Research Area:  Internet of Things

Abstract:

In MQTT, a retained message allows new subscribers to receive the latest status update without waiting for the publisher to send a new message. However, because the retained message is delivered regardless of when it was originally published, it can become outdated, leading to potential issues such as production inefficiency. To address this challenge, this paper proposes an Age-of-Information (AoI)-aware retained message policy in MQTT-based IoT networks (ARMY). In ARMY, the broker evaluates AoI (i.e., freshness) of the retained message before delivering it to new subscribers. Specifically, if the retained message is outdated, the broker requests an updated message from the publisher before forwarding it to the subscribers, thereby reducing the AoI. This approach is particularly beneficial in scenarios where real-time data is crucial, such as industrial control systems. However, this process involves additional signaling overhead, so an optimal policy balancing AoI and signaling overhead is necessary. We formulate a Markov Decision Process (MDP) model and determine the optimal policy using Q-learning (QL). Simulation results show that ARMY significantly improves the average reward and AoI satisfaction ratio without incurring significant additional signaling costs compared to the comparison schemes across various settings.

Keywords:  

Author(s) Name:  Youngjun Kim, Yeunwoong Kyung

Publisher name:  ResearchGate

DOI:  10.1109/JSEN.2024.3465025

Volume Information:  volume 35,(2024)

Paper Link:   https://www.researchgate.net/publication/384387042_AoI-aware_Retained_Message_Policy_in_MQTT-based_IoT_Networks