Design And Implementation Of An Authorization Mechanism Based On Edge Computing Platform

Access control is a basic technology to secure IoT systems,and authorization is a key part of it.To reduce the computational pressure of encryption and decryption of IoT end devices and simplify the configuration and management of access control policies,mainstream research and industrial solutions have turned to the authorization method of directly verifying device attribute information to cope with the uneven computational resources of IoT devices and heterogeneous data sources.However,such authorisation mechanisms are often highly dependent on cloud service centres and lack protection for user data privacy,which can easily lead to privacy leakage problems such as illegal collection and inference of end-user attribute information.To address the above issues,this paper designs an authorisation mechanism that deploys homomorphic encryption services to prevent enduser privacy preference attribute information from being directly exposed to the cloud service centre by introducing an edge computing platform.At the same time,this paper further assesses the privacy leakage risk of enduser non-privacy preference attribute information in conjunction with user data disclosure propensity characteristics to avoid inference of user privacy data.The assessment of end-user attribute information privacy leakage risk includes 2 core steps:firstly,this paper automates the identification and classification of attribute information in end-user requests through the BERT-CRF deep learning model to cope with the multi-source heterogeneity of IoT terminal device data,and secondly,this paper studies and analyses the user terminal privacy disclosure propensity characteristics,and designs a standardised mutual information based on this type of characteristics The algorithm calculates the degree of privacy leakage of attributes carried by user requests by quantifying user privacy preferences and the degree of exposure of user attribute information to support the decision of edge compputing nodes to encrypt user attribute information,and verifies the effectiveness of the algorithm through experiments.Finally,this paper systematically builds the corresponding cloud service platform through StarlingX open source edge cloud,designs,implements and deploys the authorization mechanism and related services based on the edge computing platform,and verifies the authorization efficiency of the mechanism through testing.