Proxying Is Enough: Security of Proxying in TLS Oracles and AEAD Context Unforgeability

Abstract

TLS allows a client to securely obtain data from a server, but does not allow the client to offer the data provenance to an external node. TLS oracle protocols are used to solve the problem. Specifically, the verifier node, as an external node, is convinced that the data is indeed coming from a pre-defined TLS server, while remaining unable to access the client’s credentials (e.g., password). Previous TLS oracle protocols such as DECO (CCS 2020) enforced the communication pattern of server-client-verifier and utilized a novel three-party handshake process during TLS to ensure data integrity against potential tempering by the client. However, this approach introduces a significant performance penalty on the client and the verifier. Most recently, some works have proposed to reduce the overhead by putting the verifier (as a proxy) between the server and the client such that the correct TLS transcript is available to the verifier. Nevertheless, these works still rely on heavy two-party secure computations or zero-knowledge proofs. In this work, we push the proxy model to the extreme, where the verifier only needs to forward messages without performing any other heavy computational operations when only the credentials should be protected and the data retrieved from the server could be open to the verifier. Surprisingly, we prove that the thorough proxy model is enough to guarantee security in some common scenarios, allowing a saving of 60-90% in running time under common scenarios.
We first formalize the proxy-based Oracle protocol and functionality that allows the verifier to directly proxy client-server TLS communication, without entering a three-party handshake or interfering with the connection in any way. We then show that for common TLS-based higher-level protocols such as HTTPS, data integrity to the verifier proxy is ensured by the variable padding built into the HTTP protocol semantics. On the other hand, if a TLS-based protocol comes without variable padding, we demonstrate that data integrity cannot be guaranteed. In this context, we then study the case where the TLS response is pre-determined and cannot be tampered with during the connection. We propose the concept of context unforgeability and show that data integrity can also be guaranteed as long as the underlying Authenticated Encryption with Associated Data (AEAD) satisfies context unforgeability. We further show that ChaCha20-Poly1305 satisfies the concept while AES-GCM does not.