基于云计算的配电自动化集群Lzo无损压缩新方法  

Cluster Lzo Lossless Compression New Method for Distribution Network Automation

作　　者：屈志坚[1] 郭亮[1] 

机构地区：[1]华东交通大学电气学院,南昌330013

出　　处：《计算机测量与控制》2014年第4期1222-1224,1229,共4页Computer Measurement &Control

基　　金：国家自然科学基金项目(51267005);博士科研启动基金项目(09122014);江西省教育厅科技研究项目(GJJ13350)

摘　　要：针对配电自动化监控系统中大数据集信息和压缩存储率低的问题,提出一种基于Hadoop云计算的信息流集群无损压缩新方法;从监控业务信息流出发,将Lzo无损压缩编码融入Map/Reduce计算任务中,实现对大数据集监控信息流的有效压缩;以配电网监控信息流的处理为例,分别取300万和1000万断面记录进行测试,结果表明:每秒可压缩处理记录数从4万提高至8.4万条,处理效率提高了52.4%,同时压缩比也从55.4%提高至62%,适用于动态量测过程的信息处理,提高了压缩处理的效果。In order to solve low efficiency problem of large data sets compression storage in distribution network dispatching monitoring system, a new processing method based on cloud computing Hadoop is proposed for information flow lossless compression. The key business information flow is extracted from monitoring system. Then the Lzo lossless compression codes fused into Map/Reduce parallel computation task, which achieving effective compression for large data sets. Taking distribution network monitoring information flows as test example. Result shows that when section measurement record sets increased from 300 million to 1000 million, the compressed record per second is in creased from 4 million to 8.4 million, and the compression ratio from 55.4% to 62%. The processing efficiency is improved by 52.4%. New compression method is faster and more suitable for processing dynamic measurement and control process information, which improves the compression effect.

关 键 词：配电自动化系统 Hadoop云计算 集群压缩 Lzo无损压缩 并行分布式计算 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置] TM73[自动化与计算机技术—控制科学与工程]