Public-key encryption with keyword search: a generic construction secure against online and offline keyword guessing attacks

In public-key setting, the problem of searching for keywords in encrypted data is handled by the notion of public-key encryption with keyword search (PEKS). An important challenge in designing secure PEKS schemes is providing resistance against variants of an attack known as the keyword guessing attack (KGA). Basically, by KGA, an adversary is able to determine the searched keyword through using the data communicated in the search process. Security against offline KGA performed by both inside/outside adversaries is well-studied in the literature. However, this is not true about the online version. In this paper, we employ a technique called ciphertext re-randomization to propose a generic construction for designing PEKS schemes which are secure against both online and offline KGAs performed by outsiders. We show that compared to existing literature, our construction is more efficient in terms of computational and communication costs.