Can any contemporary Mac boot from a CD? I think Jaguar was the last system that came on CDs. Minor quibble, but the language in this story looks just a little bit off.
If it were just a matter of getting already-infected machines to talk to each other, why use audio? The vast majority of computers nowadays have Internet connections-- It’s far more likely that you’ll be connected to the network somehow than it is that you’ll be in the same room as another computer.
I’m calling hoax, and not even a particularly good one.
Obviously, Slashdot isn’t the origin. But I am suspecting they are in on the joke and double posted it as part of their “fun”.
Also, I was using “BIOS” in the general form. UEFI (not EFI) is a BIOS system. It’s just not the same as the BIOS in IBM-compatible PCs of yore. It’s like assuming that someone writing DOS means MS-DOS as opposed to any Disk Operating System.
The OP suggests that the infection vector might be through the computer’s microphone: this is patently absurd, computers do not monitor and analyze any ultrasonic data that shows up on the porch. The article states that already infected machines appeared to be able to fall back on audio signals to communicate with each other and keep the infection updated against attack. Even that is a little far-fetched, though remotely believable.
The idea that an infection could propagate acoustically requires that one believe that operating systems are being required to be built with back-doors in the sound system, presumably for NSA to have ready access. This sounds blazingly paranoid to me.
You’re forgetting the likelihood that all computers would in the room would have multi-directional microphones that could pick up the signal.
Now, I wouldn’t be surprised if the machines do actually whisper, as part of a prank. Macs all have Wi-fi, and many PCs do, too, so just use that for actual communication. Or, even easier, just hide a second infection that has a preprogrammed way of dealing with things, so it’s not actually updating between computers at all.
I could also see this sort of thing being done as proof of concept in conditions that were well-known ahead of time. But then I’d also expect that the people who had done it would be pretty proud of it and wouldn’t make people merely guess that this is what is happening.
But no general purpose malware is going to work this way. It’s just way too much complexity of code for such a low benefit.
The actual technology of sending data (at fairly low speeds) acoustically is close to trivial at this point. I’d posit that 1 kbit/s is very doable, even in the presence of heavy noise and speaker/microphone nonlinearities. I could write a program to do it at even lower speeds (~100 b/s) in an afternoon.
Everything else is within epsilon of complete BS (though as others have said, the author never claims that an uninfected machine can become infected using this scheme–only that infected machines can talk to each other via this method).
Acoustic data transfer is trivially easy to do in controlled conditions. Doing it open air given the crappy quality of notebook speakers and microphones trying to send and receive ultrasonic messages across meters of open space in a noisy environments is IMO another kettle of fish. Why don’t you try it? It would be q spectacular experiment and would be interesting proof of concept news that would be of interest to a lot of magazines. If you can do it you’d get some major ink.
I’m too busy working on open-air communication over hundreds of kilometers :).
That said, I figured I’d do a simple test for kicks. I played alternating 20 kHz and 22 kHz waveforms (ultrasonic to people over 12 years old, at least :)) from my desktop computer while playing a song in the background. My laptop with a crappy built-in microphone received it from 4 meters away. Here’s the spectrogram of the received signal. As you can see, the signal shows up plain as day. That’s a very low bit rate (2 bps), but nevertheless proves the concept. With a bit of processing and error correction, I have no doubt one could go much higher, even in the presence of much more noise.
I have no idea if this specific story is true, but I would be very surprised if something like this didn’t already exist. If you have a sophisticated platform for delivering payloads of data, you want to be able to take advantage of every delivery path available to you. Imagine giving, say, Stuxnet another way of updating its own code in a distributed fashion.
In other words, if major intelligence agencies don’t already do something like this, they haven’t been doing their jobs.
Yes, and this applies to any kind of data. I sometimes get the impression that non-technical people imagine that merely having a computer connected to <something else> makes it vulnerable to hacking from <something else>. But in fact, there has to be some program running on the computer that, by design or by accident, responds to data coming from <somewhere else> (in the case of malware, it is often by accident), and the data has to be of the particular form that the program expects. The computer will not respond otherwise. Computers are unresponsive bricks unless you explicitly program them to respond to something.
Maybe it’s just me, but whenever I’m reading a story and I come across the phrase, “high frequency transmissions,” I translate it as “undiagnosed schizophrenia.”
