Secure Multi-Party Processing Against Malicious Parties
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multi-party calculation methods are not secure when parties exhibit malicious behavior, as they can learn information about each other's data, even if the unconscious transfer protocol is secure.
Innovation Solution
The method involves a double execution of the calculation series with role reversal between parties, followed by an equality test to detect cheating, and uses double masking with invertible random values to protect results, ensuring security without significantly increasing execution time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a standard multi-party calculation method using unconscious transfer protocol is used, then the calculation can be performed with semi-honest security, but the security breaks down when parties exhibit malicious behavior
Solution Approach 1:
The patent applies preliminary action by performing authentication and generating authentication tags before the main calculation occurs. The parties authenticate their inputs and intermediate results in advance using cryptographic authentication codes, so that when the actual calculation is performed, the security against malicious manipulation is already established. This allows the protocol to maintain security without requiring complex verification mechanisms during the calculation itself.
Solution Approach 2:
The patent introduces authentication tags as an intermediary mechanism between the parties' data. These tags act as mediators that verify the integrity and authenticity of transmitted values without revealing the actual data. The authentication tags enable the protocol to detect malicious behavior while maintaining the confidentiality of the underlying calculation, thus resolving the contradiction between security and complexity.
2Reliability
If authentication mechanisms are added to protect against malicious parties, then security is improved, but execution time increases
Solution Approach 1:
The authentication tags are generated and verified in advance during setup phases, allowing the main calculation to proceed efficiently without repeated authentication overhead. The preliminary authentication establishes trust boundaries that prevent malicious parties from manipulating results during the actual computation, thus minimizing time loss during the critical calculation phase.
Solution Approach 2:
Each party independently generates and verifies their own authentication tags for their inputs and intermediate results. This self-service approach to authentication eliminates the need for complex centralized verification mechanisms, reducing the overall execution time while maintaining security against malicious behavior.
3Reliability
If double execution with role reversal is performed to detect cheating, then security against malicious parties is improved, but the execution time significantly increases
Solution Approach 1:
The patent extracts the security verification function into a separate authentication mechanism using cryptographic tags, rather than requiring full double execution of the calculation. By taking out the verification aspect and handling it through efficient authentication code comparison, the protocol achieves malicious behavior detection without the time penalty of complete dual execution.
Solution Approach 2:
The patent replaces the mechanical approach of double execution with a cryptographic substitution using authentication tags. Instead of physically re-executing the entire calculation with reversed roles, the system uses cryptographic verification to achieve the same security goal more efficiently, substituting computational mechanics with cryptographic mathematics.
Data Source
Figure 1a
Figure 1b
Figure 2a
AI summary
The invention proposes a method comprising the evaluation of a function F obtained by applying a first operation to n sub-functions fi, the evaluation comprising: - the implementation of a series of computation steps in which a first unit takes the role of client and a second unit takes the role of server, and - the repetition of the series of computation steps in which the roles of client and server are exchanged between the units, each series of steps comprising: a) random generation, by the server, of first data, and of a second data point, b) for each sub-function fi, generation by the server of a set of elements formed by: o a result of fi evaluated in the data of the client and the server, o masked by a first data point, by application of the first operation between the result and the first data point, and o masked by the second data point.by applying a second operation, different from and distributive with respect to the first, between the masked result and the second data point; c) retrieval, by the client, through unconscious transfer, of intermediate data corresponding to one of the elements generated by the server; d) generation, by the server, of a first part of the result, by: • masking each first data point by the second data point, • applying the first operation to all the masked first data points; and e) generation, by the client, of a second part of the result, by applying the first operation to all the intermediate data points.