Synchronous DataFlow Graphs (SDF in short) is a simple model of computation introduced for the description of Digital Signal Processing Applications. This formalism is today widely used to model embedded parallel applications. This chapter aims at presenting a panorama of theoretical results and practical applications in connection with cyclic scheduling problems. We first recall that the execution of a SDF can be seen as a set of cyclic dependant tasks. The structure of precedence constraints, important dominance properties and simplifications of the SDF are then presented. For the special case of uniform precedence graph, periodic schedule are dominant and the maximum throughput can be polynomially evaluated. Main results on the resource constrained problem are presented, followed by a more recent problem issued from sensor networks. In the general case, the existence of a polynomial-time algorithm to evaluate the maximum throughput of a SDF is a challenging question. However, the determination of a periodic schedule of minimum period is a polynomial problem, and many authors limit their study to this class of schedule to express optimization problems as the total buffer minimization or to evaluate the latency of a real-time periodic system.