Research And Implementation Of Cloud Computing-Based General Platform For Processing Data Of Internet Of Things

With the continuous development of Internet of things,and the more industries combine with it,more and more universities and corporate researchers begin to pay attention to the practical application of IoT technology.But that how to receive,store,and effectively manage the big data generated by the sensor nodes have become one of the major bottlenecks of the research and application of Io T.The process of big data generated by sensor nodes has attracted the attention of researcher long time ago.To improve the back-end server capacity of storage and processing data,the traditional approach is to build a distributed server cluster,and for specific scenarios and different needs researcher must develop different applications.However,the traditional server cluster is not only expensive,but also forces developers repeatedly develop the common parts such as data reception,storage,and display no matter how simple the need is.This leads to the waste of resources.In view of the above problems,this paper takes advantage of cloud computing technology as the basic framework of platform underlying data storage and processing,designs and implements a universal IoT data management platform.The platform not only supports multiple data encoding formats,but also provides configuration features for data frame format.At the same time for the common link of data analysis,the platform provides the application of state monitoring module,the data frame monitoring and alarm module,data processing management module.Additionally the platform also provides TCP server and Hadoop cluster monitor based on web for remote management.Taking into account the special needs of some applications,the code architecture and library functions of core parts use modular and object-oriented design principles to package.Because of open source license,secondary development is easy and simple.The other researchers can freely improve the platform,while avoiding wasteful duplication of development.This paper mainly involves the following work:1.Based on Hadoop,build a distributed cluster,and a big data processing warehouse Hive.Use LAMP bundle to establish a HTTP Web server,and install Swoole extension.2.The design and implementation of the TCP data reception server.Based Swoole,achieve a high-performance server to receive,cache,and analyze data frames,and the intermediate layer which interacts with the other part of platform.3.The design and implementation of data management mechanism.This paper abstracts the common parts of data process and develop the module of frame configuration,monitoring,and management based on the web platform.4.In order to research the performance of this platform,and the main feature of Hadoop cluster,this paper designs some experiments,and also provides real examples of platform to show platform's main user interface.