Secure and Efficient Privacy-Preserving Ciphertext Retrieval in Connected Vehicular Cloud Computing

Abstract

As vehicular equipment is becoming more and more intelligent, the vehicular information service, as the main means of capturing information, has been far from able to meet the needs of occupants [1, 2]. Cloud computing, with its powerful computing and storage capabilities, convenient network access, energy saving and excellent scalability, reliability, availability, and other advantages, can be an effective solution to the limitations of existing automotive information services. Connected vehicular cloud computing, which combines cloud computing and VANETs, has the characteristics of both a cloud platform and a mobile ad hoc network, including autonomy and no fixed structure, good scalability, and so on. However, during the information retrieval, high-density node distribution and high-speed mobile nodes may directly affect the information transmission capacity of a VANET by information tampering, transmission delay, and other issues. In this article, we propose a ciphertext-based search system that exploits RSUs as super peers for connected vehicular cloud computing. The proposed system supports ciphertext retrieval for related documents. In the proposed system, all the computations and retrieval operations are handled by super stationary peers, while documents are stored in the cloud to achieve high efficiency and security of the index structure. We can also reduce the impact of vehicle dynamics on the information retrieval process in this way. In our system, the indexing efficiency is also improved by utilizing a hybrid indexing structure in which binary trees are nested in a B+ tree. Through security analysis and performance evaluation, we demonstrate that our proposal can achieve acceptable security and efficiency.