Monday, March 23, 2015

Jumping the air-gap

Computers are said to be “air-gapped” when they are not connected to the internet or any other computer by physical or wireless connections. Air-gaped computers are usually used to keep sensitive information safe from the dangers of the internet often by journalists, governments or credit card companies and banks. Now there is a new proof of concept released by Isreali researchers that allows systems to communicate across the air-gap.

There have been a few other such concepts to communicte across the air-gap but most (if not all) of them have been a one-way communication, this new method is bi-directional and it uses the heat sensors that are inside all computers.

The attack makes it possible for an air-gapped computer to communicate with another by generating heat. One computer generates and dissipates heat while the other uses its heat sensors to detect the change in heat of the other computer (this means they need to be somewhat close). The communication is in binary (a lot of heat is a 1 and normal heat is a 0) and since generating and dissapating heat takes time this communication becomes very slow.

In order for this to be carried out as an attack the air-gapped computer would need to be infected with a piece of malware as well as another computer (preferrably internet connected). Then the two computers need to be brought into close proximity in order to communicate (the computers will need to remain in proximity for some time in order to transfer anythin worthwhile. Despite these strict requirements many of them may be more possible than they sound. The article from Wired mentions that in many places a worker will have an air-gapped computer right next to an internet connected one (for easy use at a desk), then all that is needed is to infect the internet connected computer and maybe a flash drive the worker uses with the air-gapped computer. Once both computers are infected the attacker can transfer data or issue simple commands to the air-gapped computer.

I think this proof of concept is very interesting and forces us to think about what we consider secure. Is it enough for a “secure” system to be disconnected from the internet? Should it also be kept in a room that blocks radio waves? Maybe isolated from everything else possible? Then we need to solve the problem that the system is hard to work with.


Security is an ever-evolving field and there may be no way to ever achieve “perfect security” but we still need to try. Locks are not perfectly secure either, I have heard it said that locks are to keep the honest people out, and that if someone really wants to steal your things a lock is not going to stop them. The idea is that the more difficult something is to break into less people will try, an attacker needs to be more motivated to try to get into a really secure system. At this point I think air-gapped systems are still safe from all but the most powerful and motivated groups (the NSA being one of those).

No comments:

Post a Comment