Researches On Cloud Outsourcing Optimization Service For Privacy Protection

With the rapid development of network information technology and the generation of massive data sets in the big data environment,multiple participants can not only complete online transactions,but also collaborate to solve optimization problems with the powerful storage and computing capabilities of cloud servers.Cloud outsourcing optimization services can greatly reduce participants' local computing costs and improve the accuracy of calculation results.However,at the same time,the personal data left by participants on the network is easily abused by malicious organizations or units,and there is a risk of leaking participants' private information,and there are many hidden security risks.In addition,the complexity of the network environment prevents multiple participants from confirming each other's identities,so they are unwilling to disclose too much personal privacy data.Therefore,while protecting participants' sensitive information,providing participants with secure cloud outsourcing optimization services,cooperating to solve optimization problems,completing online transactions have received extensive attention and research.The secure cloud outsourcing optimization service selects the appropriate cryptography technology according to the computing needs of the participants,and designs the protocol that meets the needs,so as to solve the collaborative optimization problem with privacy protection between a group of mutually distrustful participants.In the fields of electronic auctions,machine learning,mathematical statistics,electronic voting,etc.,secure cloud outsourcing optimization services enable participants to overcome geographical restrictions,realize collaborative computing,obtain optimal solutions to optimization problems,and provide participants with convenient and efficient services.Aiming at the problem of protecting the personal privacy of participants in the big data environment and combining with the practical application of secure cloud outsourcing optimization services,this paper conducts research on the optimization problem from two aspects: electronic auction protocol and ridge regression fitting.The main research contents of this paper are as follows:(1)Aiming at the privacy problem of bidding information leakage during the electronic auction process,this paper proposes a privacy-preserving and verifiable electronic auction scheme,which uses a secure greedy algorithm to obtain the optimal solution in the auction scheme.A one-way and monotonically increasing function is used to process the bidding prices of the bidders to ensure that the auctioneers can sort the bidding prices without recovering the bid prices;The ELGamal homomorphic encryption algorithm is used to protect the sequence of goods that the bidders want to obtain,so that the auctioneer can calculate on the ciphertext and verify whether there are sold products without knowing bidder's product sequence;We propose a privacy-preserving and verifiable payment determination algorithm to calculate the payment that the winner should pay,and use the Nyberg-Rueppel blind signature algorithm so that the winnner can verify the correctness of the payment.The protocol can meet the basic characteristics and security requirements of electronic auctions in multiple application scenarios,can resist internal and external attackers effectively,and has good practicability and scalability.(2)Aiming at the problem of personal privacy leakage of sample data in the ridge regression fitting process,this paper proposes a privacy-preserving ridge regression fitting scheme,which uses a secure gradient descent algorithm to implement ridge regression model fitting.Encrypt users' privacy data through BGN encryption algorithm,and complete model training on ciphertext.In this protocol,the user encrypts the sample data and sends it to the cloud server.The cloud server completes the secure calculation and fitting in the ciphertext domain,and the crypto service provider,referred to as CSP,provides assistance in the calculation.The user obtains the ridge regression fitting model parameters with the help of CSP.No party involved can know the sample data of other participants,and there is no need for a fully trusted third party.This protocol can effectively prevent internal attacks by external attackers and semi-honest servers.Under the premise of not revealing users' privacy,relevant fitting calculations can be completed effectively,which has high practical value.(3)The actual application of cloud outsourcing optimization service for privacy protection under big data environment is deeply studied.The correctness and security of the two protocols proposed in this paper are analyzed,and simulation experiments are conducted to compare the various performances of the schemes in detail...