SkyMap: A trie-based index structure for high-performance skyline query processing

Abstract

Skyline queries have become commonplace in many applications. The main problem is to efficiently find the set of Pareto-optimal choices from a large amount of database items. Several algorithms and indexing techniques have been proposed recently, but until now no indexing technique was able to address all problems for skyline queries in realistic applications: fast access, superior scalability even for higher dimensions, and low costs for maintenance in face of data updates. In this paper we design and evaluate a trie-based indexing technique that solves the major efficiency bottlenecks of skyline queries. It scales gracefully even for high dimensional queries, is largely independent of the underlying data distributions, and allows for efficient updates. Our experiments on real and synthetic datasets show a performance increase of up to two orders of magnitude compared to previous indexing techniques. © 2011 Springer-Verlag Berlin Heidelberg.