PEREA

Abstract

Several anonymous authentication schemes allow servers to revoke a misbehaving user's ability to make future accesses. Traditionally, these schemes have relied on powerful TTPs capable of deanonymizing (or linking) users' connections. Recent schemes such as Blacklistable Anonymous Credentials (BLAC) and Enhanced Privacy ID (EPID) support "privacy-enhanced revocation" -- servers can revoke misbehaving users without a TTP's involvement, and without learning the revoked users' identities. In BLAC and EPID, however, the computation required for authentication at the server is linear in the size (L) of the revocation list. We propose PEREA, a new anonymous authentication scheme for which this bottleneck computation is independent of the size of the revocation list. Instead, the time complexity of authentication is linear in the size (K << L) of a revocation window, the number of subsequent authentications before which a user's misbehavior must be recognized if the user is to be revoked. We prove the security of our construction, and have developed a prototype implementation of PEREA to validate its efficiency experimentally.