基于并发跳表的云数据处理双层索引架构研究

发布时间：2018-01-14 22:06

  本文关键词：基于并发跳表的云数据处理双层索引架构研究　出处：《计算机研究与发展》2015年07期 　论文类型：期刊论文

  更多相关文章： 云计算 双层索引 并发跳表 范围查询 乐观并发控制

【摘要】：云数据处理在云计算基础设施中占有极其关键的地位.然而,当前的云存储系统绝大部分都采用基于分布式Hash的健-值对模式来组织数据,在范围查询方面支持不理想、且动态实时性差,有必要构建云环境下辅助动态索引.通过总结、分析云环境中辅助双层索引机制,提出一种基于并发跳表的云数据处理双层索引架构.该架构采用两层体系结构,突破单台机器内存和硬盘的限制,从而扩展系统整体的索引范围.通过动态分裂算法解决局部服务器中的热点问题,保证索引结构整体的负载均衡.通过并发跳表来提高全局索引的承载性能,改善了全局索引的并发性,提高整体索引的吞吐率.实验结果表明,基于并发跳表的云数据处理双层索引架构能够有效支持单键查询和范围查询,具有较强的可扩展性和并发性,是一种高效的云存储辅助索引.
[Abstract]:Cloud data processing plays an extremely critical role in cloud computing infrastructure. However, most cloud storage systems use a distributed Hash based key-value pair pattern to organize data. It is necessary to construct the auxiliary dynamic index in the cloud environment because of its poor support for range query and poor dynamic real-time. Through summing up the mechanism of auxiliary double-layer index in the cloud environment is analyzed. This paper presents a two-layer index architecture for cloud data processing based on concurrent hopping table, which adopts a two-tier architecture to break through the restrictions of single machine memory and hard disk. In order to extend the whole index range of the system, the dynamic splitting algorithm is used to solve the hot spot problem in the local server, to ensure the overall load balance of the index structure, and to improve the load bearing performance of the global index through concurrent table hopping. It improves the concurrency of the global index and the throughput of the global index. Experimental results show that the cloud data processing two-layer index architecture based on concurrent hopping table can effectively support single-key query and range query. With strong scalability and concurrency, it is an efficient cloud storage aided index.
【作者单位】： 云南大学软件学院;
【基金】：国家自然科学基金项目(61363021,61363084) 云南省软件工程重点实验室开放基金项目(2011SE01,2012SE304) 云南省青年基金项目(2012FD004) 云南省教育厅科学研究基金项目(2014Y013)
【分类号】：TP333
【正文快照】： 由于能够提供海量存储、可靠服务,近年来,云计算系统日益受到重视.现有的云计算基础设施主要包括亚马逊的EC2(amazon，

本文编号：1425484