Prevention of Information Security Incidents in Automated Information System

Abstract

The purpose of the article is to support the processes of preventing information security incidents in conditions of high uncertainty. Method: methods of mathematical (theoretical) computer science and fuzzy set theory. Result: an information security Incident, including a computer incident, is considered as a violation or termination of the functioning of an automated information system and (or) a violation of information stored and processed in this system, including those caused by a computer attack. Information descriptions are presented in the form of structured data about signs of computer attacks. Structured data is the final sequence of strings of symbols in a formal language. The Damerau-Levenstein editorial rule is proposed as a metric for measuring the distance between strings of characters from a particular alphabet. The possibility of presenting the semantics of information descriptions of attack features in the form of fuzzy sets is proved. Thresholds (degrees) of separation of fuzzy information descriptions are defined. The influence of semantic certainty of information descriptions of features (degrees of blurring of fuzzy information descriptions) on the decision-making about their identity (similarity) is evaluated. It is shown that the semantic component of information descriptions of signs of computer attacks presupposes the presence of some semantic metric (for its measurement and interpretation), which, as a rule, is formally poorly defined, ambiguously interpreted and characterized by uncertainty of the type of fuzziness, the presence of semantic information and the inability to directly apply a probabilistic measure to determine the degree of similarity of input and stored information descriptions of signs. An approach is proposed to identify fuzzy information descriptions of computer attacks and to apply methods for separating elements of reference sets on which these information descriptions are defined. It is shown that the results of the procedure for identifying fuzzy information descriptions of computer attacks depend on the degree of separation of the reference sets and on the indicators of semantic uncertainty of these descriptions