As commercial printing presses become faster, cheaper and more efficient, so too must the RIPs that process and feed them data to print. Previously [1], we presented a single machine GPU-accelerated RIP system that harnesses the massive parallel computational power of programmable Graphic Processing Units (GPUs). This project builds on this work and leverages the Apache Hadoop framework to construct a distributed RIP system designed to scale efficiently and meet current and future RIP requirements for high speed commercial digital presses. Hadoop is an open source package already deployed in mission critical application at major internet companies designed for massively distributed processing of large static data sets. It provides job management, load balancing, and error recovery services which are well suited to building a stable and comprehensive distributed RIP system. We found, however, that the Hadoop processing and data management models do not match the streaming media processing nature of an EPID (“Every Page is Different”[2]) printing infrastructure. We have therefore implemented custom processing and data management modules to support distributed ripping of jobs and dynamic streaming of PDFs, images and ripped jobs through the system. Lastly, a custom front end was developed which allowed the system to be integrated with a digital press. The result is a solution combining open source software and a cluster of GPU-equipped commodity PC workstations which significantly reduces the cost and energy consumption of the DFE while scaling to meet the DFE needs of a wide range of commercial digital presses.