Machine vision in the cloud is a strong fit when applications require high processing power, when computing demand varies over time, or when the use case is not real-time critical. We recommend cloud deployment when image sources are distributed but processing should be centralized, when the hardware platform is virtualized, or when an organization is actively moving toward cloud-based infrastructure.