计算机应用 ›› 2014, Vol. 34 ›› Issue (10): 2806-2811.DOI: 10.11772/j.issn.1001-9081.2014.10.2806

• 先进计算 • 上一篇    下一篇

基于Hadoop的电力地理信息系统数据管理

林碧英1,2,王艳萍2   

  1. 1.
    2. 华北电力大学 控制与计算机工程学院,北京102206
  • 收稿日期:2014-04-18 修回日期:2014-06-11 出版日期:2014-10-01 发布日期:2014-10-30
  • 通讯作者: 王艳萍
  • 作者简介:林碧英(1955-),女,湖南株洲人,教授,主要研究方向:计算机网络、电力信息化;王艳萍(1989-),女,湖南益阳人,硕士研究生,主要研究方向:电力GIS、云计算。

Data management based on Hadoop for power geographic information system

LIN Biying1,2,WANG Yanping2   

  1. 1.
    2. School of Control and Computer Engineering, North China Electric Power University, Beijing 102206, China
  • Received:2014-04-18 Revised:2014-06-11 Online:2014-10-01 Published:2014-10-30
  • Contact: WANG Yanping

摘要:

针对传统电力地理信息系统(GIS)在存储能力、分析能力和扩展能力上的不足,将云计算技术应用到电力GIS领域,提出利用Hadoop云平台对电力GIS数据进行高效存储和管理的方案。首先对电力GIS各类数据的特点进行了分析,提出了关系型数据库与非关系型数据库相结合的数据存储策略,并在此基础上设计了基于Hadoop的电力GIS数据管理整体架构、相应的数据模型以及基于MapReduce的数据并行查询分析方法。最后,在单机和集群的环境下,对空间分析与运行数据查询的性能进行了对比与验证。实验结果表明,在数据量达到一定规模时,该方案优势明显,数据分析与查询的平均时间缩短30%以上,具有较高的效率和良好的扩展性。

Abstract:

In consideration of the problem that the traditional power Geographic Information System (GIS) has inabilities in storage, computing and scalability, cloud computing was applied into the power GIS field and a solution which used Hadoop cloud platform to store and manage the massive GIS data was proposed. After analyzing the characteristics of the power GIS data,a data storage strategy was proposed which combined relational database with non-relational database. Based on this strategy, the architecture of power GIS management based on Hadoop was presented, the data model and parallel data analysis based on MapReduce were designed. Finally, a quantity of experiments were carried out including spatial analysis and operation data queries in single-machine and cluster environment to compare and validate the performance. The experimental results show that the average time of data analysis and query declines over 30% after reaching certain amount of data. The proposed scheme has more obvious advantages to deal with large-scale data and has high efficiency and good feasibility.

中图分类号: