EVALUATION OF THE PROBABILITY OF SOCIAL ENGINEERING ATTACK PROPAGATION IN A CORPORATE INFORMATION SYSTEM WITH ACCOUNT FOR THE DIRECTIONALITY OF COMMUNICATION INTERACTION

Abstract

This article addresses the improvement of an approach to evaluating the propagation of a social engineering attack in a corporate information system by accounting for the directionality of communication interaction between users. The relevance of the study stems from the fact that, in modeling the propagation of social engineering influence, attack success is determined not only by the characteristics of individual users, but also by the structure of the links between them. In this context, the communication component is of particular importance. The aim of the article is to improve the approach to evaluating the propagation of a social engineering attack in a corporate information system by accounting for the directionality of communication interaction between users and to determine the effect of this refinement on the estimated probabilities of compromise. The study employs a graph-based representation of user interaction, probabilistic modeling of attack propagation between connected nodes, and a comparative analysis of the results obtained for symmetric and asymmetric graph variants under different trajectory cutoff thresholds. A decomposition of the total communication intensity between a pair of users into two directed components using a directionality coefficient is proposed, which makes it possible to formalize the uneven distribution of communication activity across opposite directions of interaction. The modeling results demonstrate that accounting for the directionality of communication interaction leads to lower final estimates of compromise probability compared with the symmetric graph representation and increases the sensitivity of the model to the trajectory cutoff threshold. It is shown that an asymmetric representation of the communication component provides a more selective and structurally adequate description of the conditions under which a social engineering attack propagates. The practical significance of the obtained results lies in the possibility of applying the proposed approach to evaluate the probability of social engineering influence propagation in corporate information systems.

Keywords: information security management, cybersecurity, social engineering attack, corporate information system, graph model, compromise probability.