一种融合服务满意度的多因素感知云服务性能预测策略  

Performance prediction policy combining with service satisfaction and multi-factored awareness for cloud services

作　　者：胡丹丹[1] 陈宁江[1] 朱莉蓉[1] 李湘[1] 

机构地区：[1]广西大学计算机与电子信息学院,南宁530004

出　　处：《计算机应用研究》2014年第12期3663-3667,共5页Application Research of Computers

基　　金：国家自然科学基金资助项目(61063012;61363003);广西自然科学基金项目(2012GXNSFAA053222);广西高校优秀人才资助计划([2011]40);广西科学研究与技术开发计划项目(桂科攻1348020-7)

摘　　要：提出了一种多因素感知的云服务性能预测策略,它基于综合权衡服务提供商的利益,通过合理定价和性能保障来量化和优化服务满意度,并以之驱动性能预测,并引入效用理论实现服务满意度的合理定价。最后通过模拟实验表明,它能够降低SLA破坏率,有效提高云服务的资源利用率。This paper proposed a multi-factored awareness cloud service performance prediction strategy. The strategy took the interest of service providers into consideration,and service satisfaction was quantified and optimized in two ways-reasonable pricing and performance guarantee. It used service satisfaction for performance prediction. The strategy adopted the utility theory to decide reasonable price. The experimental results show that it not only can reduce service level agreement violations of service performance,but also improves the resource utilization effectively.

关 键 词：服务满意度 云服务 性能预测 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]