Research On Layered Secure Data Storage And Application Model For Edge Computing

Billions of data are generated every day on the Internet.With cloud computing,enterprises can store and leverage the huge data in the cloud to serve their users.With the development of Internet of Things(IoT)in the past few years,more and more IoT devices have access to the Internet and generate massive amounts of data,these data rely on cloud computing for centralized processing and analysis.The ever-increasing amount of data has exceeded the performance of cloud computing,which causes serious computing delays.Edge computing allows the data owner to collect and process data near the network edge by edge nodes and supports data pre-processing to overcome the low efficiency and other issues in cloud computing with IoT data processing.However,the data security issues in edge computing involve different storage and communication methods due to its heterogeneous network.Enterprises have not paid enough attention to edge computing data security issues.It has become the most important issue for enterprises and government departments to protect the security of data in edge computing.This thesis focuses on data security issues in cloud computing and edge computing.It comprehensively analyzes the shortcomings of existing data security models and systems.The thesis devotes to the study of a secure data storage and application model for edge computing.The main work of this thesis is summarized as follows:(1)A Secure and Highly Available Multi-Cloud Database(SHAMC)based on the hybrid cloud was proposed.Based on the homomorphic encryption algorithm and secret sharing mechanism,the database can directly store encrypted data and execute queries on the ciphertext.The hybrid cloud structure improves the availability and the data integrity of the encrypted database,avoids the service interruption and vendor lock-in problem.The evaluation proved the high availability and query efficiency of SHAMC than other existing encrypted databases.(2)A layered data storage and processing model for edge computing was further proposed.The model was applied to a distributed camera edge network to realize an Edge-Assisted Realtime Trajectory Prediction System(EaTP).The person re-identification and trajectory prediction algorithms were used in this system to support trajectory tracking and trajectory prediction in complex scenarios.The experiment proved the feasibility and scalability of the model.(3)A layered data security integration model for edge computing was proposed.In this model,the edge computing network was classified into the cloud layer,the edge layer and the application layer due to the characteristics of heterogeneous networks.The secret sharing and homomorphic encryption algorithm were adopted to make all the data transmission performed on the ciphertext.Most of the computational load was transferred to the cloud layer,and the ciphertext data transmitted from the edge node can be directly calculated and analyzed in the cloud.The malicious attacker cannot snoop the user's sensitive data due to the encrypted data model.The application model was also designed to help service providers integrate their own resources and provide a complete edge computing solution for different user groups(data owners).The experiment evaluated its encryption/decryption performance and system availability.