Whiteboards that compute: A workload analysis

A whiteboard that automatically identifies drawn strokes, interprets them in context, and augments drawn images with computational results, such as solutions to mathematical equations or results of circuit simulations, is a surprisingly realistic goal for systems architects. In this paper we describe the state of this emerging domain and argue that technical trends will make this a particularly attractive workload in the future. We provide a preliminary characterization of the critical loops that exist within one state of the art system, currently undergoing development, and we attempt to quantify the workloads that whiteboard-sized devices are likely to face in the future. While this work is by no means a typical workload characterization paper, given the shift in programming models that we are about to endure, it is now more important than ever before to identify and understand those applications that have the potential to drive our industry forward.