Browse by author
Lookup NU author(s): Dr Ehsan Toreini, Professor Brian RandellORCiD, Professor Feng Hao
Breaking the encrypted message traffic from the German Enigma cipher machine was one of the key allied achievements of World War II, performed at Bletchley Park by a team led by Alan Turing. The work described in this paper was motivated by the historic significance of the Enigma, and the fact that we had the possibility of gaining access to one. This led to the realisation that it would be intellectually interesting to investigate the possible effectiveness of a "side-channel" attack on the machine that exploited the noise made by the act of typing the source plaintext. Much has been written on the cryptologic aspects and the historic impact of the Enigma, but we are unaware of any previous account of an investigation into its susceptibility to an acoustic side-channel attack. The Enigma keyboard differs greatly from a modern keyboard, and is particularly noisy, due to movements of one or more of the internal rotors and associated machinery in response to each key press. We applied state-of-the-art signal processing and machine learning techniques to investigate the possibility of identifying the individual Enigma keys from the noises they made; the outcome was a demonstration that such identification could be reliably achieved with a success rate of 84% (as opposed to 3.8% by random guess), using a simple microphone and a personal computer.
Author(s): Toreini E, Randell B, Hao F
Publication type: Report
Publication status: Published
Series Title: School of Computing Science Technical Report Series
Year: 2015
Pages: 23
Print publication date: 01/03/2015
Report Number: 1455
Institution: School of Computing Science, University of Newcastle upon Tyne
Place Published: Newcastle upon Tyne
URL: http://www.cs.ncl.ac.uk/publications/trs/papers/1455.pdf
