A unified framework for concurrent security: Universal composability from stand-alone non-malleability

We present a unified framework for obtaining Universally Composable (UC) protocols by relying on stand-alone secure non-malleable commitments. Essentially all results on concurrent secure computation|both in relaxed models (e.g., quasi-polynomial time simulation), or with trusted set-up assumptions (e.g., the CRS model, the imperfect CRS model, or the timing model)- are obtained as special cases of our framework. This not only leads to conceptually simpler solutions, but also to improved set-up assumptions, roundcomplexity, and computational assumptions. Additionally, this framework allows us to consider new relaxed models of security: we show that UC security where the adversary is a uniform PPT but the simulator is allowed to be a non-uniform PPT (i.e., essentially, traditional UC security, but with a non-uniform reduction) is possible without any trusted set-up. This gives the first results on concurrent secure computation without set-up, which can be used for securely computing "computationallysensitive" functionalities (e.g., data-base queries, "proof of work"-protocols, or playing bridge on the Internet).