MEADOW: a middleware for efficient databases through openGIS wrappers

  • Authors:

  • Sang K. Cha;Kihong Kim;Byung S. Lee;Changbin Song;Sangyong Hwang;Yongsik Kwon

  • Affiliations:

  • Graduate School of Electrical Engineering and Computer Science, Seoul National University, Seoul, Korea;Graduate School of Electrical Engineering and Computer Science, Seoul National University, Seoul, Korea;Department of Computer Science, University of Vermont, Burlington;Graduate School of Electrical Engineering and Computer Science, Seoul National University, Seoul, Korea;Graduate School of Electrical Engineering and Computer Science, Seoul National University, Seoul, Korea;Graduate School of Electrical Engineering and Computer Science, Seoul National University, Seoul, Korea

  • Venue:
  • Software—Practice & Experience
  • Year:
  • 2002

Quantified Score

Hi-index 0.01

Visualization

Abstract

With the proliferation of various geographic databases on the Internet, we have seen increasing needs for accessing them concurrently and remotely via the Web for high-level decision making. In this paper, we present Middleware for Efficient Access to Databases through OpenGIS Wrappers (MEADOW), an object-oriented middleware system we have developed to meet these needs. Current OpenGIS standard addresses many interoperability issues involved in such a global utilization of geographic databases. However, existing Simple Feature specification for CORBA (SFCORBA) implementations of OpenGIS proved to be insufficient for MEADOW. The main problems are the complexity of system development and maintenance, and the inefficiency of accessing remote data servers for processing region queries. We resolved the complexity problem by automatically generating a major portion of the application code, specifically wrappers on database servers and client library modules called transparent access providers. A MEADOW view definition language was developed as a high-level specification language for this purpose. The efficiency problem was resolved by using a region-based group prefetching of spatial objects from a geographic region. In addition, we implemented an OID-based semijoin for efficient global query processing, and a region-level locking to enhance the level of concurrency among region queries.