Research On A New Framework Of Privacy Preserving Outsourced Computing In Cloud Computing

With the development of computer network,software and hardware technology,cloud computing technology is increasingly popular.Outsourced computing emerges in such a scenario.Outsourced computing technology allows users with limited computing resources to outsource data to cloud servers for processing.However,outsourcing private data to honest and curious or malicious cloud service providers brings a series of security implications and privacy breaches.Therefore,This paper devote to research methods for privacy-preserving outsourced computation,and propose a new framework for efficient and privacy-preserving outsourced computation.Using this framework,a user using mobile devices and personal computers can encrypt his data(both integer and floating point numbers)and securely upload them to the cloud for storage and computation.The cloud cannot obtain private data while completing data processing tasks.The main research work of this paper is as follows:1.This paper constructs a secure and efficient outsourcing computing framework based on a one-way trapdoor permutation and a self-developed symmetric somewhat homomorphic encryption algorithm.Different from the previous Fully Homomorphic Encryption(FHE)scheme based on RLWE(Ring Learning with Errors),Our scheme has the characteristics of higher efficiency(tens to hundreds of times),greater flexibility and easier deployment in practical scenarios.We combine two privacy protection technologies,Homomorphic Encryption(HE)and Secure Multi-Party Computing(MPC),to realize the ciphertext calculation of almost arbitrary functions.Including addition,multiplication,comparison,modulo,division,square root,etc.We first propose a basic encryption scheme,and then give more function-rich computing sub-protocols on the basis of this basic scheme,so as to propose a complete technical system.2.Since most of the data in the real world are rational numbers and floating-point numbers,we give a floating-point number processing mechanism based on the integer calculation sub-protocols.The general approach is to use scaling factors and rounding to convert floating-point numbers.It becomes fixed-point processing.This processing method greatly reduces the complexity of floating-point calculation or rational number calculation,and does not need to frequently consider the decimal point alignment in floating-point number addition and the expansion of the numerator and denominator of rational numbers.In addition to implementing fixed-point addition,subtraction,and multiplication,we also provide a division calculation algorithm with higher precision.3.In addition to the above features,our proposed framework can securely handle multi-user’s data.Because the data of multiple users is encrypted by different keys,we securely transfer the ciphertext under multiple keys to a Cryptographic Service Provider(CSP)through an ingenious conversion mechanism,and then execute the calculation protocol under the CSP’s key to achieve the calculation of multi-key data.The final calculation results are then sent to an authorized receiver for decryption after a conversion algorithm.