MQTT is widely adopted in IoT deployments but offers no built-in encryption or strong authentication. Existing solutions such as the IETF ACE-MQTT-TLS profile require dedicated authorization servers and dynamic token flows, raising deployment complexity for resource-constrained environments. This paper presents a practical secure MQTT architecture that combines Transport Layer Security (TLS) 1.3 with statically provisioned JSON Web Token (JWT) authorization, targeting ESP32-S3 and ESP32-C6 microcontrollers with hardware cryptographic acceleration. A Python-based broker management graphical user interface (GUI) automates Mosquitto configuration and JWT issuance, removing manual setup errors. Measured on real hardware, TLS 1.3 connection time is 315-380 ms and steady-state throughput decreases by only 9%, with an additional heap memory requirement of approximately 42 KB. The results confirm that multi-layered IoT security—encompassing mutual authentication, AES-GCM encrypted transport, and fine-grained access control—is feasible on commodity microcontrollers without specialized infrastructure.
H. Minasyan, Hripsime Hakobyan· 2026 6th International Confe...· 0 citations
Internet of Things deployments increasingly incorporate devices with heterogeneous capabilities connected through diverse communication technologies including WiFi and low-power wide-area networks. Current protocol architectures force designers to select either feature-rich protocols like MQTT [1] that provide reliable delivery with high overhead, or lightweight alternatives like CoAP [2] that minimize resource consumption at the cost of application features. This paper presents a security-aware protocol translation framework that enables dynamic switching between MQTT and CoAP based on real-time assessment of device battery state, network conditions, and message criticality. The framework leverages Platform Security Architecture [3] cryptographic primitives to maintain end-to-end security during protocol transitions without exposing plaintext at gateway nodes. Implementation on ESP32-C3 [5] and ESP32-C6 [6] platforms connected to Raspberry Pi gateways via WiFi and LoRa demonstrates energy consumption reductions of 42 percent compared to MQTT-only operation while maintaining message delivery rates above 99.4 percent. The system achieves protocol switching in under 380 milliseconds with hardware-accelerated cryptographic operations consuming less than 3 percent of total device energy budget. Comparative evaluation across multiple network configurations validates the effectiveness of adaptive protocol selection for resourceconstrained IoT deployments.
H. Minasyan, Mane Araqelyan· 2026 6th International Confe...· 0 citations