Abstract

Abstract. An interactive proof system (or argument) (P, V)isconcurrent zero-knowledge if whenever the prover engages in polynomially many concurrent executions of (P, V), with (possibly distinct) colluding polynomial time bounded verifiers V1,...,Vpoly(n), the entire undertaking is zero-knowledge. Dwork, Naor, and Sahai recently showed the existence of a large class of concurrent zero-knowledge arguments, including arguments for all of NP, under a reasonable assumption on the behavior of clocks of nonfaulty processors. In this paper, we continue the study of concurrent zero-knowledge arguments. After observing that, without recourse to timing, the existence of a trusted center considerably simplifies the design and proof of many concurrent zero-knowledge arguments (again including arguments for all of NP), we design a preprocessing protocol protocol, making use of timing, to simulate the trusted center for the purposes of achieving concurrent zero-knowledge. Once a particular prover and verifier have executed the preprocessing protocol protocol, any polynomial number of subsequent executions of a rich class of protocols will be concurrent zero-knowledge. 1

Citations