Lightweight Fog Attention Graph Convolutional Network for Secure Data Aggregation Mechanism to Fog based Cloud Assisted IoT Network

Ever Increasing demand of Cloud Computing paradigm has resulted in widespread development and deployment of multiple fog nodes to cloud assisted internet of things networks as it is highly capable of providing the users with on demand and low latency services through mobile collaborative devices in the edge of multiple clouds. Mobile Collaborative Devices is used for fog computing with integrated storage, computing and communication capabilities. However it offers improved efficiency and increased flexibility. Despite of multiple advantageous of fog based cloud assisted internet of things, security of the user data and their privacy can be compromised during data aggregation. In addition, it faces huge challenges in effectively aggregating the data and transmitting to the cloud server on establishing the strong security. In order to enhance the security of the user data and to preserve the privacy of the user information along establishing a secure data aggregation and transmission, a new light weight convolutional attention network is proposed in this article. It establishes secure data aggregation and data transmission to the distributed cloud data centers efficiently. Fog attention graph convolutional network leverages neighbour fog data efficiently and securely from dataset and represents in form of graph. Attention coefficient is assigned to each fog data represented in graph structure and attention score is calculated to each update of the nodes in graph. Specifically softmax function employs ID3 decision tree classifier and K anonymization mechanism to aggregates fog data on basis of updates of the attributes of the fog nodes in terms of the attention score and transmits aggregated data to the distributed cloud server. Further proposed model enhance the security of the data aggregation and transmission with index function through hash mechanism. Experimental analysis is carried out on the Fog assisted IoT medical (FIoMT) dataset. Dataset composed of patient health data collected using fog devices on various locations. Finally performance analysis of the secure data aggregation of fog IoT data using proposed architecture provides increased data integrity and increased prediction accuracy on basis of completeness, consistency, precision, recall and measures against state of art approaches.