Bandwidth-constrained distributed skyline computation

• Authors: Akrivi Vlachou;Kjetil Nørvåg
• Venue: Proceedings of the Eighth ACM International Workshop on Data Engineering for Wireless and Mobile Access
• Year: 2009

Skyline queries have been studied in centralized systems, but also in distributed environments, such as web information systems and peer-to-peer networks. Most of the existing work focuses on efficient processing of skyline queries that return the exact and complete result set. In this paper, we study skyline computation in a distributed environment under the assumption of a given upper bound on the bandwidth consumption. Supporting such a constraint is very important in a mobile environment, where data transmissions may lead to deterioration of query processing performance, while imposing high cost for the mobile device in terms of energy consumption. Therefore, the cost of transferring all data points that may contribute to the skyline result set to the querying server is often prohibitive. Our target is, given an upper bound of bandwidth consumption, to maximize the quality of the retrieved skyline approximation. For this purpose, we propose a framework for bandwidth-constrained skyline query processing, based on the intentional selection of a limited amount of data points from each participating mobile device. We propose a novel method to carefully select the most promising data points based on subspace dominations and analyze its properties. In our extensive experimental evaluation, we demonstrate that result sets of high quality are achieved, while reducing the communication cost considerably.