The pervasive, ubiquitous, and heterogeneous properties of IoT make securing IoT systems a very challenging task. More so when access and storage are performed through a cloud-based IoT system. IoT data stored on cloud should be encrypted to ensure data privacy. It is also crucial to allow only authorized entities to access and decrypt the encrypted data. In this work, we propose a ciphertext-policy attribute-based encryption (CP-ABE) scheme that enables fine-grained access control of encrypted IoT data on cloud. CP-ABE is regarded as a highly promising approach to provide flexible and fine-grained access control, which is very much suited to secure cloud based IoT systems. We first present an access control system model of CloudIoT platform by using ABE. Based on the presented system model, we construct a ciphertext-policy hiding CP-ABE scheme, which guarantees the privacy of the users. We further construct a white-box traceable CP-ABE scheme with accountability in order to address the user key abuse and authorization center key abuse. Experiment illustrates the proposed systems are efficient.

Attribute-based encryption (ABE) can guarantee confidentiality and achieve fine-grained data access control in a cloud storage system. Due to the fact that every attribute in ABE may be shared by multiple users and each user holds multiple attributes, any single-attribute revocation for some user may affect the other users with the same attribute in the system. Therefore, how to revoke attribute efficiently is an important and challenging problem in ABE schemes. In order to solve above problems, we first give a concrete attack to the existing ABE scheme with attribute revocation. Then, we formalize the definition and security model, which model collusion attack executed by the existing users cooperating with the revoked users. Finally, we present a user collusion avoidance ciphertext-policy ABE scheme with efficient attribute revocation for the cloud storage system. The problem of attribute revocation is solved efficiently by exploiting the concept of an attribute group. When an attribute is revoked from a user, the group manager updates other users’ secret keys. Furthermore, we prove that the proposed scheme is secure against collusion attack launched by the existing users and the revoked users. The security of the proposed scheme is reduced to the computational Diffie–Hellman assumption.

To increase the availability and durability of the outsourced data, many customers store multiple copies on multiple cloud servers. To guarantee the integrity of multi-copies, some provable data possession (PDP) protocols for multi-copy are presented. However, most of previous PDP protocols consider all copies to be stored on only one cloud storage server. In some degree, multi-copy makes little sense in such circumstance. Furthermore, many PDP protocols depend on the technique of public key infrastructure (PKI), which suffers many types of security vulnerabilities and also brings heavy communicational and computational cost. To increase the security and efficiency, we provide a novel identity-based PDP scheme of multi-copy on multiple cloud storage servers. In our scheme, all copies are delivered to different cloud storage servers, which work cooperatively to store the customer's data. By the homomorphic verifiable tags, the integrity of all copies can be checked simultaneously. The system model and security model of our scheme are provided in the paper. The security for our scheme is proved based on the computation Diffie-Hellman (CDH) hard problem. Analysis and experimental evaluation show that our scheme is efficient and practical. The proposed scheme is the first identity-based PDP scheme for multi-copy and multi-cloud servers.

Attribute based encryption (ABE) is a popular cryptographic technology to protect the security of users' data. However, the decryption cost and ciphertext size restrict the application of ABE in practice. For most existing ABE schemes, the decryption cost and ciphertext size grow linearly with the complexity of access structure. This is undesirable to the devices with limited computing capability and storage space. Outsourced decryption is considered as a feasible method to reduce the user's decryption overhead, which enables a user to outsource a large number of decryption operations to the cloud service provider (CSP). However, outsourced decryption cannot guarantee the correctness of transformation done by the cloud, so it is necessary to check the correctness of outsourced decryption to ensure security for users' data. Current research mainly focuses on verifiability of outsourced decryption for the authorized users. It still remains a challenging issue that how to guarantee the correctness of outsourced decryption for unauthorized users. In this paper, we propose an ABE scheme with verifiable outsourced decryption (called full verifiability for outsourced decryption), which can simultaneously check the correctness for transformed ciphertext for the authorized users and unauthorized users. The proposed ABE scheme with verifiable outsourced decryption is proved to be selective CPA-secure in the standard model.

To solve key escrow problem, we utilize the technique of certificateless signature to present a new remote data possession checking (RDPC) protocols protocol for checking the integrity of data shared among a group. In our scheme, user's private key includes two parts: a partial key generated by the group manager and a secret value chosen by herself/himself. To ensure the right public keys are chosen during the data integrity checking, the public key of each user is associated with her unique identity, for example the name or telephone number. Thus, the certificate is not needed and the problem of key escrow is eliminated too. Meanwhile, the data integrity can still be audited by public verifier without downloading the whole data. In addition, our scheme also supports efficient user revocation from the group. The security of our scheme is reduced to the assumptions of computational Diffie-Hellman (CDH) and discrete logarithm (DL). Experiment results exhibit that the new protocol is very efficient and feasible.