P2P-Based Query Processing over Linked Data

• Published in: Linked Data Management pp. 367 - 396
• Format: Book Chapter
• Language: English
• Published: United States Chapman and Hall/CRC 2014; CRC Press LLC
• Subjects: Internet guides & online services; Subject Object Joins; Range Queries; Query Plan; Logical Operators; RDF Data; Query Execution; Early Synchronization; Triple Sets; Extract Triples; Query Initiator; P2P Systems; SPARQL Query; Limited Local Knowledge; Routing Graphs; Triple Pattern; Blank Nodes; Linked Data; Query Processing; Hash Key; Query Processing Capabilities; Query Algebra; Query Operators; Efficient Query Processing; Solution Mappings
• ISBN: 9781466582408; 1466582405
• DOI: 10.1201/b16859-23

Digital Enterprise Research Institute, National University of Ireland, GalwayKai-Uwe SattlerIlmenau University of Technology, GermanyManfred HauswirthDigital Enterprise Research Institute, National University of Ireland, GalwayThe Linked Data paradigm and many other data integration efforts, e.g., in e-science and e-government, are examples of public data management where data of public interest are maintained and accessed by a community. In a community setting, it would be preferable that the costs of the infrastructure would be shared as well. So instead of again building data stovepipes with their “proprietary” structures and interfaces and varying levels of query processing capabilities which require significant efforts for access and integration at the technical level, these systems could benefit from a “the world as a database” perspective, where all data irrespective of their internal structures and functionalities could be accessed and combined in a uniform way. In turn, many users would certainly like to share their own knowledge/data in this way. For instance, in astronomy, more than 200 scientific publications have been produced which rely mainly on the ability of the Sloan Digital SkyServer to answer database queries [515], which clearly demonstrates the need for an open Linked Data solution. In this context, RDF, as one of the main pillars of Linked Data, helps to address the challenges around data organization. As it does not rely on global schemata, is self-descriptive and supports “pay-as-you-go integration” [373], RDF helps to deal with heterogeneity and offers great genericity and flexibility. In addition, semantics for informationdriven integration (e.g., schema correspondences) can be handled as standard, first-class-citizen data.