At nanometer technologies, variability has become one of the leading causes for chip failures and delayed schedules. For nanometer design implementation flows, variability is associated with design modes, power states, environmental conditions, manufacturing steps, and the behavior of devices and interconnects. Variability affects the entire physical design environment, from power management, through timing and signal integrity closure, to manufacturability.

This paper explores the different forms of variability, reviews the challenges involved in modeling variability, and discusses the requirements of an implementation system that comprehensively analyzes and optimizes a design for best results in the presence of variability.