摘要
随着数据量和数据类型的逐渐增加,用户对数据存储性能的要求和整体需求越来越高。当前数据分布算法因为受到成本及可扩展性的约束,无法达到用户对存储的要求。为此,提出一种新的行列混合存储数据库的数据分布自适应优化算法,给出数据分片、数据访问频率矩阵、拓扑结构和最小支撑树的定义,将总成本最小作为依据,判断数据是否冗余分布,获取行列混合存储数据库的利益函数。找到一个使利益最大化的数据分布方案,通过迭代求取该问题的解。输入相关参数,输出的结果即为数据分布结果。实验结果表明,所提算法读写能力和均衡性均较强。
With the gradual increase in the amount of data and data types,user requirements for data storage performance and overall demand more and more high,the current data distribution algorithm because of the cost and scalability constraints,can not meet the user requirements for storage. For this,a new adaptive distribution optimization algorithm ranks mixed storage database were put forward,define data slice and data access frequency matrix,topology and the minimum spanning tree,the minimum total cost as the basis for judging whether the data redundancy distribution,access list of interest function mixed storage database. To find a data distribution scheme that maximizes the benefits,the solution of the problem is obtained by iteration. Enter the relevant parameters,the output is the result of the data distribution. The experimental results show that the proposed algorithm has strong ability of reading and writing.
出处
《科学技术与工程》
北大核心
2017年第27期232-237,共6页
Science Technology and Engineering
基金
河南省基础前沿项目(152300410006
162102210046)资助
关键词
行列
存储数据库
数据分布
自适应
优化
ranks
storage database
data distribution
adaptive
optimization
