Rational adversary with flexible utility in secure two-party computation

摘要

Secure two-party computation evaluates a function among two distributed parties without revealing the parties' inputs except for the function's outputs. Secure two-party computation can be applied into various fields like cloud computing, which is a composition of distribute computing, parallel computing and utility computing etc. Rational secure two-party computation may achieve some desirable properties under two assumptions deriving from STOC 2004. However, the emergence of new computing paradigms like pay-as-you-go model restricts the application of rational protocols. Previous adversaries does not consider payment in secure two-party protocols. Therefore, new type of adversaries should be propose for these new paradigms. In this paper, we address this problem by proposing a new kind of rational adversary, who consider payment in his relaxed utilities. The utilities are based on economic incentives instead of standard assumptions. Furthermore, the new rational adversary is assumed to negotiate with rational parties in protocols. It's similar to cost corruption but more flexible. Our new adversary can dynamically negotiate with each rational party in different phases in order to maximize his utilities. To verify the validity of the new adversary, we model a rational secure two-party protocol, which inherits the hybrid framework of STOC 2007. We also prove the security in the presence of the new rational adversary under ideal/real paradigm.