Simulation-based Assessment of Performance-Workload Tradeoffs for System Design Evaluation

Ideally, when selecting a system design or re-design, the selected option should balance workload and performance. However, previous research suggests that the relationship between workload and performance is nonlinear, thus achieving one, does not imply achieving the other. To ensure that a system will meet performance and workload goals, system designers should investigate the expected workload and performance tradeoffs of the potential system designs. This study uses discrete event simulation to demonstrate how to perform workload-performance tradeoffs when designing or re-designing a system, using a baseline system and three alternative designs. By examining the relationship between workload and performance, the study demonstrates how to identify the impacts that design alternatives can have on workload and performance. The study finds that four types of decision circumstances can occur: 1) a single preferred alternative is identified, 2) multiple preferred alternatives are identified, 3) no preferred alternative is identified, and 4) alternatives achieve conflicting goals. Each of these circumstances has unique implications for the design selection decision.