International Association
for Cryptologic Research

In this work, we present two generic frameworks for leakage-resilient attribute-based encryption (ABE), which is an improved version of ABE that can be proven secure even when part of the secret key is leaked. Our frameworks rely on the standard assumption ($k$-Lin) over prime-order groups. The first framework is designed for leakage-resilient ABE with attribute-hiding in the bounded leakage model. Prior to this work, no one had yet derived a generic leakage-resilient ABE framework with attribute-hiding. The second framework provides a generic method to construct leakage-resilient ABE in the continual leakage model. It is compatible with Zhang et al.'s work [DCC 2018] but more generic. Concretely, Zhang et al.'s framework cannot act on some specific ABE schemes while ours manages to do that. Technically, our frameworks are built on the predicate encoding of Chen et al.'s [EUROCRYPT 2015] combined with a method of adding redundancy. At last, several instantiations are derived from our frameworks, which cover the cases of zero inner-product predicate and non-zero inner-product predicate.