Research And Implementation Of The Auxiliary Index Technology For The Real-Time Traffic Stream Data Based On HBase

The traffic data show exponential growth which with the rapid development of urban intelligent traffic and the prevalence of taxi software and shared bicycles.A reasonable analysis of traffic data can predict people's travel habits and popular traffic routes,to help the management of urban transport,but also for the relevant Internet companies to provide revenue to provide support.The current the HBase database which has the inherent advantages of large-scale data storage who application in the field of transportation has become very extensive.However,HBase can only build indexes on row keys and can not directly construct efficient multidimensional indexes and non-key indexes so that it is poorly on spatio-temporal queries of traffic data and can not achieve fast non-key query.In view of these problems,it is necessary to construct the corresponding auxiliary index for HBase.Based on the characteristics of real-time traffic stream data,this paper studies and designs two kinds of auxiliary indexing technologies,such as HBase spatio-temporal index and secondary index,and implements a query system based on HBase auxiliary index.The main research contents are as follows:First,we designed three spatio-temporal indexes based on the temporal and spatial characteristics of traffic data which improved the spatial and temporal query of traffic data.We are analyzed their advantages,disadvantages,applicable scenarios and so on.Second,we are designed two secondary indexes based on the inverted index which solve the problem of low performance of HBase non-rowkey's query and improve the traffic data query's flexibility of HBase.We are analyzed their related applicable scene.Third,we were implemente a query system based on HBase auxiliary index.The system implements the optimal scheme of the two auxiliary indexes designed in this paper.It also realizes the functions of SQL analysis,index building,index matching and index management,and simplifies the query operation of HBase database.Finally,the experimental results show that the proposed index technology can effectively improve the query performance of traffic stream data,and also verify that the implementation of the query system has a very good index management capabilities.