The Internet of Things (IoT) represents a new class of applications that can benefit from cloud infrastructure. However, directly connecting smart devices to the cloud has multiple disadvantages and is unlikely to keep up with the growing speed of the IoT or the diverse needs of IoT applications. Here, the authors argue that fundamental IoT properties prevent the current approach from scaling. What's missing is a well-architected system extending cloud functionality and providing seamless interplay among heterogeneous components closer to the edge in the IoT space. Raising the level of abstraction to a data-centric design-focused around the distribution, preservation, and protection of information-better matches the IoT. To address such problems with the cloud-centric architecture, the authors present their early work on a distributed platform, the Global Data Plane.