Joint edge caching and dynamic service migration in SDN based mobile edge computing
Chunlin Li,Lei Zhu,Weigang Li,Youlong Luo
DOI: https://doi.org/10.1016/j.jnca.2020.102966
IF: 7.574
2021-03-01
Journal of Network and Computer Applications
Abstract:<p>In recent years, with the rapid growth in the demand for online streaming services, video streaming platforms are becoming more and more popular, and users' demand for low latency and high-quality services is increasing. Therefore, in order to be able to allocate caching resources reasonably to serve as many user requests as possible, and reduce latency and energy consumption, the edge cooperative caching method based on delay and energy consumption balance in SDN based mobile edge computing is proposed. In the proposed caching method, firstly, the multilayer perceptron neural network is used to predict the video content requested by the mobile user. Secondly, an objective function to minimize delay and energy consumption is established, and an edge cache optimization model is constructed. Finally, the branch-and-bound algorithm is used to obtain the optimal edge cache strategy. Meanwhile, to provide users with seamless service migration to ensure service continuity and high-quality services, the dynamic service migration method based on deep Q learning is proposed. In the proposed service migration method, firstly, the service migration problem is expressed as a Markov decision process. Secondly, the service migration process is analyzed and a service migration reward function is constructed. Finally, deep Q learning is used to obtain the optimal service migration strategy. In the experiment, the proposed edge caching algorithm can effectively improve the cache hit rate, reduce the backhaul traffic load, and control the average access delay and energy cost. Moreover, the proposed service migration algorithm can effectively reduce the number of service migrations and transmission cost, improve the success rate of service migration, and reduce the average traffic consumed by migration.</p>
computer science, interdisciplinary applications, software engineering, hardware & architecture