GridIR GGF Charter

Administration

Grid Information Retrieval (GridIR)

Chairs

Kevin Gamiel

Organization	MCNC
Postal	3021 Cornwallis Road, Research Triangle Park, NC 27709-2889
Phone	+1-919-248-1915
Email	kgamiel at cnidr.org

Dr. Gregory B. Newby, Research Faculty

Organization	Arctic Region Supercomputing Center University of Alaska Fairbanks
Postal	910 Yukon Drive Suite 106 PO Box 756020 Fairbanks, AK 99775
Phone	+1-907-474-7160
Fax	+1-907-474-5494
Email	newby at arsc.edu

Secretary

Sousan Karimi

Organization	MCNC
Postal	3021 Cornwallis Road, Research Triangle Park, NC 27709-2889
Phone	+1-919-248-9297
Email	sousan at mcnc.org

Mailing List

gir-wg@gridforum.org

Subscription details available at: http://www.gridir.org/mailing_list.html.

Description and Objectives

The GridIR WG will establish a specific set of requirements, an architecture, and detailed specifications for a particular Information Retrieval (IR) system on the OGSA Grid. GridIR will provide document collection management, indexing/searching, and query processing services to OGSA grid users and applications.

Goals

The GridIR WG will:

1. Establish requirements for GridIR.

The WG will describe specific functional system requirements. Initial requirements are identified as:

• Distributed, asynchronous and independent of content and query type
• Collection management capability
• Indexing and searching capability
• Query processing capability
• Metadata capabilities to identify and interact with collections, indexes and query processors

The WG will describe a specific architecture in support of satisfying the IR system requirements in an efficient, modular manner. The initial architecture includes three major components:

• Collection Manager services - Provides generalized methods for describing logical record collections, instantiating those collections (by crawling, copying, accepting streams, etc), providing metadata services on those collections, provide access methods to those collections, and provide event services for collection change notification.
• Indexing/Searching services - Provides generalized methods for building indexes/representations of Collections for subsequent rapid search. Also provides generalized complex query and retrieval methods.
• Query Processing services - Provides methods for distributed searching, merging result sets, query expansion, and event-driven change notification.

The WG will describe a set of specifications based on the OGSA suitable to implement GridIR systems. We will describe a specific set of OGSA services, initially including:

• GridIR PortType
  • Includes method(s) for discovering metadata about available IR capabilities and collections
• CollectionManagement PortType
  • Includes methods for defining logical document collections (including document harvesting and transformation rules) and delivering them (as full collections or partial updates)
• SearchAndPresentation PortType
  • Includes methods for submitting a structured query and manipulating result sets.
• InformationRetrieval PortType
  • Inherits SearchAndPresentation PortType.
  • Adds methods for index administration.
  • Supplies core IR functionality.
• QueryProcessing PortType
  • Inherits SearchAndPresentation PortType.
  • Handles distributed asynchronous event-driven queries
  • Presents super-sets of InformationRetrieval PortTypes to user clients
  • Provides methods for result set merging
  • Provides query expansion and other processing services

Milestones

• GridIR Requirements Document - Stakeholder-driven list of functional IR system requirements. Also includes use cases, historical, and background material. Draft by GGF7, finalize by GGF8.
• GridIR Architecture Document - Describes overall system comprised of integrated grid services in support of the identified system requirements. First draft by GGF8, revisions by GGF9, finalize by GGF10.
• GridIR Specifications Document - Describes each OGSA grid service in detail. For each OGSA grid service identified, includes complete method API with inputs and outputs, query structures, and schema descriptions. First draft by GGF9, revisions by GGF10, finalize by GGF11.

Website

http://www.gridir.org