Dynamic Adaptive Bloom Filter Based Privacy-Preserving Phrase Search for Encrypted Iot Cloud Data

The use of IoT devices in huge numbers results in massive amounts of sensitive data being outsource into cloud storage systems, making it necessary to design efficient search schemes over such encrypted data. Phrase Search allows users to retrieve their documents by matching exact sequences of keywords, but existing solutions have not found a good balance between search time, storage space, and privacy loss due to their static design. In this work, we propose an efficient Dynamic Adaptive Bloom Filter based Privacy-Preserving Phrase Search (DABF-PPS) scheme for encrypted IoT cloud data. We introduce a dynamic parameter optimization method to adaptively adjust Bloom filters according to document set properties and queries. A hierarchical n-gram Bloom filter index is designed to allow phrase searches without exposing keyword locations. Furthermore, our solution supports dynamic updating of indices with the help of counting Bloom filter structures and lazy deletion. Our security analysis proves that our solution provides forward and limited backward privacy against chosen-keyword attacks. Our experiments on real-world IoT datasets show that our approach decreases search time and storage space while maintaining acceptable false-positive rates.