Article: LSTM-CNN: a deep learning model for network intrusion detection in cloud infrastructures Journal: International Journal of Critical Infrastructures (IJCIS) 2024 Vol.20 No.6 pp.505 - 523 Abstract: In cloud computing, resources are shared and accessed over the internet to perform intended computations remotely to minimise infrastructure costs. The usage and dependency on the cloud network have increased, and the chances of invasion and loss of data and challenges to develop a reliable intrusion detection and prevention system (IDPS). The existing machine learning-based approaches require the manual extraction of features, which produces low accuracy and high computational time. Providing a secure network involves a framework based on multi-fold validation and privacy in information transmission. The deep learning-based network IDPS model has been proposed to handle the large volume of network traffic in the cloud. This paper proposes a tailored long short-term memory and convolution neural network (LSTM-CNN)-based approach to design a new IDS. The proposed model productively examines intrusions and generates alerts proficiently by incorporating users' information and conducting examinations to detect intrusions. The model's performance is assessed using accuracy, precision, F1-score and recall measures. The proposed model achieves outstanding performance with a test accuracy of 99.27%. Inderscience Publishers - linking academia, business and industry through research

Title: LSTM-CNN: a deep learning model for network intrusion detection in cloud infrastructures

Authors: Doddi Srilatha; N. Thillaiarasu

Addresses: School of Computing and Information Technology, REVA University, Bangalore, 560064, India; Department of CSE, Sreenidhi Institute of Science and Technology, Hyderabad, 501301, India ' School of Computing and Information Technology, REVA University, Bangalore, 560064, India

Abstract: In cloud computing, resources are shared and accessed over the internet to perform intended computations remotely to minimise infrastructure costs. The usage and dependency on the cloud network have increased, and the chances of invasion and loss of data and challenges to develop a reliable intrusion detection and prevention system (IDPS). The existing machine learning-based approaches require the manual extraction of features, which produces low accuracy and high computational time. Providing a secure network involves a framework based on multi-fold validation and privacy in information transmission. The deep learning-based network IDPS model has been proposed to handle the large volume of network traffic in the cloud. This paper proposes a tailored long short-term memory and convolution neural network (LSTM-CNN)-based approach to design a new IDS. The proposed model productively examines intrusions and generates alerts proficiently by incorporating users' information and conducting examinations to detect intrusions. The model's performance is assessed using accuracy, precision, F1-score and recall measures. The proposed model achieves outstanding performance with a test accuracy of 99.27%.

Keywords: cloud intelligent infrastructures; convolution neural network; intrusion detection and prevention; long short-term memory; random forest; neural network; security.

DOI: 10.1504/IJCIS.2024.142451

International Journal of Critical Infrastructures, 2024 Vol.20 No.6, pp.505 - 523

Received: 30 Jan 2023
Accepted: 16 Mar 2023
Published online: 01 Nov 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: LSTM-CNN: a deep learning model for network intrusion detection in cloud infrastructures

Keep up-to-date