We present some fundamental communication limits of wireless networks under two different system architectures. Design guidelines arising from the analysis are then briefly discussed in both cases. First, we consider a wireless "ad-hoc" network in which nodes communicate without a pre-existing infrastructure. We present an outer bound to the classic throughput scaling result of Gupta and Kumar using electromagnetic theory techniques and without relying on specific stochastic models for the propagation channel. Our results are information-theoretic, but also take into account fundamental limitations due to the physics of wave propagation. We also discuss the relationship between our physical outer bounds and achievability results based on hierarchical cooperation, interference alignment, and specific stochastic fading and path loss models. Second, we consider a wireless random-access system where multiple nodes transmit to a common access point in an uncoordinated fashion. The main contribution in this setting is an information-theoretic formulation of the problem, which allow us to characterize the maximum achievable throughput for the collision channel and the AWGN channel. An optimum distributed algorithm for rate allocation at the transmitters is provided.