Publications scientifiques
VCUFE-IP: Verifiable Ciphertext-Updatable Inner-Product Functional Encryption for Auditable Cloud Data Transfers (ouvre dans un nouvel onglet)
Auteurs
Université de Yaoundé I
Autres auteurs
Serge Alain Ebele
Publications scientifiques
Serge Alain Ebele
Chantal Marguerite Mveh-Abia
Clément Marie Atangana Amougou
Arthur Ulrich Ewane
René Ndoundam
HAL
Ciphertext-updatable functional encryption lets an authorized party convert a ciphertext from one tag to another without decrypting it. This is useful in cloud data-sharing workflows, but existing ciphertext-updatable functional encryption models assume that the update server performs the conversion correctly. In adversarial deployments, a compromised server may return a malformed ciphertext, omit or misapply an update token, or report a forged functional decryption result. We present VCUFE-IP, a verifiable ciphertext-updatable functional encryption construction for bounded inner products. The construction uses asymmetric bilinear groups and separates fresh and updated ciphertexts into two tag-hash levels. Update tokens are vectors of group elements and updates are deterministic, which makes update verification a public recomputation. We additionally give Fiat-Shamir proof systems for ciphertext well-formedness and delegated result correctness. The public transcript authenticates the encryption bases, functional-key view, and update token before checking group equations. The result proof is formulated over public group elements and is bound to the signed verification view of the issued functional key; it does not require the verifier to know discrete logarithms of ciphertext components or the hidden function vector. We prove algebraic correctness, state explicit game-based confidentiality and verifiability reductions, and provide a reference implementation with executable tests over both an exponenttracing backend and a BLS12-381 backend. The artifact also includes symbolic identity checks, an analytical state-of-the-art comparison, BLS12-381 single-dimension and scaling benchmarks, optimized-pairing measurements, container and CI metadata, and a local reproducibility archive prepared for public deposit.