Privacy-Preserving Secret Sharing using Fully Homomorphic Encryption

This paper proposes a new approach to improving the security and confidentiality of data exchanged in a network, by combining Shamir's secret sharing and fully homomorphic encryption. This approach is based on the partitioning of infor-mation using Shamir's threshold scheme, enriched by the addition of noise to improve confidentiality. The segments thus created, protected by homomorphic encryption, are sent via distinct paths to an intermediate proxy node capable of directly manipulating the encrypted data and reconstructing the encrypted secret using Lagrange interpolation, then forwarding it to its recipient for de-cryption. Experimental results, obtained on a Windows platform using Python and the NetworkX library, demonstrate that this method not only guarantees data confidentiality and security, but also preserves confidentiality throughout the routing process. This approach represents a significant advance on existing techniques, enhancing data security while maintaining confidentiality at every stage.