Proper

Abstract

One of the important requirements in the Internet of Things (IoT) environment is to provide a differentiated Quality of Service (QoS) for each message. However, since IoT nodes generally have limited computing power and memory resources, and the amount of messages exchanged in the system is very large, applying existing scheduling algorithms to the IoT environment is not suitable. In this paper, we propose a Weighted Unfair Queue Scheduling (WUFQ) algorithm, which is a lightweight scheduling algorithm suitable for theIoT environment. WUFQ divides incoming and outgoing messages into three groups, UNC (Unconditional), RT (Real Time), and DT (Delay Tolerant), in order to perform the task of scheduling. Messages in the UNC group are sent without any scheduling, while for the messages in the remaining two groups, the transmission time is determined by increasing or decreasing the weight of the queue to which they belong. In order to verify the efficiency of the proposed WUFQ algorithm, the results of a performance comparison with the multi-class Q-Learning (MQL) scheduling algorithm are presented. The experimental results show that the proposed WUFQ scheduling algorithm transmits messages of UNC group faster than MQL.