Sign in / up

back to article Sniffing your storage could lead to sensitive leaks, warn infosec bods

Data from storage devices leaks through electromagnetic radiation to a much greater extent than previously thought, according to new research. Near-field analysis allowed security researchers at MWR Security to infer (or ‘sniff’) data transferred internally within a device. The finding means that resilient systems are far …

COMMENTS

Post your comment
House rules Send corrections
Add to 'My topics' unstar *
  1. CAPS LOCK

    I'm starting to get tired of these 'Your data is leaking down the side of your device' PR bollocks.

    None of these scenarios seem the least bit practical. In this current one, the amount of RF noise from gubbins inside the box (switch mode PSU anybody?) and nearby makes this seem on the edge of fantasy.

    6 0 Reply
    1. Alister
      Reply Icon

      Re: I'm starting to get tired of these 'Your data is leaking down the side of your device'

      I was going to post exactly the same thing, the inside of these devices are so electronically noisy that the chances that you can effectively isolate the particular signal associated with encryption / decryption seems remote.

      2 0 Reply
    2. Anonymous Coward
      Reply Icon
      FAIL

      Re: I'm starting to get tired of these 'Your data is leaking down the side of your device'

      Don't bother me with anything that doesn't work from the parking lot 50 metres away.

      If someone is going to get into "close vicinity of the [device]”, it will be to do something more effective than installing something which will let them infer statistically some information about its internal cryptographic operation. If they want to know what's inside, they will probably just walk off with it.

      5 1 Reply
    3. Aodhhan
      Reply Icon

      Re: I'm starting to get tired of these ...

      This is because you haven't done a lot of research.

      Gathering information via electromagnetic signals from computer systems was being done by various intelligence agencies back in the 70s, without having to have internal access to the building which housed the computer system.

      Today, computer devices are everywhere. Most concerning are point of sale and point of interaction devices, ATMs, etc. They also give off EM signals, and I can stand in line next to you while you use them and pick up the signals. I don't need to be 50 feet away.

      At the bank, while a teller enters your information I can be in line, and again, pick up EM signals or at the ATM. While you're tapping away at Starbucks, etc. Laptops are made to be light weight, and have nothing which interferes with EM signals. Picking up keyboard signals can provide a malicious individual with a lot of information. Like credit card numbers, passwords, etc.

      Automatically discounting something without conducting research on it doesn't make a lot of sense. Just because you "believe" or "think" something cannot happen, doesn't make it true.

      3 1 Reply
      1. Anonymous Coward
        Reply Icon
        Stop

        Re: I'm starting to get tired of these ...

        Automatically discounting something without conducting research on it doesn't make a lot of sense. Just because you "believe" or "think" something cannot happen, doesn't make it true.

        This article talks about "near-field analysis". Near-field means they are looking at the radio signals no more than within about a wavelength from the emitter. Here the radiation from different parts of the device can be separately distinguished by its direction. Further than about a wavelength away, you are in the "far-field" where everything kind of merges together and the whole thing looks like one big emitter. Only in the near-field is it true that "These physical quantities and structures must necessarily have a ... spatial-extent."

        However, at 1 GHz clock speed the wavelength of the radio waves produced is about 15 cm, so you have to get closer than that - probably much closer. Higher clock speeds have even shorter wavelengths. And if you have a big earthed metal case in the way then that is going to completely mess up the near-field signal, by hugely weakening it, and making spacial information about anything that gets out (mainly through any holes in it) dominated by the characteristics of the case rather than what is inside. And what do we get for all this effort - this isn't your PIN number, the code being run, or anything useful like that. It is "statistical information". If you can get within six inches of my server AND get the lid off, it's not going to be to put a big directional receiver next to it.

        0 0 Reply
    4. The Man Who Fell To Earth Silver badge
      Reply Icon
      Boffin

      Re: I'm starting to get tired of...

      This type of stuff has been around for years, used by pirates to crack & then duplicate smart cards. There's a whole chip hardening industry based on it. One of the most famous academic papers on this came out of the University of Cambridge in 1999.

      https://www.cl.cam.ac.uk/~mgk25/sc99-tamper.pdf

      0 1 Reply
  2. Anonymous Coward
    Anonymous Coward

    Read history, spy history.

    Thinking these methods are not real suggests one has not read or does not know history. Reading EM fields is decades old and if such methods are not used today it would be due to even better methods being available.

    Of course that would be discounted by those who in the past would have argued against using lasers and other methods to listen in on conversations behind closed doors.

    1 1 Reply
  3. allthecoolshortnamesweretaken

    What about jamming devices added in order to drown any interestig signal in random noise?

    0 0 Reply
  4. Bucky 2
    Coat

    I've never tried to sniff my own storage. I doubt I'm flexible enough.

    0 0 Reply
  5. Ken 16 Silver badge
    Holmes

    Virtualised, tiered storage shared across a VM farm?

    It's a pain in the arse trying to find which application is using which storage device at the best of times, let alone using a pair of dowsing rods.

    0 0 Reply
  6. Badger Murphy

    Is not the complexity of the machine a major factor here?

    It would seem to me that it would be less trivial to isolate and capture the desired electromagnetic signal on a very simple machine, such as a credit card chip or a discrete component, but surely an entire server humming along has numerous sources of electromagnetic fields in its housing, all of which transmit at varying, cumulative, and harmonic amplitudes and frequencies.

    I am definitely not an expert on this subject, but this would definitely be non-trivial, as the article does state, and would require extensive knowledge of the specific machine being snooped.

    While I don't want to dismiss these findings out of hand as alarmist nonsense, I also don't think we should all start running for the bomb shelters, either. Those old examples cited are of much older, simpler, and slower machines with far fewer parts (albeit much larger ones).

    1 0 Reply
    1. Anonymous Coward
      Reply Icon
      Anonymous Coward

      Re: Is not the complexity of the machine a major factor here?

      Also toss in that much higher frequencies are generally involved which mitigates against propagation distance. Energy and information are mutable, where you have leakage of one, you have leakage of the other. It can be RF, thermal, etc. How you deal with it in your environment is what matters. A good survey with high-end equipment can give you a good idea if there is a problem. No way I can afford that here, though I've had use of it in the past.

      0 0 Reply
      1. Number6
        Reply Icon

        Re: Is not the complexity of the machine a major factor here?

        That can work against you if you've got a resonant slot in the case, or a cable the correct length. That might cause a fairly strong signal to be radiated at that frequency, with the data modulated on it.

        As for demonstrations, didn't the BBC do a programme some years back where they parked up a van full of technology and could generate an image of someone's CRT in an adjacent building? I've seen a TEMPEST demo, if they could do that sort of thing twenty years ago, with modern DSP and improvements in semiconductors generally, it wouldn't surprise me that such things can be done at an even finer level than before.

        0 0 Reply

POST COMMENT House rules

Not a member of The Register? Create a new account here.

Forgotten password ?

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like