Many recent studies focus on dynamic searchable encryption (DSE), which provides efficient data-search and dataupdate services directly on outsourced private data. Most encryption schemes are not optimized for update-intensive cases, which say that the same data record is frequently added and deleted from the database. How to build an efficient and secure DSE scheme for update-intensive data is still challenging. We propose UI-SE, the first DSE scheme that achieves single-roundtrip interaction,

more »

... near-zero client storage, and backward privacy without any insertion patterns. UI-SE involves a new tree data structure, named OU-tree, which supports oblivious data updates without any access-pattern leakage. We formally prove that UI-SE is adaptively secure under Type-1 − backward privacy, which is stronger than Type-1 backward privacy proposed by Bost et al. in CCS 2017. Experimental data also demonstrate UI-SE has low computational overhead, low local disk usage, and high update performance on scalable datasets.