Comprehensive anomaly detection model for distributed systems using spectral and biometric methods

Introduction. The article discusses the integration of spectral analysis of network behavior and modern biometric authentication methods.

Goal. The goal of the research is to improve the security of distributed computing systems. Further research prospects include the use of wavelet transformations for analyzing non-stationary traffic, and the adaptation of detection thresholds based on the user's context.

Materials and methods. The model includes the use of Fourier series to analyze the periodicity of network processes, as well as the use of FaceNet, VGGFace, and IrisCode for the correct identification of users. A correlation model is presented that combines the biometric and spectral characteristics of a user, allowing for the formation of an integrated risk indicator.

Results and discussion. A numerical analysis has been conducted to demonstrate the possibility of separating normal users and attackers based on the spectral characteristics of their behavior. The graphical results confirm the effectiveness of the proposed model.

Conclusion. Based on the results of this study, it can be concluded that the use of Fourier series decomposition allows for the formalization of a user's behavioral profile in the frequency domain and the identification of high-frequency components characteristic of automated attacks.

1. Oppengeim A. Tsifrovaya obrabotka signalov. Prentice Hall, 2010.

2. Schroff F., Kalenichenko D., Philbin J. FaceNet: Unified Embedding for Face Recognition and Clustering. Proc. IEEE CVPR. 2015. pp. 815–823.

3. Parkhi O., Vedaldi A., Zisserman A. Deep Face Recognition. Proc. BMVC. Oxford Univ., 2015.

4. Daugman J. How IrisCode Works. IEEE Trans. Circuits and Systems Video Technology. 2003. Vol. 14. pp. 21–30.

5. Tanenbaum EH., Van Stin M. Raspredelennye sistemy: printsipy i paradigmy. Pearson, 2020.

6. Stollings V. Osnovy setevoi bezopasnosti. Pearson, 2018.

7. Chen L., Sun M., Zhao Q. Hybrid Fourier-Wavelet Models for Encrypted Traffic Analysis. Computers & Security. 2023.

8. Mallat S. A Wavelet Tour of Signal Processing. Academic Press, 2008.

9. Nguyen T., Vo M., Huynh D. FaceNet++: Enhanced Embedding Representations. Pattern Recognition Letters. 2024.

10. ISO/IEC 19794-5:2011. Informatsionnye tekhnologii. Biometricheskie formaty obmena dannymi. Chast' 5: Geometricheskie dannye litsa.

11. Zhang Y., Liu H., Wang P. Fourier-Based Profiling of Network Behavior. IEEE Access. 2021.

12. Kumar S., Patel R. Deep Biometric Fusion for Anomaly Detection. Journal of Network and Computer Applications. 2022.