Optimality of non-adaptive strategies: The case of parallel games

Most cryptographic security proofs require showing that two systems are indistinguishable. A central tool in such proofs is that of a game, where winning the game means provoking a certain condition, and it is shown that the two systems considered cannot be distinguished unless this condition is pro...

Full description

Saved in:
Bibliographic Details
Published in2014 IEEE International Symposium on Information Theory pp. 1707 - 1711
Main Authors Demay, Gregory, Gazi, Peter, Maurer, Ueli, Tackmann, Bjorn
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.06.2014
Subjects
Computer science
Cryptography
Games
Information theory
Probability distribution
Random variables
Online AccessGet full text

Cover

More Information
Summary:Most cryptographic security proofs require showing that two systems are indistinguishable. A central tool in such proofs is that of a game, where winning the game means provoking a certain condition, and it is shown that the two systems considered cannot be distinguished unless this condition is provoked. Upper bounding the probability of winning such a game, i.e., provoking this condition, for an arbitrary strategy is usually hard, except in the special case where the best strategy for winning such a game is known to be non-adaptive. A sufficient criterion for ensuring the optimality of non-adaptive strategies is that of conditional equivalence to a system, a notion introduced in [1]. In this paper, we show that this criterion is not necessary to ensure the optimality of non-adaptive strategies by giving two results of independent interest: 1) the optimality of non-adaptive strategies is not preserved under parallel composition; 2) in contrast, conditional equivalence is preserved under parallel composition.
ISSN:2157-8095
2157-8117
DOI:10.1109/ISIT.2014.6875125