Design And Implementation Of IoT Data Streaming Computing Platform Based On Serverless Architecture

The wide application of the Internet of Things has effectively promoted the development of the industry.However,most of the functions provided by the Internet of Things platforms are still mainly based on device data management and connection management.Due to the increasing number and types of connected devices,the amount and types of data generated are also increasing,and these platforms generally lack analysis and computing services for device data.In addition,the provision of computing resources requires the support of cloud computing technology.However,the traditional cloud computing model has high usage costs and complex application deployment processes.Developers need to devote most of their energy to operation and maintenance,which reduces development efficiency.Therefore,this paper designs an IoT data streaming computing platform based on a serverless architecture,which extends the functions of the existing IoT platform,provides users with IoT data analysis and computing services,reduces the use cost of the platform,and improves development efficiency.Because IoT data processing has the characteristics of high timing,structure,and traffic stability,the platform uses the Flink streaming computing engine to manage data with low end-to-end latency and high throughput.At the same time,in order to efficiently use computing resources,we choose to use serverless architecture to build the platform,and use the Knative framework to simplify the construction and deployment process of big data applications,improve development efficiency,realize automatic scaling of applications,and reduce operation and maintenance costs.The platform provides users with automatic deployment,operation and elastic scaling of task codes,as well as corresponding resource management functions.Secondly,in order to solve the cold start problem of Knative,the Mixer component is used to improve the original architecture to speed up the cold start of instances;Platform functionality was tested to verify the usability of the platform.