Research Of Industrial Internet Of Things Data Storage Strategy Based On HBase

With the progress of the Internet of things technology, lots of sensors and intelligent terminals are applied to traditional industrial area, it is bound to bring the huge growth of data. How to manage these huge amounts of data effectively to improve the efficiency of industrial production is becoming an urgent problem to be solved. In view of the large industrial real-time data storage and processing, traditional solution is to adopt the centralized storage systems, however, in the face of industrial networking data, centralized storage system has limited capacity and scalability of the shortcoming, in the system flexibility and fault tolerance, traditional distributed database is difficult to meet the demand for data storage of industry networking data.After a lot of research and industrial real-time data storage system approach, this paper presents an approach to storage management against real-time data for a massive industrial. Based on the characteristics of industrial data, real-time historian commonly used in industrial systems approach to data processing is chosen as adaptation layer, it can filter, monitor and cache industrial collected data, at the same time to guarantee the data write smoothly to HBase. The open source system HBase is chosen as a data persistence layer to store eventually archiving historical data of industrial systems, The scalability,fault tolerance, data consistency, high support concurrent query performance of HBase can ensure the efficient management of industrial data. Finally the approach of offline compressing after data has stored is raised,it can make industrial system to compress more flexible and dynamic controllable.Finally, the rowkey of HBase is designed according to the requirements of storage, simultaneously simulating database on child nodes and query processor instance to test all aspects of the strategy. The experimental results prove that the massive data storage strategy has excellent archives performance, good expansibility and fault tolerance. Performance of parallel query strategy enhances as the number of query processors increase, but the trend will gradually be slow, continuous time range queries have excellent performance. At the same time, to select a particular day of the compression processing performed, it can be seen, the compressed data is drastically reduced, but the change trend of the data is preserved.