Curriculum Vitae - The University of Texas at Dallas

8. Design and Implementation of a Database Inference Controller, December 1993, Data andKnowledge Engineering Journal (North Holland), Vol. 11, No. 3, p. 271 - 297 (co-authors: W.Ford, M. Collins, J. O'Keeffe); (Article reprinted by the MITRE Journal, 1994). Lead AuthorMy Contribution: I designed the entire system. My colleagues Mr. Ford and Mr. O’Keeffeimplemented the query processor. My colleague Ms. Collins implemented the update processor. Iwrote the entire paper from the technical reports.9. Security Constraint Processing in a Multilevel Secure Distributed Database Management System,IEEE Transactions on Knowledge and Data Engineering, April 1995 (coauthor: W. Ford). LeadAuthorMy Contribution: I designed the entire system. My colleague Mr. Ford implemented the system.I wrote the entire paper from the technical reports.10. The Use of Conceptual Structures to Handle the Inference Problem, November 1991,Proceedings of the 5th IFIP WG 11.3 Conference on Database Security, p. 333-362,Shepherdstown, VA. (Also published by North Holland, 1992, Database Security V: Status andProspects, p. 333 - 362). Sole AuthorMy Contribution: This is 100% my work.11. A Nonmonotonic Typed Multilevel Logic for Multilevel Database Management Systems, June1991, Proceedings of the 4th IEEE Computer Security Foundations Workshop, Franconia, NH.My Contribution: This is 100% my work.17.3 DEPENDABLE DATA MANAGEMENT SYSTEMSI began my research on dependable data management systems around 1992 and this research stillcontinues. I describe seven of my papers on this topic.Evolvable Real-time Systems: My research in the early to mid 1990s was on building an object-orienteddata manager and infrastructure for next generation command and control systems. At that time the legacysystems were hardcoded and running on mainframes. The goal was to develop flexible systems that couldaccommodate changes. There were no commercial systems available at that time. Data was arriving at avery rapid speed. This data had to be captured, stored, analyzed and decisions had to be made. Togetherwith my colleagues at MITRE, we were the first to develop an object-based, real-time infrastructure anddata manager. The infrastructure consisted of several services including inter-process communication,memory management and scheduling. The real-time data manager was a main memory data manager andwe designed a real-time priority ceiling protocol for transaction processing. The results are documented inPaper #1 (IEEE WORDS 1996). Then we integrated the infrastructure and data manager with the multisensorfusion applications and carried out an integrated design and implementation. This integration workis discussed in Paper #2 (IEEE WORDS 1997). The research was demonstrated to the Air Force and thetechnology was transferred to the AWACS program and subsequently Boeing and Lockheed took many ofthe ideas to implement into the operational systems.Real-time Transaction Processing: One of the challenges in designing a real-time data manager is todesign transaction processing algorithms that meet timing constraints. There were a number of algorithmsdeveloped that had to meet the requirements of AWACS. After examining and evaluating severalalgorithms, we felt that the priority ceiling algorithm would be most suited. However, the priority ceilingalgorithm had to be adapted to meet the different criteria. I presented the initial design to the team inMarch 1995. Then I collaborated with a professor at the University of Rhode Island and co-supervisedstudents who carried out the detailed design and implementation of the algorithm in 1996 - 1997. Thisalgorithm is presented in Paper #3 (Real-time Systems Journal).Adaptive Real-time Systems: In this paper, we introduced some novel ideas into the infrastructuredesign. While in the earlier papers we considered only non-adaptive protocols, in our research in the late104