Update Recovery Attacks on Encrypted Database Within Two Updates Using Range Queries Leakage

Jianting Ning, Geong Sen Poh, Xinyi Huang, Robert H. Deng, Shuwei Cao, Ee-Chien Chang

Abstract: Recently, reconstruction attacks on static encrypted database supporting range queries have been proposed. However, attacks on encrypted database within two updates in the similar setting have not been studied extensively. As far as we know, the only work is the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">update recovery attack</i> presented by Grubbs <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">et al.</i> (CCS 2018). Following their seminal work, we present new update recovery attacks for <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">dense</i> dataset (i.e., at least one record corresponding to each value in the range), which enable a deeper understanding of the impact caused by leakages due to updates on dynamic encrypted database. Our first attack aims at recovering the value of a newly added record in the case of one database update. We further demonstrate that the attack can fully reconstruct the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">database counts</i> if the updated value is either the minimum or maximum in the range. We then consider a setting where two distinct records are added separately, which leads to our second attack. We next extend our attacks to the setting where the update operation is deletion. To the best of our knowledge, update recovery attack on database supporting deletion has not been considered before. We demonstrate practicality of our attack via extensive simulations using real dataset.