In many applications involving multiple criteria optimal decision making, users may often want to make a personal trade-off among all optimal solutions for selecting one object that best fits their personal needs. As a key feature, skyline in a multi-dimensional space provides a minimal set of candidates for such purposes by removing every object that is not preferred by any (monotonic) utility/scoring function; that is, the skyline removes all objects not preferred by any user no matter how their preferences vary. Due to its importance, the problem of skyline computation and its variants have been extensively studied in the database literature. In this paper, we provide a comprehensive survey of skyline computation techniques. Specifically, we first introduce the skyline computation algorithms on traditional (exact) data where each object corresponds to a point in a multi-dimensional space. Then, we discuss the skyline models and effcient algorithms to handle uncertain data which is inherent in many important applications. Finally, we briefly describe a few variants of the skyline (e.g., skycube, k-skyband and reverse skyline) in this paper.