This plenary will introduce a class of control problems where the feedback loop is closed over unreliable communication links. The motivation for such problems comes from growing industrial applications that demand remote control of objects over Internet-type or wireless networks where links are prone to failure. Depending on the availability of acknowledgment (ACK) signals, two different types of networking protocols can be considered in these problems: TCP (Transmission Control Protocol) and UDP(User Datagram Protocol). Under the TCP structure, ACK signals are transmitted, whereas in the UDP structure no ACK packets are present. One objective in these control problems is to mean-square stabilize the system while minimizing a performance criterion, when the information flow between the controller and the plant is disrupted due to link failures or packet losses. The talk will present necessary and sufficient conditions for mean-square stabilizability of LTI (linear time-invariant) systems in such configurations, as well as sufficient conditions for the existence of stabilizing optimal controllers. It will also elaborate on a characterization of such controllers. Finally, extensions to other types of network-control problems will be discussed, such as the class of problems where there are restrictions on the frequency of usage of communication links.