Secure and Efficient Keyword-Based Threat Detection

Public Key Encryption with Keyword Search (PEKS) is a common cryptographic primitive that facilitates secure keyword search over encrypted data so that the keywords hidden inside are concealed from unauthorized parties. For facilitating more complex search operations, Expressive PEKS (EPEKS) generalizes the fundamental PEKS to accommodate expressive queries like conjunctive and disjunctive keyword search. Provision of these expressive features is usually facilitated by AttributeBased Encryption (ABE). With additional features, though, current EPEKS systems are susceptible to keyword guessing attacks that compromise the confidentiality of the search queries. This work examines Shen et al.'s (2019) modified expressive PEKS scheme from a keyword guessing attack vulnerability perspective. To mitigate such security attacks, we present a new improved encryption-decryption model with expressive search functionality but improved guessing attack resistance. Our approach employs a secure means of key sharing, e.g., Diffie-Hellman or Elliptic Curve Diffie-Hellman (ECDH), to facilitate secure communication. We deploy the suggested framework and evaluate its performance and security attributes, demonstrating its capacity to provide expressive search ability as well as resist keyword inference attacks successfully