Аналіз сучасних систем виявлення та запобігання вторгнень

Abstract

This work is devoted to comparative analysis of modern intrusion detection and prevention systems (IDS/IPS), both open source and commercial. The systems were evaluated according to various characteristics, including ML/AI support, the presence of DPI L7, and control types. As a theoretical basis for evaluating the effectiveness of ML-based IDS, a confusion matrix is introduced, with an emphasis on the trade-off between recall and false positive rate inherent in each ML paradigm. The result demonstrates a clear functional division between open and commercial solutions and provides practical recommendations for choosing a system according to your needs.