基于K-means的动态资源分配策略

Dynamic resource allocation strategy based on K-means

为提高移动边缘计算(MEC)环境下大型云计算中心资源的利用率和用户体验,提出基于K-means算法的动态计算资源和频谱资源分配算法(KDSAA)。分析传统的资源平均分配方式和虚拟机分配方式现状,研究用户的综合需求,将资源模拟成“流体”,采用拍卖算法进行分配,线性求解出边缘云吞吐量和传输时延的最优值。实验结果表明,该算法可有效提高边缘云吞吐量和降低传输时延。

To improve the utilization and user experience of large cloud computing center resources in the mobile edge computing(MEC)environment,a dynamic computing resource and spectrum resource allocation algorithm(KDSAA)based on the K-means algorithm was proposed.The current status of the traditional average resource allocation method and virtual machine allocation method was analyzed,the user’s comprehensive needs were studied,resources were simulated as fluid,an auction algorithm was used to allocate,and the optimal values of edge cloud throughput and transmission delay were linearly solved.Experimental results show that the algorithm can effectively improve edge cloud throughput and reduce transmission delay.