Low-Latency Intrusion Detection Using a Deep Neural Network

Intrusion detection systems (IDSs) must be implemented across the network to identify and avoid attacks to counter the emerging tactics and techniques employed by hackers. In this research, we propose a lightweight IDS for improving IDS efficiency and reducing attack detection execution time. We use...

Full description

Saved in:
Bibliographic Details
Published in:IT professional 2022-05, Vol.24 (3), p.67-72
Main Authors: Ahmad, Umair Bin, Akram, Muhammad Arslan, Mian, Adnan Noor
Format: Article
Language:English
Subjects:
Algorithms
Artificial neural networks
Computer hacking
Deep learning
Intrusion detection
Intrusion detection systems
Network latency
Neural networks
Radio frequency
Tactics
TCP/IP (protocol)
TCPIP
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Intrusion detection systems (IDSs) must be implemented across the network to identify and avoid attacks to counter the emerging tactics and techniques employed by hackers. In this research, we propose a lightweight IDS for improving IDS efficiency and reducing attack detection execution time. We use an RF algorithm to rank features in order of importance and then reduce data dimension by selecting the top 15 important features. We then implement a deep neural network (DNN) architecture to classify anonymous traffic by analyzing TCP/IP packets on the network security laboratory-knowledge discovery in databases (NSL-KDD) dataset. The results indicate that our proposed technique of applying RF to identify important features can improve the DNN-based IDS system’s execution time and performance and has low latency for intrusion detection as compared to other state-of-the-art techniques.
ISSN:1520-9202
1941-045X
DOI:10.1109/MITP.2022.3154234