The emerging trends in computing have increasingly had a network-centric focus. Networked services offered through cloud computing paradigms have replaced applications that would traditionally run on local machines. In addition, the growing usage of applications such as social networking and platforms such as smartphones has resulted in greater need for ubiquitous network access. The consequent heightened demand for networked computing warrants efficient utilization of the limited network resources and more intelligent resource control algorithms, with a focus on providing an enhanced user experience. This thesis examines quality-of-service aware resource control for both wireless and wired networks. The first part of the thesis focuses on smartphones, which have become the de-facto mobile computing platform. A smartphone typically has access to multiple types of wireless networks, such as cellular networks and WiFi. Moreover, the functionality of smartphones can be expanded by installing applications. These two core characteristics of smartphones also reveal their most significant limitations: lower available bandwidth and limited computing power. Both of these limitations are addressed in this thesis. Available bandwidth on wireless networks fluctuates over time and is also shared among all users connecting to the same base station. In this work we present a dynamic bandwidth prediction model that makes short term predictions on the evolution of bandwidth. The model is dynamic and adjusts to the latest measurements provided by Zeus, a bandwidth measurement tool we designed and implemented on Nokia phones. The bandwidth predictions of our model are utilized in a novel rate control scheme, which we demonstrate to offer better performance than existing schemes. We next investigate the computing limitations of smartphones. A novel framework is considered, where computational tasks may be transferred to a central server and the results are fetched back at a later time. The central server has ample computing resources compared to the smartphones and the computing speedup outweighs the communication delays. The goal is to minimize the latency experienced by computational tasks, while judiciously utilizing the scarce memory resources available at the smartphone. Given the fluctuating nature of wireless bandwidth, there is a tradeoff between limited connectivity and congestion at the mobile. The second part of this thesis investigates resource control issues in wireline computing and more specifically in packet switches. Packet switches are essential parts of the Internet backbone and are also present in every data center. Modern data centers are severely constrained by their power consumption and power saving schemes would enable their further expansion. We propose novel power-aware scheduling algorithms for switches that offer significant power savings while sacrificing minimal performance. Finally, we examine a novel, scalable two-stage ingress memory switch architecture and we add backlog awareness to the scheduling algorithm to improve performance and fairness as perceived by the user.