Noises promise the ultimate cipher

While quantum-based cryptography is still at its very infancy, Kish and his team at the University of Szeged (Hungary) proposed a new device that is believed to enable an even stronger level of security.

The basic principle of quantum cyrptography relies on quantum particles to avoid the information being eavesdrop. The idea sounds good, but the engineering process seems to be moving much slowlier. Furthermore, new researches already indicated some potential vulnerabilities with such systems, had it not been properly implemented.

Kish's solution relies very much in physics, but Newton/Einstein type of physics. He proposed to use thermal noises to securely transmit a message (could be the entire message, or mostly the encryption key). Basically, when an electric current passes through a wire (from point A to point B), the thermal noise is generated and can be measured at both points. In Kish's design, sender and receiver both possess two resistors (one with low resitance, and the other with high resistance).

To transmit a bit, for example, A and B randomly (and independently) chose one resistor to use. Half of the time, they both chose high or low resistor, causing very high or very low thermal noises. Both parties would be aware of that situation and re-transmit. When the pair of resistors are different, an intermediate/friendly level of thermal noises will be measured at both points. Depending on its chosen resistor, one can determine what resistor the other party has chosen on the other side. Consequently, one bit is transmitted successfully.

This technique is resistant to eavesdropping, because the attacker can only know if a message is being transferred. He can not deduce which bit, as he does not know which pair of resistors were being used. Futhermore, even an attempt of eavesdropping would change the level of thermal noises. In other words, it alarms A and B that someone is listening, and they may decide to terminate the communication.

The idea sounds much simpler (and more effective) than the quantum-based technique, and it already had some prototype device, costing less than $100. There are still issues (mostly engineering-related, i think); but it may well replace the quantum solutions, and probably many currently established cryptographic protocols.

(Source: NewScientist).