基于用户行为反馈的云资源调度机制  

User behavior-based resource scheduling mechanism for cloud computing with feedback control

作　　者：丁丁[1] 艾丽华[1] 罗四维[1] 徐保民[1] 

机构地区：[1]北京交通大学计算机与信息技术学院,北京100044

出　　处：《系统工程与电子技术》2018年第1期209-216,共8页Systems Engineering and Electronics

基　　金：国家自然科学基金(61300176;61473031;61472029;61672086);中央高等学校基本科研业务费(2016JBM019)资助课题

摘　　要：云计算中不断增长的资源种类、数量以及不断变化的用户需求给资源调度带来了极大的挑战,其中之一就是如何在复杂纷纭的云计算环境中为用户寻找真实所需的资源。针对这一问题,从用户的角度着手,基于用户行为对云资源调度问题进行深入的研究,提出了基于用户行为反馈的资源调度机制(user behavior-based resource scheduling mechanism with feedback control,UBRSM-FC)。UBRSM-FC利用用户交互行为,将用户行为信息融入到资源调度过程中,并结合相关反馈机制,建立用户需求的主动发现网络和反馈网络对调度过程不断微调和优化,使资源调度的结果与用户的主观感知更加接近。实验结果表明,UBRSM-FC通过反馈控制能够更好地满足不同用户不同程度的资源请求,提高了用户满意度,而且能够兼顾系统资源的利用率。The ever-increasing quantity and variability of both user requirements and resources poses many challenges to cloud resource scheduling.One challenge is how to identify the real needed resources for users with diverse requirements.In response to this issue,the problem of cloud resource scheduling is studied from the perspective of users is studied,which is based on user behaviors.And a user behavior-based resource scheduling mechanism with feedback control(UBRSM-FC)is further developed.UBRSM-FC integrates the user behaviors information collected from scheduling history into the resource scheduling via the relevance feedback mechanism.UBRSM-FC establishes an active discovery network and feedback network for user requirements to perform continuous feedback on the scheduling process which can make scheduling results closer and closer to the real purpose of users.Simulation results confirm that this user behavior-based scheduling mechanism is much more efficient in satisfying the diverse requirements of users,and also performs well on the resource utilization rate of the cloud system.

关 键 词：云计算 资源调度 用户行为 反馈控制 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]