Research Of Private Set Intersection Computation Method In Multi-party And Multi-scene

With the advent of the Internet era and the application of artificial intelligence technology,applications on terminal devices are often distributed across multiple service devices that interact over the Internet to jointly realize program functions.However,the way existing distributed applications handle interaction data no longer meets the needs of terminal-user’s for personal data privacy protection.Private set intersection(PSI),a protocol specific to secure multi-party computation,allows participants to input private set and compute the set intersection interactively without any information leakage other than the intersection result.PSI is a privacy computing technique that specializes in dealing with distributed set intersections and is now widely followed by academia and industry.Traditional two-party PSI protocols have been extensively studied nowadays,mainly by solving data encryption primitives and security comparisons to achieve privacy-preserving intersection computation.Two-party PSI protocols have been designed for all types of distributed scenarios(e.g.,small set scenario,unbalanced scenario,large set scenario)with efficient cryptographic primitives and security comparisons applicable to the scenario.Only a small number of studies exist on multi-party PSI protocols.In addition to the problem of two-party PSI protocols,the problem of multi-party collusion to leak the intersection of some participants needs to be addressed.Existing multi-party PSI protocols focus only on large sets of distributed scenarios with a small number of participants,which may be due to the unavoidable communication overhead between participants.However,in real-life scenarios,multi-party PSI protocols have more and meaningful distributed scenarios than two-party PSI protocols.This paper extends the scenario of multi-party PSI protocols in terms of redefining the intersection of multi-party PSI and the relationship between the set size and the number of participants.The main work accomplished is as follows:(1)For existing efficient multi-party PSI protocols based on oblivious programmable pseudo-random function and zero-sharing framework to solve data transmission and multi-party collusion problems,resulting in their applicability only to scenarios with large sets and a small number of participants.Design a framework based on one-round key agreement protocol and embedded unconditional zero sharing to resist data transmission and multi-party collusion problems.Since the cryptographic primitive is a one-round key agreement protocol,embedded unconditional zero sharing requires no additional communication overhead and is therefore suitable for scenarios with small sets and a large number of participant.A small set-large participant private set intersection computation protocol is designed based on a round key agreement protocol and an embedded unconditional zero sharing framework.The security of the protocol under the semi-honest model is demonstrated by the ideal-reality model,which is upgraded to malicious security by only a small changes.Compared to existing multi-party PSI protocols,it has the lowest number of communication rounds and the communication complexity is only related to statistical security parameters.Our protocol has the best operational efficiency and communication load under the small set(n<512),and the advantages of our protocol become more obvious as the number of participants gradually increases.(2)After investigating the definition of intersection in existing multi-party PSI protocols and their practical application scenarios,it was found that the computational protocol for the Over-Threshold Multi-party Private Set Intersection is still stuck in a public-key cryptography-based implementation,and the computational efficiency is extremely low leading to impractical deployment.To efficiently solve the Over-Threshold intersection computation problem,we design a novel cryptographic component,oblivious programmable pseudo-random function with secret sharing(OPPR-SS)only rely on symmetric key based cryptographic operations.It has the security and programmability of oblivious programmable pseudo-random function and(t,m)-reconfigurability of threshold secret sharing.And a semi-honest secure multiparty private set over-threshold intersection computation protocol is designed based on oblivious programmable pseudo-random secret sharing.The theoretical analysis and experimental comparison with existing protocols show that this protocol has more efficient computational efficiency and lower communication overhead.