Research Of Sensor Data Management Technology Based On WOT

The Internet of Things based on web (WoT) takes advantage of the openness and usability of web technology to formulates framework and interface standard. It is not only benefit for unified resource management and open but also benefit for rapid application development and access.The interaction between things and things, humans and things relies on intelligent data processing which requires effective data management. Data management includes data acquisition, data compression, data transmission, data storage, data query and so on. Because the sensor data in WoT has characteristics of heterogeneous, mass and uncertainty, designing a WoT system with efficient data management has strong realistic meanings.Based on the research of WoT business requirements, system architecture, middleware technology and sensor data characteristics, this paper designs sensor data management system, discusses the function and design principles of data management modules in WoT embedded gateway and WoT platform.This paper mainly studies gateway interfaces and sensor communication protocols, especially RS-232/RS-485interface and Modbus protocol, designs request-response driver model and direct-read driver model which are compatible with heterogeneous sensors to realize data acquisition, data conversion, data identification; mainly studies sensor data compression algorithm based on piecewise linear fitting, puts forward LRC-EC (Linear Regression Compression for Endpoint-Continuous) algorithm and compares LRC-EC with other algorithms, the result shows that LRC-EC can realizes great lossless compression and loss compression and is fit for business with no harsh real-time requirement; mainly studies database storage engines and cache replacement algorithms, theoretical researches and practical tests show that MyISAM engine is fit for short term historical data, Archive engine is fit for archiving historical data, memory cache based on FIFO-LRU is fit for high frequency access real-time sensing data, these technologies can realize optimum match between query speed and storage space.The major research achievements of this paper are high compatible device driver models, high compression algorithm, high matching storage scheme. This paper realizes these technologies, completes sensor data processing from data generation to data open.The development efficiency and system performance of WoT system is significantly improved by the research of WoT sensor data management technology, so that it can provides efficient services to meet the increasing needs from users and markets.