The functionality of future devices and services will require ever improving network speed and capacity. Whether new innovations will be available to all, or just a privileged few, hinges on the ability of communication networks to scale with the growth of demand. However, our current understanding of how to design and manage networks is ill-suited to respond to changing traffic trends, such as more prevalent inter-user data and real-time communications. The result is grossly inefficient network resource allocation decisions that threaten long-term network scalability. To prevent shortages of network resources in the future, we must eliminate these inefficiencies and proactively evolve communication networks to new types and growing volumes of Internet traffic. The research work presented in this dissertation makes the case for several principles that lead scalable network design. Most importantly, I argue that scalable network designed must not only match growing network demands in magnitude, but also provide the right kind of connectivity with respect to the continual evolution of network traffic patterns.