Real-Time CPU Scheduling Approach for Mobile Edge Computing System.

Xiaoyi Yu,Ke Wang,Wenliang Lin,Zhongliang Deng
DOI: https://doi.org/10.1007/978-3-319-94965-9_4
2018-01-01
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 45% latency and improve the reliability of transmission, comparing with popular fixed-priority CPU scheduling algorithm.
What problem does this paper attempt to address?