实时历史数据库中压缩技术的并行化研究  

Research on Parallel Compression Technology in Real-Time Historical Database

作　　者：祝君[1,2] 林庆农[2] 徐造林[1] 

机构地区：[1]东南大学计算机学院,江苏南京210096 [2]南京南瑞继保电气有限公司,江苏南京211100

出　　处：《计算机技术与发展》2010年第7期36-39,共4页Computer Technology and Development

摘　　要：实时历史数据库系统需要极高的数据压缩速度,但传统的压缩算法都使用串行处理,数据压缩和文件读写都要耗费大量的时间。为了提高数据压缩速度,提出一种并行压缩算法,首先把待处理数据分成若干小块,然后通过对块数据压缩、解压缩、文件读写并发处理,充分发挥多核处理器的高性能。并设计出一种适合并行压缩算法的压缩数据存储文件格式和一个数据段信息循环链表,通过多线程编程技术来实现并行压缩算法。使用工程实测数据在多核处理器系统上的实验表明,并行压缩算法可以极大地提高数据压缩速度,满足实时历史数据库的需求。Real-time historical database system needs a high data compression speed.The traditional compression algorithms with serial processing technique require large amount of time to compress data and read or write file.In order to improve data compression speed,a parallel compression algorithm is presented.The high performance of multi-core processors can be fully exploited through the concurrent processing of data compression,data decompression,reading and writing of data files.A file format and a circular linked list for data segment information are designed to fit the proposed parallel compression algorithm.Multi-thread programming techniques are employed to achieve the parallel compression algorithm.Experiment is conducted on a real engineering data with a multi-core processor system.The results show that data compression speed can be greatly improved with the parallel compression algorithm and the requirement of real-time historical database can be met.

关 键 词：实时历史数据库 数据压缩 并行编程 

分 类 号：TP31[自动化与计算机技术—计算机软件与理论]