Authors: Norah Saleh Alghamdi; Wenny Rahayu; Eric Pardede
Addresses: Department of Computer Science and Computer Engineering, La Trobe University, Melbourne, VIC 3083, Australia. ' Department of Computer Science and Computer Engineering, La Trobe University, Melbourne, VIC 3083, Australia. ' Department of Computer Science and Computer Engineering, La Trobe University, Melbourne, VIC 3083, Australia
Abstract: Due to the rapid growth of XML representation for information exchange, XML databases have been widely adopted in a variety of applications. This paper presents two layers of optimisation for dealing with large XML databases: (1) OXDP (Object-Based Methodology for XML Data Partitioning) which has been developed to partition XML data efficiently and (2) OXiP (Object-Based XML Indexing for Partitions) which is an indexing and linking mechanism for partitioned data. OXDP provides optimal XML data partitioning based on an object's semantic features which improves XML data query performance. The OXiP method tokenises all rooted label paths and preserves the pathways within each XML object partition. The semantic-based data partition ultimately enhances the notion of a frequently accessed data subset which is an advantageous feature in our proposed methods to decrease the time to answer queries. Experimentally, OXDP and OXiP can achieve an order of magnitude performance improvement for querying XML data.
Keywords: large databases; XML databases; semantic query processing; indexing; optimisation; path query; object-based methodology; XML data partitioning.
International Journal of Grid and Utility Computing, 2012 Vol.3 No.2/3, pp.112 - 125
Received: 04 Jul 2011
Accepted: 11 Sep 2011
Published online: 20 Dec 2014 *