Mass Data Middleware Within The Internet Of Things

With the qualitative change and development of Internet of Things(IoT)technology,IoT products have become more and more fashionable in daily life in recent years,and the demand in IoT is also growing fast,IoT technology has already spread into sports,security,education and many other fields,and has a wide variety of products.IoT technology collects the sensor's sensory data and then couples it with the cloud platform.For example,environmental monitoring uses temperature and humidity sensors,and shared bicycles use GPS sensors.However,the use of IoT products is increasing,and it must be accompanied by a large amount of perception.Data is generated,while traditional IoT technology cannot perform machine learning in a mass data environment.The cost of secondary development is high,and the functions of IoT products are relatively simple,and the deployment is relatively cumbersome.As a result,middleware technologies for mass data IoT have emerged.A general,lightweight IoT middleware platform for mass data is proposed in this paper.It successfully integrates the perception of intelligent hardware side of the IoT with the application layer.It conducts cloud computing on a large amount of perceived data.Hardware has given new missions,and the IoT middleware platform is subdivided into a versatile mass data server platform and a lightweight algorithmic server platform.Aiming at the large amount of perceptual data provided by the current IoT intelligent hardware terminal group,a general perceptual mass data server network model is proposed,which combines multiple load balancing algorithms to process mass data messages concurrently.In view of the single function of traditional IoT products,the platform also incorporates remote intelligent hardware device upgrade and control technology.Aiming at the situation that traditional IoT products can't be trained in mass data machine learning,a dynamically scalable cloud computing algorithm server architecture model is proposed.The algorithm server uses the Linux Container idea to realize cloud computing with a large amount of perceptual data training,and mass data training.The algorithm image supports second-level update registration,and the overall deployment is light and fast.Aiming at the high situation of the secondary development of traditional IoT products,an IoT middleware platform with end-cloud integration ideas is proposed.In terms of use,the user only needs to connect to the intelligent hardware device to transmit the protocol,and connect the data interface with the cloud,and combine the IoT terminal with the cloud computing cloud to create a situation of end cloud integration.In the area of IoT sports and security,two IoT products based on this middleware technology have been applied in 18 colleges and many places in Beijing,Zhejiang,Chongqing,Xi 'an,etc.It has more than 200000 users,and it has good effect,stable operation,strong flexibility and good feedback.