Efficient verifiable databases with additional insertion and deletion operations in cloud computing

Abstract

Verifiable database (VDB) schemes allow the data owner to outsource a large database to the cloud so that any resource-constraint client could later retrieve the database records and check whether the cloud returns valid records or not. Meanwhile, the database can be updated by the data owner. To the best of our knowledge, there is no secure and efficient VDB scheme supporting public verification and all kinds of update operations (i.e., insertion, modification, and deletion) simultaneously. To resolve this problem, we propose an efficient VDB scheme by incorporating vector commitment (VC) with Merkle Interval Hash Tree (MIHT). To enhance the security (i.e., resist the forward automatic update attack), we re-formalize the syntax of VC and present a new construction of VC based on modified Generalized Bilinear Inversion (mGBI) assumption. Our proposed VC scheme is used for guaranteeing the integrity of data. Like the traditional VC schemes, our proposed scheme remains publicly verifiable. MIHT is a new primitive introduced by us, which is mainly used to support all update operations. Security analysis shows that our proposed VDB scheme can achieve real-world security requirements. The detailed performance analyses and simulations show that our proposed schemes are more practical in the real world.