CCE Theses and Dissertations
Date of Award
2012
Document Type
Dissertation
Degree Name
Doctor of Philosophy in Information Systems (DISS)
Department
Graduate School of Computer and Information Sciences
Advisor
Junping Sun
Committee Member
Marlyn K. Littman
Committee Member
James Cannady
Keywords
Benchmarking, Fine-Grained Access Control, Information Security, Performance, Privacy Preservation, Relational Databases
Abstract
Fine-grained access control is a conceptual approach to addressing database security requirements. In relational database management systems, fine-grained access control refers to access restrictions enforced at the row, column, or cell level. While a number of commercial implementations of database fine-grained access control are available, there are presently no generalized approaches to implementing fine-grained access control for relational database management systems.
Fine-grained access control is potentially a good solution for database professionals and system architects charged with designing database applications that implement granular security or privacy protection features. However, in the oral tradition of the database community, fine-grained access control is spoken of as imposing significant performance penalties, and is therefore best avoided. Regardless, there are current and emerging social, legal, and economic forces that mandate the need for efficient fine-grained access control in relational database management systems.
In the study undertaken, the author was able to quantify the performance costs associated with four common implementations of fine-grained access control for relational database management systems. Security benchmarking was employed as the methodology to quantify performance costs. Synthetic data from the TPC-W benchmark as well as representative data from a real-world application were utilized in the benchmarking process.
A simple graph-base performance model for Fine-grained Access Control Evaluation (FACE) was developed from benchmark data collected during the study. The FACE model is intended for use in predicting throughput and response times for relational database management systems that implement fine-grained access control using one of the common fine-grained access control mechanisms - authorization views, the Hippocratic Database, label-based access control, and transparent query rewrite. The author also addresses the issue of scalability for fine-grained access control mechanisms that were evaluated in the study.
NSUWorks Citation
David Harold Kumka. 2012. Quantifying Performance Costs of Database Fine-Grained Access Control. Doctoral dissertation. Nova Southeastern University. Retrieved from NSUWorks, Graduate School of Computer and Information Sciences. (204)
https://nsuworks.nova.edu/gscis_etd/204.