Research And Implementation Of Distributed Data Fusion Management System Based On IoT Platform

With the development of the IoT(Intemet of things),supported by technologies such as data fusion,information interaction and data sharing,intelligent IoT applications have been innovated.The terminals devices of these applications come from all walks of life,have created massive and complex data sets.However,the inconsistent information exchange standards applied by different types of bottom terminals,the data of various industries are difficult to aggregate and share,which hinders the development of applications in related industries.It is important to find,query,retrieval and access the data you want quickly and correctly.For example,to take multi-source sensor sensing data,the paper designs and implements a distributed data fusion management system.The major contributions of the dissertation include:1)We design a data fusion platform.The platform can be connected to a variety of monitoring terminals of different standards.The attribute name is redefined according to the different monitoring attribute names in the Internet of things platform according to their meanings.2)To meet the requirement of strong timeliness and consistency in data access of distributed IoT,it can be implemented in three levels.First,the system can be used to set the data slice,and the data alignment of neighbor memory database is completed.Second,we achieved all basic CRUD(create,retrieve,update,and delete)operations and refresh the library structure in real time and ensure strong consistency through multicast reply,that is,users can check the latest data at any time.Thirdly,after the data is updated,it broadcasts to the neighbor,each slice subscribe to the relevant modification and update automatically to ensure weak consistency,that is,the neighbor related slice will gradually reach the consistent state and tolerate partial data loss over time.3)We designed the distributed data fusion management architecture.In order to solve the problem of data storage location diversity and user dyslexia,the client agent and distributed cluster are designed.The client agent transferred user requests to the distributed cluster and it provides opaque support to upper layer business application.For R/W Splitting,master and slaver model,load balancing,ensuring data reliable,we realized distributed cluster.In order to solve the problem of large amount of real-time data,data sampling is designed,and the memory snapshot is used to store the historical data to the persistent database to compress the data amount.Finally,a central management architecture is provided for managers,who are responsible for overall management of the whole system,providing necessary configuration download,update and monitoring the normal operation of each module.Finally,the movement control performance of this system is stable.It can realize the accurate functions of data fusion management system in the cluster environment.