Research On Data Privacy Protection Technology For Edge Computing

With the development of Fifth Generation Mobile Communication and Industrial Internet,edge computing plays a more and more important role in practical applications.Edge node devices provide personalized services with collecting user information which may lead to leakages of privacy.If data publisher publishes data without any changes,it is benefit for data user to analyze or mine data and easy to cause privacy leakage.Therefore,ensuring availability and privacy of data at the same time is hot research direction in privacy protection of edge computing during data publishing period.This paper introduces the development of edge computing and its research status of data privacy protection,and focuses on the data privacy issues in data publishing.Considering the situation where edge computing model handles lots of data,this paper proposes a static data publishing algorithm named KSV-MDAV which combine cluster algorithm and sensitive attributes weight on the basis of Variable-Maximum Distance Average Vector(V-MDAV)algorithm.A privacy protection model for continuous data publishing in edge computing is also proposed.The main work of this paper is as follows:(1)Considering the large amount of data in edge computing scenarios,this paper proposes to introduce the clustering algorithm into the micro-aggregation process.The algorithm divides the data set into small data sets for processing.In order to improve the clustering effect,the distance between different records is calculated through distance metric based on weights,and each litter dataset is divided as uniformly as possible.(2)A weight-based sensitivity division method is proposed.This paper considers the importance of different sensitive attributes and the influence of the distribution of sensitive attributes on privacy leakage.The weight of sensitive attribute consists of sensitive attributes distribution and sensitivity of different sensitive attribute.It can avoid the phenomenon that higher sensitive attribute values are aggregated in a QI group,which makes the algorithm more flexible and effective in protecting sensitive attribute values.(3)This paper proposes a dynamic dataset update privacy protection model called CDP-VMDAV for continuous data publishing.Based on the anonymous dataset released last time,the model performs dynamic redistribution of the dataset by incremental clustering to reduce computational redundancy.It can also protect sensitive information from inference attacks.At the end of chapter three and chapter four,this paper shows simulation and comparison results.Comparing with traditional algorithms,the algorithm proposed in this paper reduces the running time and risk of privacy leakage,and gets better balances between data availability and privacy.Finally,this paper summarizes whole research results,and discusses the future research directions.In this paper,we use 8 figures,19 tables,55 references.