Secure Multi-party Computation Protocol Design Toward Massive Data

With the development of emerging technologies,such as: cloud computing,Internet of things,social networks and so on,the data produced in the Internet environment grows extensively.Big data is precious treasure which contains tremendous research value,so how to dig out the information from big data has attained growing attention both from academic and industrial fields.What is more,it has become a hotspot issue for the government home and abroad.However,the rapid development of big data also has caused some undeniable troubles for the traditional personal privacy issues.Secure multiparty computation problems are widely regarded as one of the tough problems.Under present network environment,the basic requirement of protecting the safety of users' data is of essential importance.Thus secure multi-party computation has attracted widespread attention under this kind of consideration.In the massive data environment,secure multi-party computation problem coalesce cryptography and distributed computing technology,which is an important research direction in information security.A.C.Yao,the Turing Award winner,put forward the secure multi-party computation problem in 1982 firstly and gives its definition.Because secure multi-party computation can solve many practical problems,this issue got the worldwide attention of the researchers,which has a broad application scenarios,such as solve machine learning,data mining,DNA sequencing,scientific computing,statistical analysis and many other computing problems.Now there are a lot of research results of secure multi-party computation,but some neglected points are worth of studying deeply and widely.We do some further researchs based on the existing research results in this thesis,and the main works and achievements are mainly embodied into the following aspects:1.To give a related review of massive data set and secure multi-party computation: Mainly including the related review of massive data privacy protection,the overview of secure multi-party computation problem which is massive-data-set-oriented,the relations of secure multi-party computation protocol and cryptography.We further introduce the related theories of secure multi-party computation;2.To design secure multi-party computation protocol based on cosine similarity: As an important method of similarity measurement,cosine similarity has been widely used in dealing with high-dimensional data.By combing the cosine similarity and the calculation of massive data sets,we have designed an effective secure multi-party computation protocol,which can be used for solving scientific computing problems under the scenario of privacy protection,such as keyword matching,similarity analysis and DNA sequence query and so on.Finally,we give the performance evaluation of the algorithm;3.To design secure multi-party computation protocol based on homomorphic encryption: First of all,we have got the basic algorithm by fully analyze homomorphic encryption,and then put forward a kind of secure multi-party computation protocol with high efficiency and security based on the vector homomorphic encryption.The secure multi-party computation protocol can be used for solving scientific computing problems under the scenario of cryptographic data,such as disease forecasting,the privacy comparison of the vector collection,and the encrypted data statistical analysis,etc.The whole procedure includes the design,analysis and performance evaluation of the algorithm;4.To design secure multi-party computation protocol based on private set intersection: First of all,we have to analyze the research background of privacy-preserving set intersection and analyze its practical significance.Furthermore,we get to put forward a secure multi-party computation protocol based on private set intersection.The secure multi-party computation protocol can be used for intersection calculation under the scenario of privacy protection,such as location-based services,social networking,complete test of the human genome and collaborative zombie network monitoring,etc.The whole project includes the design,analysis and performance evaluation of the protocol.