Paper 2023/1377

Janus: Fast Privacy-Preserving Data Provenance For TLS 1.3

Abstract

Web users can gather data from secure endpoints and demonstrate the provenance of sensitive data to any third party by using privacy-preserving TLS oracles. In practice, privacy-preserving TLS oracles are practical in verifying private data up to 1 kB in size selectively, which limits their applicability to larger sensitive data sets. In this work, we introduce a new oracle protocol for TLS, which reaches new scales in selectively verifying the provenance of confidential web data. The novelty of our work is a construction which combines an honest verifier zero-knowledge proof system with a new secure validation phase tailored to an asymmetric privacy setting between collaborative TLS clients. Compared to previous works, our construction proves non-algebraic TLS algorithms faster while retaining equivalent security properties. Concerning TLS 1.3, we optimize end-to-end performances and show how the garble-then-prove paradigm can benefit from previously established authenticity to employ semi-honest secure computations without authentic garbling. Our performance improvements show that 8 kB of sensitive TLS data can be verified in 6.7 seconds, outperforming related works significantly. With that, we enable new boundaries to verify the provenance of confidential documents of the web.