Modeling and Performance of Privacy Preserving Authorization Mechanism for Graph Data
Author
Zahid Pervaiz, Arif Ghafoor, Walid G. Aref
Abstract
There has been significant interest in the development of anonymization schemes for publishing graph data. However, due to strong correlation among users’ social identities, privacy is a major concern in dealing with social network data. In this paper, we propose a privacy-preserving mechanism for publishing graph data to prevent identity disclosure. The framework is a combination of access control and privacy protection mechanisms. The access control policies define selection predicates available to roles/queries and their associated imprecision bounds. Only authorized role/query predicates on sensitive data are allowed by the access control mechanism. For this framework, we define the problem of k-anonymous Bi-constraint Graph Partitioning (k-BGP) and provide its hardness results. We present heuristics for graph data partitioning to satisfy the imprecision and information loss bounds for k-BGP problem. The privacy-protection mechanism anonymizes the graph data with minimal information loss while simultaneously meeting the QoS requirement in terms of satisfying the bounds on the number of roles being satisfied. This approach provides an anonymous view based on the target class of role-based workloads for graph data. We present detailed performance evaluations to demonstrate the effectiveness of our algorithms w.r.t. both meeting both the QoS requirements and global information loss on real-world data sets.
Booktitle
Modeling and Performance of Privacy Preserving Authorization Mechanism for Graph Data
School
Electrical and Computer Engineering
Affiliation
Purdue University
Publication Date
2016-01-01