Back to Search
Start Over
Slow Subscribers: a novel IoT-MQTT based denial of service attack.
- Source :
-
Cluster Computing . Dec2023, Vol. 26 Issue 6, p3973-3984. 12p. - Publication Year :
- 2023
-
Abstract
- MQTT, a popular IoT messaging protocol, is frequently associated with numerous vulnerabilities, the majority of which are critical. Many IoT devices that utilize MQTT are susceptible to cyberattacks such as denial-of-service and buffer overflow. In this paper, we unveil a novel Denial of Service (DoS) attack in the MQTT protocol, referred to as Slow Subscribers, which has the potential to cause MQTT brokers to become single points of failure. Unlike existing MQTT DoS attacks, Slow Subscribers can occur on a single compromised node and could potentially disrupt a MQTT broker with minimal subscription permissions. We evaluated the reliability of Mosquitto and NanoMQ, two popular MQTT messaging brokers, to determine the effect of Slow Subscribers. According to the findings of our investigation, NanoMQ outperforms Mosquitto in response to the Slow Subscribers attack at QoS level 0. We also determine that the response to Slow Subscribers at QoS 2 is the worst for both broker implementations. In addition, the results of our experiments indicate that Eclipse Mosquitto achieves a higher rate of reliability than NanoMQ on cloud deployments whereas NanoMQ has proven to be well-suited for edge environments, especially edge IoT devices that require the use of QoS levels 0 and 1. Finally, we propose a Resilient Middleware for Message Queue Telemetry Transport (Remistry) framework that is capable of detecting misconfigurations while providing granular support of resource commitment errors, in particular the out-of-memory (OOM) problems for effectively mitigating the impact of Slow Subscribers attacks on MQTT brokers. [ABSTRACT FROM AUTHOR]
- Subjects :
- *DENIAL of service attacks
*CYBERTERRORISM
*QUANTUM cryptography
Subjects
Details
- Language :
- English
- ISSN :
- 13867857
- Volume :
- 26
- Issue :
- 6
- Database :
- Academic Search Index
- Journal :
- Cluster Computing
- Publication Type :
- Academic Journal
- Accession number :
- 173017175
- Full Text :
- https://doi.org/10.1007/s10586-022-03788-9