Securing Communications after Device Compromise

密码学中的传统对手模型主要关注无法访问端点的网络对手。我们现在知道，这是天真的-由于当前的安全架构，具有物理访问权限的对手可以访问所有过去和未来的秘密密钥。以前解决这个问题的尝试包括前向保密的概念，这是一种加密属性，它确保秘密（例如密钥）的泄露不会导致所有未来秘密的泄露。这可以通过密钥棘轮或密钥轮换来完成，这会在每次会话后更新密钥。一个相关的认证技术是恐慌密码，它会提醒第三方的胁迫-这当然是假设对手不知道这种隐蔽机制，这使得一个弱模型。该项目将开始开发一个新的安全模型，以对抗强大的物理对手-我们认为，以前的模型不满足我们的新模型的某些安全属性。反过来，我们将开发新的加密技术来减轻威胁，并提供可证明的安全保证。该项目将特别针对内部威胁，亲密伴侣威胁，以及在认证过程中的通信胁迫，包括电子投票。该项目属于EPSRC全球无障碍研究领域的福尔斯。

Traditional adversary models in cryptography focus primarily on a network adversary with no access to endpoints. We now know that this was naïve - due to current security architectures, an adversary with physical access may have access to all past and future secret keys for their communications. Previous attempts at solving this problem include the notion of forward secrecy, a cryptographic property which ensures that compromise of a secret (e.g. key) does not lead to compromise of all future secrets. This can be done by key ratcheting or key rotation, which updates the secret key after every session. A related technique on authentication is panic password, which would alert a third party of the duress - this is of course assuming that the adversary isn't aware of this covert mechanism, which makes for a weak model.The project will start with developing a new security model against a strong physical adversary - we argue that the previous models do not satisfy certain security properties in our new model. In turn, we will develop new cryptographic techniques to mitigate the threats, with provably secure guarantees. This project will be relevant in particular against insider threat, intimate partner threat, as well as communicating duress during authentication, including e-Voting.This project falls within the EPSRC Global Uncertainties research area.