Skyline and reverse skyline query processing in SpatialHadoop

Abstract In this paper, we study the problem of skyline and reverse skyline computation using SpatialHadoop, an extension of Hadoop that enhances its capabilities with spatial awareness. The exploitation of spatial indexing structures and the spatial properties of data can exploit MapReduce-based methods by reducing the reading, writing, computational and communicational overhead. Through our study, we propose two methods for skyline and reverse skyline computation, which operates in the spatial aware environment that SpatialHadoop provides. This environment allows for performing filtering on the initial dataset to retrieve an answer efficiently by using existing state-of-the-art indexing approaches. The proposed algorithms make use of the full capabilities of the indexing mechanisms provided by the SpatialHadoop and have been tested against large-scale datasets including a real-life, large-scale OpenStreetMap dataset. To the best of our knowledge, this is the first work that studies reverse skyline over SpatialHadoop.