A framework for multidimensional skyline queries over streaming data

Abstract

Skyline query has attracted a great deal of interest during last years because of its ability to help decision makers when multi-criteria objectives are to be handled. Several authors have pointed the interest of multidimensional skylines, i.e., the set of criteria become a parameter of the query. In order to efficiently evaluate these queries, index structures have been proposed. In this paper, we address the problem of efficiently handling multidimensional skyline queries in the context of streaming data. The appended records have a validity time interval after which they become outdated and hence, can be discarded. To that end, we propose a framework that handles an index structure periodically updated. Then the queries consider just the indexed data. This is the price we pay to deal with the streaming nature of the data we consider.Through extensive experiments, we demonstrate our framework’s ability to handle multidimensional skyline queries with challenging streaming data. The main criteria we consider to assess the performance of our solution are query execution time and both index structure maintenance time and its memory consumption.