Many people, including the OP, are misinterpreting the article’s claims about sonic data transmission.
No one is claiming that a computer can be infected over an audio channel. What the article is claiming is that two infected computers can communicate over an audio channel. That’s certainly a novel feature, if true, but it doesn’t require any kind of technomagic. If both systems are compromised, then the badBIOS can control both speakers and microphones and send/receive data between them.
You use audio because people “air gap” sensitive systems, keeping them isolated from networks, or keeping two different networks isolated from each other.
This is one way around that problem. Yes, it requires an initial, separate infection vector, but it can allow for some data transfer types.
For example, a company may have a machine that can be used for employee internet access, and then a system used to store billing records. They’re kept separate per PCI requirements, but say someone with an infected USB stick manages to infect both. Bad guy now has access to the internet PC, but the interesting stuff is on the billing network. An acoustic link would allow you to bridge to that network.
Now, in this scenario, you wouldn’t be pulling down the entire database or anything, but if you were interested in specific records, it seems like it would work.
Now, that’s not to say that this particular bit of malware is actually doing any of this, that’s not yet known, but there would definitely be great utility in being able to do it.
And if a computer is already infected with a malware program, that program could be listening to the microphone input and acting upon it. Conversely, it could be sending out data over the speaker, although it’s hard to imagine it not being heard by a nearby human, since computers speakers don’t have a wider audio range than human hearing.
I could almost see this being practical, because many PC systems have Realtek onboard audio chipsets so there would be a standard device family to interface to. Or really Windows API system calls and/or DirectX API would probably be much easier, and not dependent upon specific devices.
Whether the machines have speakers or more importantly microphones is another matter though. Laptops would probably have both, but it seems unlikely that business PCs would have microphones.
I read that article a couple of days ago and it triggered all my skeptical instincts. The virus they were describing did at least three things that were either impossible or insanely technically difficult. It somehow crept into BIOS and persisted despite reflashing of the BIOS. The article I read implied that it could creep onto USB sticks that weren’t even mounted and irredeemably corrupt them. And then there’s the air gap claim. And the claim that the “virus” self-repairs itself.
Well, the thing of it is, even technically possibly feasible stuff like ultrasonic jumping of air gaps would need to run some fairly sophisticated code. Think about what’s involved – you’d need some daemon process running to monitor the microphone and pick up transmissions, do error checking, patch the received transmissions into some kind of coherent message, and then execute that message, presumably with administrator privileges. So even if hidden within gigabytes of memory, that code would be accessible and readable, particularly if it were placing itself on USB drives.
If anyone remembers, say, faxes or acoustic modems, the baud rate of air gap transmission would be incredibly low. Not to mention that it would be trivially easy to detect whether there * were * ultrasonic transmissions happening once you suspected it. What one machine can detect, another can as well.
To me, it sounds like a computer version of Morgellon's syndrome. They seem to be blaming every software and hardware malfunction on an unknown virus that's very, very good at hiding its tracks.
So, not being a specialist in computer security, I am skeptical -- anything this sophisticated would almost have to be baked into the hardware of the system, which (when you consider the recent revelations about the NSA) is not, I suppose, utterly beyond the realms of possiblity.Ever tried making a Hackintosh, ie. install MacOS on a PC?
On some systems it requires some heavy tweaking just to get the installation started and in the end it is not guaranteed that it will work. On others systems is it outright impossible to even start the installation.
I find it hard to believe that there’s a single piece of code that can effortlessly jump from PC to Mac to BSD to Linux etc.
thanks to all for the comments.
This story certainly is outlandish. Whether it is based on fact or not, we may never know.
What bothers me is the most recent update to most of the apps on my tablet included the <permission to modify add and delete data from storage> - I don’t want something able to delete without my specific permission. I am considering returning my tablet to factory and removing my apps and very carefully selecting what goes on. I think I can get nonstandard music playing and book reading apps that won’t be able to screw around with my data without asking permission. Alternately, I can get my buddy Chris to write me some apps.
That sounds bizarrely paranoid. Don’t most apps have to add and delete data from storage in order to do anything useful? That certainly, for example, sounds like the definition of a game’s save file.
If you’re really worried, your best bet would be to back the tablet up on a regular basis.