Abstract:
This paper presents a method to partition the resources for each service class with its resource demand and service level agreement requirement, then models different service requests’ states of each class. Feedback control theory, together with resource tuning policy is used to regulate resources of each class in order to adapt to paroxysmal requests. This mechanism guarantees QoS of different classes and achieves self-management in Web clusters. Experimental results indicate that the self-managing model could support differentiated QoS effectively and control resources’ utility.
Key words:
Quality of Service (QoS),
self-management,
control theory
摘要: 基于服务类别的资源请求和服务等级给出一种资源划分策略,对各类别的服务请求状态进行建模,使用反馈控制理论的方法对各类别资源进行动态调整,与资源调整规则相结合,以适应访问的突发性和保证服务质量要求,实现了Web机群服务质量的自管理。实验表明自管理模型能有效支持区分服务质量和资源效用控制。
关键词:
服务质量,
自管理,
控制理论
CLC Number:
LIU Tao; ZENG Guo-sun;. Self-management Model for Control Differentiated QoS in Web Clusters[J]. Computer Engineering, 2007, 33(24): 32-34.
刘 涛;曾国荪;. Web机群服务质量控制自管理模型[J]. 计算机工程, 2007, 33(24): 32-34.