Decentralized and stochastic systems with information constraints among system components are becoming increasingly common. Optimal design of such systems entails an interdisciplinary analysis of control, information and probability theories. In this talk, we review some recent contributions on the stabilization and optimization of such systems. We first consider stabilization of linear systems controlled over communication channels, where tight necessary and sufficient conditions on information channels for stochastic stabilizability of such systems are obtained. The stability notions include ergodicity, asymptotic mean stationarity and the existence of finite moments. We then consider the optimization problem, where structural and topological properties of information structures in networked control are studied and optimal coding and control policies are established for a large class of such systems. Some future research directions will also be presented.