Doctoral Thesis Defense: Gaby Dagher

Speaker: Gaby Dagher

Supervisors: Drs. B. C. M. Fung, J. Clark

Examining Committee: Drs. T. Eavis, W. Hamou-Lhadj, S. Matwin, L. Wang, S. Li (Chair)

Title:  Secure Protocols for Privacy-preserving Data Outsourcing, Integration and Auditing

Date: Friday, December 18, 2015

Time: 14:00

Place: EV 1.162

ABSTRACT

As the amount of data available from a wide range of domains has increased tremendously in recent years, the demand for data sharing and integration has also risen. The cloud computing paradigm provides great flexibility to data owners with respect to computation and storage capabilities, which makes it a suitable platform for them to share their data. Outsourcing person-specific data to the cloud, however, imposes serious concerns about the confidentiality of the outsourced data, the privacy of the individuals referenced in the data, as well as the confidentiality of the queries processed over the data. Data integration is another form of data sharing, where data owners jointly perform the integration process, and the resulting dataset is shared between them. Integrating related data from different sources enables individuals, businesses, organizations and government agencies to perform better data analysis, make better informed decisions, and provide better services. Designing distributed, secure, and privacy-preserving protocols for integrating person-specific data, however, poses several challenges, including how to prevent each party from inferring sensitive information about individuals during the execution of the protocol, how to guarantee an effective level of privacy on the released data while maintaining utility for data mining, and how to support public auditing such that anyone at any time can verify that the integration was executed correctly and no participants deviated from the protocol.

In this thesis, we address the aforementioned concerns by presenting secure protocols for privacy-preserving data outsourcing, integration and auditing. First, we propose a secure cloud-based data outsourcing and query processing framework that simultaneously preserves the confidentiality of the data and the query requests, while providing differential privacy guarantees on the query results. Second, we propose a publicly verifiable protocol for integrating person-specific data from multiple data owners, while providing differential privacy guarantees and maintaining an effective level of utility on the released data for the purpose of data mining. Next, we propose a privacy-preserving multi-party protocol for high-dimensional data mashup with guaranteed LKC-privacy on the output data. Finally, we apply the theory to the real world problem of solvency in Bitcoin. More specifically, we propose a privacy-preserving and publicly verifiable cryptographic proof of solvency scheme for Bitcoin exchanges such that no information is revealed about the exchange's customer holdings, the value of the exchange's total holdings is kept secret, and multiple exchanges performing the same proof of solvency can contemporaneously prove they are not colluding.