Research On NoSQL Database For Trajectory Big Data Storage And Query

In recent years,with the development of GPS(Global Positioning System)and the popularity of the Internet,the acquisition of location data is becoming more and more diversified and the amount of trajectory data is increasing.More and more location-based services platform provides a wealth of services for us.These online service platforms require higher data access efficiency and the traditional relational geographic database encounters bottlenecks in the management of massive trajectory data.Based on NoSQL database,this paper proposes a scheme to manage large scale trajectory data and establishes the spatial data index of track data to achieve efficient spatio-temporal query.This paper from the following three aspects researches on the trajectory data storage scheme based on NoSQL database.(1)In the aspect of database storage architecture,Take advantage of the high data writing efficiency of Redis and the high data reading efficiency of LevelDB,Combining redis and LevelDB to manage trajectory data.First,the real-time data is stored on Redis,and then transfer the data from memory to LevelDB database.This storage scheme can not only improve the data access efficiency,but also reduce the cost of data management by storing data on disk rather than in memory.(2)Using the sorted set to store the trajectory data.The time is divided into several time intervals according to the time gap of an hour,The key is composed of object identifier(ship number)and the time interval.The sets of ship dynamic information is taken as value.The dynamic information is represented by the coordinate pair,the ground navigation speed,the course,the rotation rate and so on.The Unix timestamp is used as the scores and the trajectory data is sorted according to scores.(3)In the aspect of the optimization of spatial index,the spatial grid index is created.The design of grid index also consider the spatial and temporal characteristics of trajectory data.Dividing the latitude and longitude grid with 0.25 degrees interval.The lower left corner of the grid where the latitude and longitude as grid number.Divide the time according to the interval of an hour,the key is composed of the grid number and the merged time interval.The values is corresponding to the ships number that appear in the grid within this time period.The scores are Unix timestamp.In the spatio-temporal query,the index set is searched to get the corresponding ship number set and time information according to the grid number which intersects the query window and the temporal information in the query condition.The results of the query are serialized according to the score values which is used to form trajectory.By comparing the data storage and query efficiency with the traditional object relational database Geodatabase,we can conclude that the proposed data storage architecture can effectively improve the efficiency of data access.The Sorted Set storage model of trajectory data can effectively reduce data redundancy and the space required for data storage.Index containing spatio-temporal information can improve the efficiency of large scale trajectory data management and spatio-temporal query.