Performance Evaluation of k-means using CPU and GPU with oneAPI and OpenMP for Network Intrusion Detection
Resumo
This article describes the application of the k-means algorithm for network intrusion detection, evaluating both sequential and parallel CPU and GPU versions using oneAPI and OpenMP. Based on a network intrusion dataset (CIC-IDS2017), which contains millions of instances of modern attacks, the experiments showed that the parallel versions achieved significant performance gains over the sequential version. The good scalability of the CPU solutions and GPU acceleration stand out, with oneAPI performing slightly better than OpenMP. The results reinforce that parallelization is essential for applying k-means in contexts where the data volume is large.
Referências
L. Andrade Maciel, M. Alcântara Souza, and H. Cota de Freitas. Reconfigurable fpga-based k-means/k-modes architecture for network intrusion detection. IEEE Transactions on Circuits and Systems II: Express Briefs, 67(8):1459–1463, 2020. DOI: 10.1109/TCSII.2019.2939826.
J. Bhimani, M. Leeser, and N. Mi. Accelerating k-means clustering with parallel implementations and gpu computing. In 2015 IEEE High Performance Extreme Computing Conference (HPEC), pages 1–6, 2015. DOI: 10.1109/HPEC.2015.7322467.
Codeplay Software Ltd. / Intel. oneapi for nvidia® gpus. [link], 2025. Access on August 21, 2025.
J. MacQueen. Multivariate observations. In Proceedings ofthe 5th Berkeley Symposium on Mathematical Statistics and Probability, volume 1, pages 281–297, 1967.
oneAPI. oneapi programming guide. [link], 2025. Access on August 21, 2025.
OpenMP. The openmp api specification. [link], 2025. Access on August 21, 2025.
V. Pathak and V. S. Ananthanarayana. A novel multi-threaded k-means clustering approach for intrusion detection. In 2012 IEEE International Conference on Computer Science and Automation Engineering, pages 757–760, 2012. DOI: 10.1109/ICSESS.2012.6269577.
I. Sharafaldin, A. H. Lashkari, A. A. Ghorbani, et al. Toward generating a new intrusion detection dataset and intrusion traffic characterization. ICISSp, 1(2018):108–116, 2018.
C. Yin and S. Zhang. Parallel implementing improved k-means applied for image retrieval and anomaly detection. Multimedia Tools Appl., 76(16):16911–16927, Aug. 2017. ISSN 1380-7501.
J. Bhimani, M. Leeser, and N. Mi. Accelerating k-means clustering with parallel implementations and gpu computing. In 2015 IEEE High Performance Extreme Computing Conference (HPEC), pages 1–6, 2015. DOI: 10.1109/HPEC.2015.7322467.
Codeplay Software Ltd. / Intel. oneapi for nvidia® gpus. [link], 2025. Access on August 21, 2025.
J. MacQueen. Multivariate observations. In Proceedings ofthe 5th Berkeley Symposium on Mathematical Statistics and Probability, volume 1, pages 281–297, 1967.
oneAPI. oneapi programming guide. [link], 2025. Access on August 21, 2025.
OpenMP. The openmp api specification. [link], 2025. Access on August 21, 2025.
V. Pathak and V. S. Ananthanarayana. A novel multi-threaded k-means clustering approach for intrusion detection. In 2012 IEEE International Conference on Computer Science and Automation Engineering, pages 757–760, 2012. DOI: 10.1109/ICSESS.2012.6269577.
I. Sharafaldin, A. H. Lashkari, A. A. Ghorbani, et al. Toward generating a new intrusion detection dataset and intrusion traffic characterization. ICISSp, 1(2018):108–116, 2018.
C. Yin and S. Zhang. Parallel implementing improved k-means applied for image retrieval and anomaly detection. Multimedia Tools Appl., 76(16):16911–16927, Aug. 2017. ISSN 1380-7501.
Publicado
28/10/2025
Como Citar
COSTA, Laura Caetano; FRASSI, Luiz Fernando Antunes da Silva; FREITAS, Henrique Cota de.
Performance Evaluation of k-means using CPU and GPU with oneAPI and OpenMP for Network Intrusion Detection. In: WORKSHOP DE INICIAÇÃO CIENTÍFICA - SIMPÓSIO EM SISTEMAS COMPUTACIONAIS DE ALTO DESEMPENHO (SSCAD), 26. , 2025, Bonito/MS.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2025
.
p. 73-80.
DOI: https://doi.org/10.5753/sscad_estendido.2025.16262.
