A Web Service Model for Climate Data Access on the Grid

  • Authors:
  • Andrew Woolf;Keith Haines;Chunlei Liu

  • Affiliations:
  • -;-;ENVIRONMENTAL SYSTEMS SCIENCE CENTRE, UNIVERSITY OF READING, RG6 6AL, UK

  • Venue:
  • International Journal of High Performance Computing Applications
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

The problem of sharing environmental and climate data (from measurements or models) across networks does not yet have a standard solution. Ad hoc approaches are common, with complications arising through a number of different file formats and conventions. The state of the art in the climate research community is the Distributed Oceanographic Data System (DODS, also "OPeNDAP"), which maps a file or aggregation of files onto a URL. Data subsets may be retrieved, and limited (and non-standard) metadata queries made. Data are abstracted from storage only to a limited degree. We present an alternate data access mechanism, the Grid Access Data Service (GADS), better suited to Grid applications. Requirements of data abstracted from storage, rich metadata models, flexible delivery, security, and orchestrated workflows all suggest a Web Service solution. The GADS Web Service has two operations: a querying mechanism, and a request mechanism. Three metadata models are required: one for the abstract representation of climate data, one for characterizing data usage, and one for describing data storage artifacts. A compatibility interface has been layered on the Web Service to provide DODS/OPeNDAP functionality. A visualization Web portal has also been built to interface with GADS to demonstrate extendable functionality.