Dynamic authenticated keyword search in hybrid-storage blockchain

초록

Many real -world applications outsource their storage and computation to a cloud. But the cloud cannot be fully trusted. Blockchain has become a promising solution to secure data storage and retrieval. To address scalability problem in permissionless blockchain systems, a hybrid -storage model is a prevailing approach where meta -data is stored on blockchain, and raw data are outsourced to off -chain storage, e.g., a cloud service provider (SP). This hybrid -storage model can enable clients to securely verify completeness, soundness, and freshness of query results without putting trust in SP. Existing works in such a hybrid -storage model suffer from high on -chain costs and relatively low query speeds. With this model, we propose a novel authenticated data structure for authenticated keyword search in a dynamic environment (named DAKS) where data are frequently added or updated. First, we design a membership proof scheme having constant proof size (CMPS) as a building block to guarantee soundness of query results. CMPS gives our DAKS high off -chain efficiency in both client and server. Besides the soundness, our scheme guarantees completeness of the results by designing a data structure and further proposing algorithms for authenticated keyword searches. In addition to high off -chain efficiency, DAKS minimizes the on -chain cost because only one digest needs storing on blockchain. Extensive evaluations on benchmark datasets show that DAKS not only has much cheaper on -chain costs than prior work, also has faster SP processing time and client verification time.

키워드