摘要
为了分析系统中产生的海量数据,研究并分析了一个面向对象数据分析框架ROOT系统。首先分析了ROOT系统的核心模块和主要功能,然后阐述了ROOT系统的文件存储结构和输入输出系统。ROOT系统的文件存储采用压缩的二进制格式,大大减少了数据的存储空间和传输时间;针对数据分析领域中的局部性特点,提供了数据对象自动拆分机制,减少了数据分析时数据的吞吐量,显著提高了数据分析效率。最后通过实验验证了ROOT系统在海量数据分析领域的高效性。
To analyze the huge amounts of data generated in the system, an object-oriented data analysis framework ROOT is re- searched. Firstly, the ROOT system's core module and main function is analyzed, and then its file storage structure and I/O sys- tem is expounded in detail. ROOT system files are stored using a compressed binary format, and the data storage space and data transmission time is greatly reduced. Aimed at the locality of data analysis fields, the ROOT system provide the data object to be automatic resolution mechanism. This mechanism reduces the data throughput, improves the efficiency of data analysis greatly. Lastly, the experiments verify the ROOT system's high efficiency in mass data analysis field.
出处
《计算机工程与设计》
CSCD
北大核心
2012年第12期4594-4597,4602,共5页
Computer Engineering and Design
作者简介
张卫星(1987-),男,河南周口人,硕士研究生,研究方向为并行计算与并行软件。E-mail:wxzhang1987@gmail.com