Jump to content

Physical unclonable function

From Simple English Wikipedia, the free encyclopedia
Revision as of 08:22, 2 December 2017 by 5.147.251.186 (talk) (Created page with "Physical Unclonable Functions (PUFs) are based on "locked" randomness (chaos), stemming from tiny imperfections in the manufacturing process of hardware, which result in the p...")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Physical Unclonable Functions (PUFs) are based on "locked" randomness (chaos), stemming from tiny imperfections in the manufacturing process of hardware, which result in the production (output) of a bitstring which appears random and can be considered as unique per hardware instance (device).

Essentially, electrical current is allowed to go in two distinct ways through a circuit, such as two wires, and if it goes mostly through one, it signifies a logical zero, and if it goes through the other, it signifies a logical one. However, the prevalent path is chosen according to tiny details, such as which wire is thicker or longer. However, the two wires have been designed to be exactly the same, and only differ because of imperfections of the manufacturing process, which cause random tiny mistakes, such as one wire being 5 mm, and the other 5.00001 mm. Therefore, and as the choice is dependent on a large number of tiny details, which are very difficult to determine beforehand, this choice appears as random, and can be characterised as chaotic. These tiny details that affect the choice, cannot be predicted as they happen randomly in the manufacturing process. Nevertheless, after they happen, they stay like that, e.g., the small difference in length between two wires will always exist after it is manufactured like this, so the aforementioned choice is "locked", although it is "random" (chaotic), and will always give the same result.

Since, most contemporary hardware consists of millions of circuits, millions of individual choices can be made at the same time, resulting in a string (series) of bits (zeros and ones) of a very large length. Since each choice is independent and "random", the probability that two such strings will be the same becomes extremely small due to their very large length. Therefore, a PUF can be used to produce a large bitstring that will be unique per device and always the same, when measured under exactly the same conditions. Such a bit string can then be used in cryptography as an inherently occurring cryptographic token, such as a key for encryption. As external environment conditions may affect the result of a PUF (its "random" unique bit string output), error correction may be needed to force this result to be the same between different measurements.