Smart Grid and Innovative Frontiers in Telecommunications. Third International Conference, SmartGIFT 2018, Auckland, New Zealand, April 23-24, 2018, Proceedings

Research Article

Real-Time CPU Scheduling Approach for Mobile Edge Computing System

  • @INPROCEEDINGS{10.1007/978-3-319-94965-9_4,
        author={Xiaoyi Yu and Ke Wang and Wenliang Lin and Zhongliang Deng},
        title={Real-Time CPU Scheduling Approach for Mobile Edge Computing System},
        proceedings={Smart Grid and Innovative Frontiers in Telecommunications. Third International Conference, SmartGIFT 2018, Auckland, New Zealand, April 23-24, 2018, Proceedings},
        proceedings_a={SMARTGIFT},
        year={2018},
        month={7},
        keywords={SVM EDF Low latency High reliability of transmission},
        doi={10.1007/978-3-319-94965-9_4}
    }
    
  • Xiaoyi Yu
    Ke Wang
    Wenliang Lin
    Zhongliang Deng
    Year: 2018
    Real-Time CPU Scheduling Approach for Mobile Edge Computing System
    SMARTGIFT
    Springer
    DOI: 10.1007/978-3-319-94965-9_4
Xiaoyi Yu1, Ke Wang1,*, Wenliang Lin1, Zhongliang Deng1
  • 1: Beijing University of Posts and Telecommunications
*Contact email: wangke@bupt.edu.cn

Abstract

Mobile edge computing (MEC) system has outstanding advantages of providing smart city applications with relatively low latency and immediately response. How to guarantee the QoS of the services in MEC system is consequently becoming a hot issue. This work focuses on solving the problem by real-time CPU scheduling. The proposed scheduling algorithm considers different services arrival profiles, computation time consumption and deadline requirements simultaneously. Specifically, the combination and optimization of support vector machine (SVM) and earliest deadline first (EDF) algorithm is designed, which could automatically classify services type and efficiently allocate the computation time in real-time manner. By deploying the traffic trace from the real world, the proposed scheduling algorithm could reduce latency and improve the reliability of transmission, comparing with popular fixed-priority CPU scheduling algorithm.