Essays in real-time distributed transaction processing and graphical decision support systems
Abstract
The present research was undertaken to study two diverse topics--analysis performance of a real time distributed database system and investigating the usefulness of three dimensional graphics techniques in understanding business data. In some database applications, transactions have economic value only if they can be completed within a stipulated time frame. The first part of the present study investigates the design considerations for such systems and studies the impact of some crucial policies on system performances. A scalable, general purpose simulation system was developed to study the behavior of distributed systems. This system was used to observe behavior of a real time distributed database system under different scheduling related policies. The investigation extended from a simple one node to a complex eight node system. The resulting behavior of such systems is related to the scheduling and proiritization policies, load on the system, concurrency control issues, network delays, volatility of transaction mix, and complexity of the system. The second part of the study investigates the applicability of three dimensional graphics with color and animation to understand financial data. These techniques are used to illustrate price movement of future stock options. The study generates images that describe the behavior of the Black-Scholes model, the binomial approximation and detect mispricing opportunity and the time frame within which each mispricing opportunity is valid. These results are a first step in validating the usefulness of 3D graphical animation as a business data analysis tool.
Degree
Ph.D.
Advisors
Koehler, Purdue University.
Subject Area
Management|Computer science
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.