Cardinality-based inference control in data cubes*

Abstract

This paper addresses the inference problem in on-line analytical processing (OLAP) systems. The inference problem occurs when the exact values of sensitive attributes can be determined through answers to OLAP queries. Most existing inference control methods are computationally expensive for OLAP systems, because they ignore the special structures of OLAP queries. By exploiting such structures, we derive cardinality-based sufficient conditions for safe OLAP data cubes. Specifically, data cubes are safe from inferences if their core cuboids are dense enough, in the sense that the number of known values is under a tight bound. We then apply the sufficient conditions on the basis of a three-tier inference control model. The model introduces an aggregation tier between data and queries. The aggregation tier represents a collection of safe data cubes that are pre-computed over a partition of the data using the proposed sufficient conditions. The aggregation tier is then used to provide users with inference-free queries. Our approach mitigates the performance penalty of inference control, because partitioning the data yields smaller input to inference control algorithms, pre-computing the aggregation tier reduces on-line delay, and using cardinality-based conditions guarantees linear-time complexity.

A secure multiparty computation privacy preserving OLAP framework over distributed XML data

“…Such a problem, which significantly impacts on the trustworthiness and reliability of OLAP server platforms, has motivated the development of recent approaches addressing the problem of devising meaningful privacy preserving OLAP techniques (e.g., [2,11,16,23,25,35,36,37]). This is a fundamental problem in Privacy Preserving Data Mining [1] research and has recently attracted the interest of a large community of Database and Data Warehousing researchers [4,10,33].…”

Privacy preserving OLAP

“…There has been recent work [22,23] to specify authorization and control inferences for OLAP data cubes. However the model assumes that the data resides at a single server, unlike our problem, where private data is integrated from multiple clients.…”

“…Consider the tree in Figure 3 built on randomized [21][22][23][24][25][26][27][28][29][30] ∧ salary[25k-100k] ∧ Q ) are reconstructed for T . The number of columns in the count query did not increase at this split on age, which was already present among the original set of queried columns.…”

Auditing sum-queries to make a statistical database secure

“…Answering this question raises some computational problems (recognizing evaluable sum-queries, updating the information model, computing a feasibility range), whose solutions depend on the data type of the response variable. If it is of real type, then standard algebraic methods can be used to solve all of them efficiently; moreover, if sum-queries contain all the "group-by" clause, that is, if they are table queries (or "cuboids" [24,25]), then there also exist cardinality-based conditions that are sufficient for them to be inference free [24,25,28]. If it is of nonnegative type, then we can resort to linear-programming or integer linear-programming methods depending on the specific data type.…”