Experimental Investigation in the Impact on Security of the Release Order of Defensive Algorithms

Abstract

AbstractIn the practical use of security mechanisms such as CAPTCHAs and spam filters, attackers and defenders exchange ‘victories,’ each celebrating (temporary) success in breaking and defending. While most of security mechanisms rely on a single algorithm as a defense mechanism, we propose an approach based on a set of algorithms as a defense mechanism. When studying sets of algorithms various issues arise about how to construct the algorithms and in which order or in which combination to release them. In this paper, we consider the question of whether the order in which a set of defensive algorithms is released has a significant impact on the time taken by attackers to break the combined set of algorithms. The rationale behind our approach is that attackers learn from their attempts, and that the release schedule of defensive mechanisms can be adjusted so as to impair that learning process. This paper introduces this problem. We show that our hypothesis holds for an experiment using several simplified but representative spam filter algorithms—that is, the order in which spam filters are released has a statistically significant impact on the time attackers take to break all algorithms.KeywordsRelease Order of Defensive AlgorithmslearningExperimentationSecurity and Protection