Perfect Hashed Skyline Analytics on Big Data

Abstract: Sky line is an important function in many programs to come back a set of exciting factors from a possibly large information area. Given a desk, the function discovers all tuples that are not covered with any other tuples. It is found that the current methods cannot process skyline on big information effectively. This document provides a novel skyline criterion SSPL on big information. SSPL uses categorized positional catalog details which require low area expense to decrease I/O cost considerably. The categorized positional catalog list Lj is designed for each feature Aj and is organized in climbing order of Aj . SSPL includes two stages. In stage 1, SSPL determines check out detail of the engaged categorized positional catalog details. During accessing the details in a round-robin style, SSPL works trimming on any applicant positional catalog to eliminate the applicant whose corresponding tuple is not skyline outcome. Phase1 finishes when there is an applicant positional catalog seen in all of the engaged details. In Phase 2, SSPL uses the acquired applicant positional indices to get skyline outcomes by a particular and successive check out on the desk. The trial outcomes on artificial and real information places show that SSPL has a important benefits over the current skyline methods.

Keywords: ZINC, SDC+, ZB-Tree, Skyline Computation.