Abstract

In Ciphertext Policy Attribute-Based Encryption (CP-ABE) system, a set of attributes is associated with the private keys of each user. Also, the ciphertext is attached with a policy which is defined over that set of attributes. A user can decrypt the ciphertext if the ciphertext’s policy is satisfied by the attributes associated hith her private key. Traditional CP-ABE schemes, based on number theoretic problems, rely on a trustworthy central authority. But in many distributed applications it is expected that such authorities should be decentralized to avoid the risks of single-point failure. While the number theory-based hardness problems are prone to quantum attacks, lattice-based hardness problems can resist such attacks. In this paper, we construct a Decentralized Ciphertext-Policy Attribute-Based Encryption (DCP-ABE) scheme. Under this scheme, any participating entity can act as an authority by creating a public key. The athority utilizes the users’ attributes to generate the private keys for them. Any user can encrypt data in terms of any monotone access structure over attributes issued from any chosen set of authorities. Hence the protocol does not depend on any central authority. We utilize Learning With Errors over Rings (R-LWE) as the underlying hardness assumption for te protocol. The proposed post-quantum protocol achieves security under selective-set model whereby adversaries are allowed to corrupt any authority only statically through adaptive key queries.