Smart Networks Traffic Analysis and Anomaly Detection using Bidirectional LSTM Networks with CUDA Kernel Utilization

• Francisco Erialdo Domingos Freitas IFCE
• Antônio Wendell de Oliveira Rodrigues IFCE

DOI: https://doi.org/10.5753/erad-ne.2026.26676

Resumo

Modern networks, particularly those involving the Internet of Things (IoT), face significant challenges in traffic management, classification, and security. Efficient traffic classification and anomaly detection are critical for protecting these networks against malicious activities. In this work, we propose an approach based on Bidirectional Long Short-Term Memory (Bi-LSTM) networks, leveraging CUDA acceleration in PyTorch to optimize training performance and model accuracy. Our method captures contextual information from both past and future sequences, enhancing detection capabilities. Experimental results on an industrial IoT dataset demonstrate superior accuracy, recall, and F1 score compared to conventional LSTM models, highlighting the potential of the proposed solution for improving security and reliability in smart network environments.

Referências

Abbas, S., Alsubai, S., Ojo, S., Sampedro, G., Almadhor, A., Hejaili, A., and Bouazzi, I. (2024). An efficient deep recurrent neural network for detection of cyberattacks in realistic iot environment. The Journal of Supercomputing, 80:1–19.

Azab, A., Khasawneh, M., Alrabaee, S., Choo, K.-K. R., and Sarsour, M. (2024). Network traffic classification: Techniques, datasets, and challenges. Digital Communications and Networks, 10(3):676–692.

Bai, H. (2023). Iot-anomaly-detection. Accessed: 2024-05-17.

Gale, T., Zaharia, M., Young, C., and Elsen, E. (2020). Sparse gpu kernels for deep learning. In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, SC ’20. IEEE Press.

Keirsbilck, M. V., Keller, A., and Yang, X. (2019). Rethinking full connectivity in recurrent neural networks. ArXiv, abs/1905.12340.

Rajasinghe, N., Samarabandu, J., and Wang, X. (2018). Insecs-dcs: A highly customizable network intrusion dataset creation framework. In 2018 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE).

Rizi, M. H. P. and Seno, S. A. H. (2022). A systematic review of technologies and solutions to improve security and privacy protection of citizens in the smart city. Internet of Things by Elsevier, 20:100584.

Tayfour, O. E., Mubarakali, A., Tayfour, A. E., Marsono, M. N., Hassan, E., and Abdelrahman, A. M. (2023). Adapting deep learning-lstm method using optimized dataset in sdn controller for secure iot. Soft Computing.

Yang, M., Moon, J., Yang, S., Oh, H., Lee, S., Kim, Y., and Jeong, J. (2022). Design and implementation of an explainable bidirectional lstm model based on transition system approach for cooperative ai-workers. Applied Sciences (Switzerland), 12(13).

Zhao, Y. (2023). Complete guide to rnn, lstm, and bidirectional lstm. last accessed: 2024-05-23.