Implementation of Event-Based Dynamic Authentication on MQTT Protocol

Autor: Rizka Reza Pahlevi, Parman Sukarno, Bayu Erfianto
Jazyk: English<br />Indonesian
Rok vydání: 2019
Předmět:
Zdroj: Jurnal Rekayasa Elektrika, Vol 15, Iss 2 (2019)
Druh dokumentu: article
ISSN: 1412-4785
2252-620X
DOI: 10.17529/jre.v15i2.13963
Popis: This paper proposes an authentication mechanism on the MQ Telemetry Transport (MQTT) protocol. The exchange of data in the IoT system became an important activity. The MQTT protocol is a fast and lightweight communication protocol for IoT. One of the problems with the MQTT protocol is that there is no security mechanism in the initial setup. One security attack may occur during the client registration phase. The client registration phase has a vulnerability to accept false clients due to the absence of an authentication mechanism. An authentication mechanism has been previously made using Transport Layer Security (TLS). However, the TLS mechanism consumes more than 100 KB of data memory and is not suitable for devices that have limitations. Therefore, a suitable authentication mechanism for constraint devices is required. This paper proposes a protocol for authentication mechanisms using dynamic and event-based authentication for the MQTT protocol. The eventbased is used to reduce the computing burden of constraint devices. Dynamic usage is intended to provide different authentication properties for each session so that it can improve authentication security. As results, the applied of the event-based dynamic authentication protocol was successful in the constraint devices of microcontrollers and broker. The microcontroller, as a client, is able to process the proposed protocol. The client uses 52% of the memory for the proposed protocol and only consumes 2% higher than the protocol without security. The broker can find authentic clients and constraint devices capable of computing to carry out mutual authentication processes to clients. The broker uses a maximum of 4.3 MB of real memory and a maximum CPU usage of 3.7%.
Databáze: Directory of Open Access Journals