Operator performance at network scheduling with dynamic pricing and limited capacities

As organizations become increasingly reliant on their network infrastructures for competitive advantage, it becomes apparent that network policies must be dynamically adapted to reflect organizational goals. Human operators often prove to be better suited to handling the ambiguity and unquantifiable phenomenon present in this sort of dynamic environment. Thus, this paper examines the effect of user interface design on an operator's ability to effectively schedule network jobs. Specifically, the relative merit of the Mind's Eye Planner (MEP) design methodology is examined through experimentation. The MEP methodology employs analogies and an ecological epistemology to construct dynamic GUIs. In this experiment a WAN simulation with dynamic pricing and limited link capacities was developed to compare subject performance using MEP and conventional interfaces. Subjects were measured on their ability to schedule jobs on time, on budget, and to maximize throughput. In each of these categories, subjects showed superior performance when using the MEP interface.