【作者】 何伟； 李庆忠； 郑永清； 崔立真；

【Author】 He Wei,Li Qingzhong,Zheng Yongqing,and Cui Lizhen (School of Computer Science and Technology,Shandong University,Jinan 250010)

【机构】 山东大学计算机科学与技术学院；

【摘要】 业务数据搜索是商业社区云面向云用户或第三方应用提供的基本服务,在动态、开放的社区云环境中,业务数据更为复杂,商业目标对数据搜索服务也提出了更高的要求.以面向旅游的社区云为背景,讨论一种透明、虚拟化的商业数据搜索服务的实现方法,基本思想是在构成社区云的服务节点中引入相似节点簇和等价节点簇两种虚拟节点簇模型,并分别给出其更新算法,使数据在云中的服务节点之间动态、自主地调整以趋于合理分布;然后,提出一种基于上述模型的3-阶段数据搜索策略.模拟实验表明,以此方法为核心的数据搜索服务,在数据搜索质量、效率和服务节点负载均衡等方面能够获得预期效果,满足社区云环境对数据搜索服务提出的商业目标.更多还原

【Abstract】 Community cloud is one of the deployment models of cloud computing,in which the infrastructure is shared by several organizations and a specific community that has common concerns is supported.In business-oriented community clouds,providing transparent and virtualized data search services for users or third-party applications is one of the basic services,and the business objectives in community clouds with characteristics of dynamic and complicated data bring forward higher requirements to data search services.Based on the environments of e-travel community clouds and treating Web services deployed on the cloud peers of the providers as business data,this paper discusses a strategy of virtualization for data search services considering multiple aspects such as search quality,efficiency and load balance of service peers.The basic idea is to introduce 2 virtual peer clusters for the service peers constructing community clouds,which are called similar peer cluster and equivalent peer cluster,and their updating algorithms make the allocation of business data be more efficient and proportional.Then,we propose a 3-phase search strategy for business data based on the model of virtual peer clusters.Simulation experiments show that our strategy can achieve anticipated results in search quality,efficiency and load balance among cloud peers,and so it can support the business goals of data search in community clouds.更多还原