scientists would notice it is damaged). – k1308517 May 27 '16 at 09:55
  • @k1308517 That is interesting. Hmmm... – mostlyinformed May 29 '16 at 03:47

6 Answers6

9

Several motherboards have overheating protection, if the computer or especially the CPU / graphics card gets too hot, the computer will switch off. So ... turn up the heat in the room (with a big heat-source) until the computer starts rebooting indefinetly. It will be hard not to damage other components I think.

In order not to target other nearby components, you could use another box, that you put over the target-computer and fill it with hot air or something similar.

hamena314
  • 2,007
  • 1
  • 15
  • 23
  • Comments are not for extended discussion; this conversation has been [moved to chat](http://chat.stackexchange.com/rooms/39902/discussion-on-answer-by-hamena314-crash-a-computer-using-waves-or-similar-techni). – schroeder May 17 '16 at 16:05
  • I'm pretty sure modern house heaters cannot bring a computer to anywhere near its maximum safe operating temperature. – forest Apr 05 '18 at 08:10
3

For wireless attacks, you're limited by having to induce currents in the machine, and it's nigh impossible to do that predictably, without going overboard (i.e. EMP style) and causing permanent damage.

You could try doing something the (maybe) installed wireless card, but you'd have to know something of the software, and you'd basically be attacking using wifi as the carrier.

Probably the easiest route would be to disrupt the power supply (either turning it off, or modifying it to be something other than the 60Hz sine wave at 240V, or whatever the device is expecting), but going between the extremes (turning it off, or completely overloading it and causing damage), would require you to be very careful, and would be very involved.

Main point is that it's quite easy to kill a device you don't like. Temporarily disabling it would require access, resources and/or expertise that your alleged perpetrator doesn't have.

Jozef Woods
  • 1,247
  • 8
  • 7
  • Comments are not for extended discussion; this conversation has been [moved to chat](http://chat.stackexchange.com/rooms/39902/discussion-on-answer-by-hamena314-crash-a-computer-using-waves-or-similar-techni). – Rory Alsop May 26 '16 at 16:14
3

It sounds like you're trying to cause/prevent a resettable denial of service condition. If you are ruling out network access to attack the software running on the system, you might consider interfering with the data going through the various devices on the mother board. I'm not an RF engineer or an EE (Electrical Engineer), but it seems though it might be possible to disrupt data flow within the motherboard or other device if you had a transmitter tuned to the resonant frequency of the CPU or RAM. A few bits out of place could cause a kernel panic or similar event.

To protect against an attack like that you would shield the system electromagnetically by using sufficient wire mesh, sheet metal, or foil and properly grounding same. Depending on the power of interfering signal you expect to encounter, you might consider adding chokes to prevent stray RF from entering the enclosure. It will also have the added benefit of reducing the amount of signals emitted from device you wish to protect.

Edit:

The antenna used to transmit or receive signals like a smart meter can still be within the device enclosure, but does not have to be enclosed within the shielding. However, in order to limit the amount of interference received by that antenna and then induced to other internal circuits, I would recommend an appropriate bandpass filter with a passband that is as narrow as the transceiver allows. Depending on your budget, you could also incorporate optical coupling between the transceiver and other circuitry. Because your device requires a transceiver of some kind, you won't totally be able to cut it off from outside interference, but those two precautions would help.

Edit 2:

There's a whole world of information about RF theory and transmission - however - now is a good time to note that to perform this type attack you'll probably need a fair amount of power and will probably end up breaking the law, unless you're testing inside a Faraday cage yourself. Transmitting outside of assigned frequencies and over the very small power limits for non-licensed transmitters will land you in trouble. Having said all of that, here is a small device, that would NOT achieve your desired result, but might be adapted to do so.

Edit 3:

I just remembered that you listed prices in pounds, Ofcom appears to be the UK equivalent of the FCC and will have similar restrictions.

Liczyrzepa
  • 373
  • 2
  • 7
2

Electromagnetic radiation?

Many parts of computers (such as Hard Disks) rely on magnets. Provide a sufficient amount of electromagnetic radiation, and the computer is bound to stop working.

If the exposure to these waves isn't for too long, there (probably) wouldn't be any permanent damage, the computer would just crash as soon as the hard disk started malfunctioning.

Update
The OP mentions on chat that such an attack successfully took place. (S)he also mentions that it was a small embedded device and that the attacker placed a small object on top of it in order to achieve this. As far as my imaginations goes, I still think that (electro)magnets are the most viable cause of attack...

undo
  • 2,075
  • 2
  • 12
  • 18
1

What are you describing is an EMP emitter.

If the EMP emitter is not strong enough it will not destroy the device, it will only cause an interruption. The EMP pulse power depends on the number of wire wraps, wire thickness and voltage so the EMP power can range.

Best way to protect against it is by creating a Faraday cage, this will not provide a full blockage or attenuation but will heavily attenuate the EMP pulse.

My advice is not try to move the device in a secured area and to research before doing this test and also to work in an isolated lab.

Is very easy to create a small EMP device, you don't need any knowledge, harder part is to control the test, to generate the amount of power that you need and not to destroy the device(s).

I think you are focusing to much on the attack. If an electrician has access to this device he can find other methods of attack so don't focus to much on testing/reproducing, focus on protecting and if this will not work then restrict the access, put this device in another box and lock it.

lauda
  • 136
  • 1
  • 8
0

As your aim is to "prevent" such an attack, I would focus on physically bolting the case inside a suitable tamper-resistant enclosure and then to the (ideally concrete) wall/floor, with tamperproof fixing. With the power cable inside the housing and not exposed, ventilation made inaccessible, and ideally steel in the enclosure against electromagnetic radiation in/out, this should be all the protection you reasonably need for this attack. The rest is too much "just theoretical what-if" to go into.

Stilez
  • 1,664
  • 8
  • 13