Decentralized and Secure Cooperative Edge Node Grouping to Process IoT Applications in Heterogeneous Smart Cyber-Physical Systems

Smart Cyber-Physical Systems (SCPS) has enhanced use of smart devices with numerous applications including smart cities, smart traffic management, smart cars, smart health care and smart grids. Core logic behind these applications usually require huge processing or massive data handling which are normally performed at cloud but the application will suffer from latency. Edge computing offers solution for latency aware computations at edge of the network but with limited resources available at edge nodes. This problem can be resolved by leveraging edge resources in a group and executing resource intensive task. Most previous studies deploy centralized methods for clustering but comes with overhead of cluster formation and management. In this article, key idea is to group heterogeneous edge nodes on task arrival in a decentralize way, and handle task allocation and execution in parallel on group devices to achieve its deadline. Our methodology will help to reduce traffic amount travelling towards cloud in case of resource intensive big tasks of SCPS applications. We have proposed an algorithm for decentralized group formation and presented task division and allocation methodology for parallel execution. Our results show that our technique is working while providing desired goals of reducing overall latency, and limiting network traffic as well as achieving higher ratio for number of tasks meeting their deadline.