A novel Decision Support System for modern Relational and Non-Relational Database Systems

Abstract

Data-based application design and development can be complex and difficult to navigate. With relational and non-relational database technologies, each with divergent paradigms, it is not apparent how to make the correct technology and design choices. Cloud-based environments add additional complexity to the decision-making process. Traditional methodologies and paradigms often neglect development requirements that must be derived from higher-level operational and maintenance requirements. A need for a guide or strategy that can direct a designer according to a defined process became evident. Procedures, methods, and techniques were further needed to assist the designer in making complex choices. This thesis follows the Quality Research Management method in concert with the Design Science Research paradigm to synthesise a pragmatic Decision Support System in the form of a process model from literature, experimentation, subject matter expertise, and best practice methods. Empirical results showed that the performance difference between the competing database technologies evaluated in this research is not the determining factor. The correct design procedure, choices, and methods in the preliminary design phase are critical performance determinants. The artefact developed in this research is a database development process model that includes a process based on the systems engineering development process, with associated procedures and methods that support decision-making. The process model was applied to the well-known TPC-C test case to validate its effectiveness and its pragmatic value. The Decision Support System, with the novelty of being mostly technology agnostic, improved database performance by up to 20% relative to traditional methods on the TPC-C test case and provided evidence that it has effectively addressed the initial research problem.