SafeTECKS: Protect Data Safety in Things-Edge-Cloud Architecture with Knowledge Sharing

With the rapid development of Internet of Things (IoT) technology, massive data are generated by IoT terminals. Cloud computing is the current mainstream method to process these massive data effectively. However, the centralized cloud computing has the risk of data leakage during data transmission. To address the issue of data safety, we propose a novel edge computing architecture, called SafeTECKS. It consists of three-layer functional structure: IoT Devices, Edge Nodes, Cloud Center. Each edge node converts private data to knowledge by Agent component, and keeps them in Knowledge Store. Edge node shares knowledge with each other, instead of raw data, which can prevent data leakage caused by data transmission. An algorithm named MKF (Multi-Knowledge Fusion) is presented to integrate all knowledge learned from edge nodes. We use taxi demand prediction as a case to verify the effectiveness of our SafeTECKS on a real-world large scale data generated by taxis in Beijing. Results show that our method not only outperforms the baselines, but also can ensure data security.