论文部分内容阅读
数据网格提供了地理分布的大量共享数据资源,然而互联网的高访问延迟降低了数据访问的效率,创建副本是一个有效的方法,而在何处放置副本是一个具有挑战性的问题。从应用环境和用户访问特征出发,讨论了副本创建位置、粒度和时机,提出了一种基于副本共享组的副本创建模型,可以获得优化的副本创建位置,构造了一个副本创建代价函数,由系统的传输速率、拓扑结构和用户访问特征共同决定,并给出了确定副本创建位置的有效算法。分析和仿真实验表明,自适应副本管理方案具有动态性、适应性和可扩展性,能较好地适应数据网格的特性,可以有效降低访问延迟,提高数据访问效率。
Data grids provide a large amount of shared data resources that are geographically distributed. However, the high access latency of the Internet reduces the efficiency of data access. Creating replicas is an effective method, and placing replicas there is a challenging issue. Based on the application environment and the characteristics of user access, the location, granularity and timing of replica creation are discussed. A replica creation model based on replica sharing group is proposed. An optimized replica creation location can be obtained and a copy creation cost function is constructed. The transmission rate, topology and user access characteristics are determined jointly, and an efficient algorithm for determining the location where the replica is created is given. The analysis and simulation results show that the adaptive replica management scheme is dynamic, adaptable and scalable and can better adapt to the characteristics of data grids. It can effectively reduce the access delay and improve the data access efficiency.