Bucket‐based authentication for outsourced databases

Abstract

SUMMARYWhen outsourced database owners delegate their data to service providers, which might be untrusted or compromised, two issues of data security emerge, including data confidentiality and data integrity. Most of the previous research focuses on only one issue and the solution to integrate two approaches is expensive. In this paper, we propose bucket-based authentication that can keep data confidentiality and meanwhile guarantee data integrity. Specifically, we first propose a new approach based on bucket checksum, which can be used for the authentication of multiple tuples at one time. We then apply bucket checksum to the authentication of various types of queries in static scenarios, including range queries and aggregation queries, such as MIN, MAX, SUM and COUNT queries. In the authentication of aggregation queries, several pruning rules have been proposed to improve performance further. We also extend our approach to dynamic scenarios based on incremental hash. Cost analysis shows the advantages of our approach over previous ones in terms of construction and verification cost. Experimental results show that our approach is both efficient and effective.

Clustering-Based Query Result Authentication for Encrypted Databases in Cloud

“…Then, the user issues a query to the server (line 2). Upon receiving a query, the service provider retrieves the hash index and the private authentication index to return an encrypted original data set for the query (line [3][4] and their signatures (line [5][6][7][8]. Finally, the user generates signatures of query results by using the RSA key sent from the data owner, and compares the signature with result signature (line 9-13).…”

“…Bucket-based index was proposed by Hacigumus [16] and its variants has been studied in [4,[17][18]. A bucket-based index contains {bucket id, data range, number of tuples in the bucket, checksum}.…”

“…For the performance analysis, we compare the performance of the proposed scheme with the existing work, proposed by J. Wang et al [4]. Table 1 represents our experimental environment.…”

“…Previous data authentication researches [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18] can be categorized into three classes: signature-based approaches, authenticated data structures and bucket-based authentication. First, signature-based approaches assign one signature to each data tuple and verify data integrity by comparing all the signatures of data within the results.…”

“…In addition, tree-based index suffer from data update overhead and verification object transmission costs. Finally, J. Wang et al [4] proposed a bucket-based authentication where a bucket contains a bucket id, data range (upper-lower bound) and the number of tuples in a bucket and a checksum. A checksum is similar concept to the data signature and generated by using a Hash function.…”