一种QoS收益感知的服务请求动态调度算法

发布时间：2018-11-26 07:20

【摘要】：随着云计算产业的发展,云计算逐步走向实际应用,许多传统企业正在将其应用向云服务模式迁移.云服务模式中能够有效减小应用负载,降低服务成本的方法就是多租户模式.在云环境下多租户模式中如何合理安排各个租户的服务请求,通过动态调度租户服务请求来保障各租户的服务质量(Quality of Service,QoS)成为了亟待解决的关键问题.服务请求动态调度是为了保障租户的服务质量,提高服务提供商的资源利用率,因此,需要制定合理的调度方案来达到双方满意的效果.该文从服务请求的服务类型以及租户两方面入手,制定了多租户服务请求动态调度策略,提出了一种以动态QoS收益值为优化目标的服务请求调度算法,该算法能够根据不同租户对服务各种QoS指标的性能需求,对各种类型的多租户服务请求按照动态QoS收益值划分优先级,按照优先级进行调度,与一些常用的方式进行调度相比,能够让平台处理更多重要性较高的服务请求,从而使平台整体的动态QoS收益值更高.
[Abstract]:With the development of cloud computing industry, cloud computing is gradually moving towards practical application, and many traditional enterprises are migrating their applications to cloud service model. The method of reducing application load and service cost in cloud service mode is multi-tenant mode. How to arrange the service request of each tenant reasonably in the multi-tenant mode under the cloud environment and how to ensure the quality of service (Quality of Service,QoS) of each tenant through the dynamic scheduling of the tenant service request have become the key problem to be solved urgently. The dynamic scheduling of service requests is to ensure the quality of service of the tenants and improve the resource utilization of service providers. Therefore, it is necessary to formulate a reasonable scheduling scheme to achieve satisfactory results. In this paper, the dynamic scheduling policy of multi-tenant service request is formulated from the aspects of service type and tenant, and a service request scheduling algorithm based on dynamic QoS revenue is proposed. The algorithm can prioritize and schedule various types of multi-tenant service requests according to the dynamic QoS revenue value according to the performance requirements of different tenants for various QoS indicators, compared with some common scheduling methods. It allows the platform to handle more important service requests, thus making the dynamic QoS revenue of the platform higher.
【作者单位】：哈尔滨工业大学(威海)计算机科学与技术学院;
【基金】：国家科技支撑计划项目(2014BAF07B02)资助山东省重大科技专项项目(2015ZDXX0201B02)资助山东省自然科学基金项目(ZR2015FM006)资助
【分类号】：TP301.6

【相似文献】