基于Hadoop的云计算在清分系统中的应用

摘要

针对轨道交通行业客流量逐年增大而带来的大数据和在清分系统中采用大中型计算机和关系型数据库导致成本高与容错低的问题，本文首次提出了采用Hadoop云计算解决该问题的一个技术途径，包括系统设计与实现以及测试结果等。实践表明，于Hadoop的云计算完全适用于轨道交通售检票清分系统的处理实时数据业务和非实时数据业务，具有成本低，容错好，运行稳定和效率高的优点，硬件投资仅占单台服务器的十分之一，其扩展性与容错性均优于单台服务器。%According to the problems of big data caused by yearly increased rail transit passenger flow and cost as wel as fault tolerance of using large and medium scale computers and RDB in the ACC system,this paper presents a technical way,including system design and implement as wel as testing results and so on,to solve the problems based on Hadoop cloud computing technology at first.Practice indicates that cloud computing based on Hadoop is totaly suitable for real time and non-real time data processing services in the ACC system of rail transit Automatic Fare Colection system(AFC)with advantages of lower cost,better fault tolerant capability,stable operation as wel as higher efficiency.Covered hardware investment is only a tenth of single server,but its expansibility and fault tolerance are both superior to single server.