摘要
物联网环境下的数据融合要求采集和处理各种具有不同时间特性的数据,包括实时数据和非实时数据。数据的实时交付作为数据融合的前提对融合结果有着极其重要的影响。论文提出了一种基于动态队列感知的分布式调度算法(QACSMA),该机制考虑了数据的丢包率、时延约束和队列预测占用率,利用贪婪算法动态调整竞争窗口,以在链路队列长度变大时,提高长队列的服务效率,降低队列长度和时延,提高吞吐量。仿真结果表明,与其他典型调度方案相比,该算法的吞吐量和延迟性能均有所提高。
Data fusion in the Internet of Things environment requires the collection and processing of various data with mixed time characteristics,including real-time data and non-real-time data.The real-time delivery of data as a prerequisite for data fu⁃sion has an extremely important impact on the results of the fusion.The paper proposes a dynamic queue-aware distributed schedul⁃ing algorithm(QACSMA).In this algorithm,it takes into account the packet loss rate of data,delay constraints and queue prediction occupancy rate and uses a greedy algorithm to dynamically adjust the contention window to improve the service efficiency of long queues,reduce the queue length and time delay when the link queue length becomes longer.The simulation results show that com⁃pared with the typical scheduling schemes.Both throughput and latency performance of the algorithm are improved.
作者
王成杰
徐九韵
朱兰芳
李世宝
WANG Chengjie;XU Jiuyun;ZHU Lanfang;LI Shibao(College of Oceanography and Space Informatics,China University of Petroleum(East China),Qingdao 266580;College of Computer Science and Technology,China University of Petroleum(East China),Qingdao 266580;Dongying Science and Education Park,China University of Petroleum(East China),Dongying 257061)
出处
《计算机与数字工程》
2023年第1期181-185,共5页
Computer & Digital Engineering
基金
国家自然科学基金项目(编号:61972417)资助