Cloud-edge MQTT messaging for latency mitigation and broker memory footprint reduction.

Area of Science:

• Computer Science
• Networking
• Internet of Things (IoT)

Background:

• Smart city applications increase the number of connected Internet of Things (IoT) devices.
• Broker-based communication using MQTT is common for IoT data delivery.
• Default MQTT lacks size differentiation, leading to latency and memory issues.

Purpose of the Study:

• To address size-dependent latency and memory overload in MQTT brokers for cloud-edge IoT.
• To propose and evaluate novel message scheduling and memory allocation schemes.
• To improve the performance and efficiency of MQTT in smart city environments.

Main Methods:

• Analysis of problems in cloud-edge messaging for IoT.
• Design and implementation of three subscriber scheduling policies and a broker memory allocation scheme.
• Validation using Eclipse Mosquitto, with empirical and analytical performance evaluation.

Main Results:

• Demonstrated performance improvements in latency mitigation.
• Empirically showed up to 75% reduction in runtime broker memory footprint.
• Validated the effectiveness of the proposed cloud-edge messaging design.

Conclusions:

• The proposed design effectively mitigates latency and reduces memory consumption in MQTT brokers.
• The optimized MQTT approach is suitable for demanding smart city IoT applications.
• The design's applicability to other messaging services warrants further investigation.