Spatio-Temporal Data Distributed Storage Research

Spatio-Temporal data is a multi-dimensional data. Its structure is very complex, with spatial and temporal characteristics. It can record in detail the state of things in space and temporal and spatial variation, and can display the correct object in the past, present and future of the state. In the era of rapid technological development, the type of data collection devices more and more, the amount of data also increases rapidly, leading to difficulties in data storage management. The spatio-temporal data storage management module design and implementation of the merits of the decision of the working capacity of the entire data management system. Therefore, the application would affect the efficiency of the system to other upper. With the proposed distributed framework, its efficient parallel computing power, storage capacity, high scalability, high stability, etc. attracts us. based on previous studies of spatial and temporal data distributed storage research explored.This paper begins with a distributed spatio-temporal data and theory to proceed, research techniques and principles of the relevant parts of the proposed space-time index based on R-tree, and then the open source cloud platform Hadoop, HBase database support, and efficient use of the computing power of the Map Reduce spatio-temporal data management, and finally through some experiments indexing performance. The main contents are as follows:1) In-depth analysis of the advantages and disadvantages of the classical space-time temporal data model and indexing; a brief analysis of the characteristics of a distributed platform and related technologies to provide theoretical and technical support for the thesis.2) Systematic analysis of the core components of Map Reduce parallel computing framework Hadoop, an open source cloud platform, HDFS distributed file storage system to HDFS as the carrier of column-key database HBase data model. For large spatial and temporal characteristics of the data, proposed the use HBase big table to store spatio-temporal data management. Binding properties of space-time data with HBase, and elaborated on the design process and how to build the table row keys, define a column family.3) According to the current appearance of temporal data indexing, we propose a spatio-temporal data to build an index on the basis of the R-tree, the index data of the past and present time are stored in their tree manages the start and end, respectively, time, improve the utilization of trees to improve query efficiency. Finally, a comparative experiment to test the efficiency of this paper insert and query temporal index.4) Finally, the experimental data generated by the GPS simulator, then stored managed in HBase.