Abstract Cognitive radio has received considerable interest as a communication paradigm to provide dynamic spectrum allocation. However, there are still no realtime, at-scale validations of the key concepts in cognitive wireless. In this chapter we present the Wireless Open-Access Research Platform for Networks (WARPnet), a platform aimed at making cognitive radio a practical reality. The platform is designed from ground-up to be deployable and provides all the functionality needed to control and observe large-scale networks while enabling iteration on low-level algorithms. In addition, several design flows have been developed for isolated as well as integrated experiments. The platform supports both traditional research and cognitive protocol research, and thus provides uniform hardware and software frameworks to compare traditional approaches to the new clean-slate designs. During the course of the chapter, we will identify the difference between cognitive radio and traditional wireless networks to motivate the various features of the platform itself. We will also explore platform architecture to enable quick prototyping of next generation algorithms. Finally, we shall showcase WARPnet with an example cognitive protocol which exercises every part of the platform.