An efficient intrusion detection system for MQTT-IoT using enhanced chaotic salp swarm algorithm and LightGBM

Internet of Things (IoT) networks are becoming increasingly popular for monitoring critical environments of various types. For the communication of IoT devices, several lightweight protocols have been developed. MQTT (message queuing telemetry transport) is a widely used messaging protocol in IoT us...

Full description

Saved in:
Bibliographic Details
Published in:International journal of information security 2022-12, Vol.21 (6), p.1263-1282
Main Authors: Prajisha, C., Vasudevan, A. R.
Format: Article
Language:English
Subjects:
Accuracy
Algorithms
Classifiers
Coding and Information Theory
Communications Engineering
Computer Communication Networks
Computer Science
Cryptology
Internet of Things
Intrusion detection systems
Management of Computing and Information Systems
Networks
Operating Systems
Optimization
Queueing
Regular Contribution
Telemetry
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:Internet of Things (IoT) networks are becoming increasingly popular for monitoring critical environments of various types. For the communication of IoT devices, several lightweight protocols have been developed. MQTT (message queuing telemetry transport) is a widely used messaging protocol in IoT using the publish–subscribe technique. The openness of publish–subscribe model and the limited built-in authentication capabilities makes it vulnerable to intruders. Hence, an intrusion detection system (IDS) has a vital role in MQTT-IoT security. This paper proposes an efficient intrusion detection mechanism for the MQTT-IoT networks using an enhanced chaotic salp swarm optimization algorithm (ECSSA) and LightGBM classifier. Traditional IDS uses a lot of irrelevant data and undesirable attributes, resulting in long detection times and low performance. To overcome the limitations, the proposed IDS uses ECSSA for feature selection and LightGBM classifier for better detection accuracy. The experimental verification on MC-IoT, MQTT-IoT-IDS2020, and MQTTset datasets demonstrates that ECSSA and LightGBM improve the overall accuracy rate. The proposed technique outperforms the existing approaches with an accuracy of 99.38%, 98.91%, and 98.35% in the three test sets, MC-IoT, MQTT-IoT-IDS2020, and MQTTset, respectively.
ISSN:1615-5262
1615-5270
DOI:10.1007/s10207-022-00611-9