An efficient authenticated data structure for dynamic data set based on B+ tree

Abstract

In cloud storage system, the client stores large volume of data at remote directories which is out of the client's control and may be malicious. In such setting, the ability to efficiently check the integrity of remotely stored data is an important security property. In this paper, we focus on a simple and very important form of remote data authentication problem, where we authenticate membership queries over a dynamic set of n data elements that is stored at untrusted directories. Some existing verification methods for membership queries, such as Merkle Hash Tree (MHT), Skip List and RSA Tree are overviewed first. These methods all achieve the goal for verification, but their data structures are not agile enough for data update. They may have high time complexity or cause reconstruction of the whole data structure frequently when update operation is preceded. So we turn to B+ tree data structure and make it work with RSA accumulator to form a new verification scheme, which implements the verification and incurs low computational costs for membership query of dynamic data set.