"Seegrid will be due for a migration to confluence on the 1st of August. Any update on or after the 1st of August will NOT be migrated"

Grid Technology for Geosciences (APAC)


Related pages (edit)
Home
Overview
Project Scenarios
Project Outputs
System Specification
Issue Tracking
Project Journal
APAC Technical Discussion
Team Meetings
Development Toolkits
External Links
APAC computational infrastructure project
Gridsphere Portal toolkit
Globus toolkit

Aims

The aim of this project is to identify and establish the geoscience community specific interoperability “standards” for the APAC Grid network. This will be achieved by collaboration with the APAC grid infrastructure projects and test bed application development. The project will act as a connection point between the APAC Grid and the emerging geoscience communities being established under the Solid Earth and Environmental Sciences Grid (SEE Grid) and other members of the geoscience research community including Monash, University of Sydney, the pmd*CRC and ACcESS MNRF.

Description

The development strategy for this APAC Grid project draws on significant amounts of background intellectual property and related R&D occurring in the Australian geoscience research community. Significant portions of this community have been working together through the Solid Earth and Environmental Sciences Grid (SEE Grid), pmd*CRC and ACCESS MNRF. SEE Grid is of particular relevance since it represents a concerted effort by the community to form a geoscience Community of Practice for the development of open standards and interfaces for the provision of Grid and Web enabled services and applications. The SEE Grid community website hosts the community collaboration, standards, architectures, patterns and the ongoing development.

This Geosciences APAC Grid project seeks to ensure that services useful to the Geoscience community are provided via the APAC Grid network and partners by ensuring they support the appropriate open standards and interfaces. Several APAC partners and proposed APAC Grid projects include components that are useful to members of the geoscience community including:
  • Numerical codes designed for HPC and Grid computing – Snark, eScript, Finley, SPModel, Snac
  • Libraries for mesh generation
  • Management of large scale jobs and workflow engines
  • Federated data access
  • Large scale data storage
  • Visualisation
  • Grid middleware services – single sign on, security, remote access and so on.
  • Cross-domain interoperability – Connecting some geoscience information with other earth sciences (eg. marine) via service chaining functionality that converts compatible data on the fly.

For these components to be broadly used by the geoscience community it is not sufficient to simply make the “codes” available via open source or a common programming API. The components must exhibit conceptual integrity and have a common notion of semantics. The open standards and interfaces being developed under the SEE Grid community provide this conceptual integrity. What is required is that APAC Grid services be exposed via these standards and interfaces so that the geoscience community can make direct use of them. Exposure of a particular service via the geoscience open interfaces does not prevent the exposure using other interfaces. This issue is specifically addressed in the Conceptual, Mathematical, Numerical (CmnPattern) design pattern utilised in the SEE Grid computational services architecture.

The establishment of such a geoscience enabled APAC grid would allow the community to utilise APAC partners Grid services as a part of their research without fear of directly tying their projects to “other people’s codes”. Should an alternate code or line of research become available the core application components would require limited modification to “redirect” to the new services. The interfaces, problem description and information model would remain the same for the community regardless of who develops the underlying code. The approach differs from programming framework development in that applications are developed from loosely coupled services and workflow coordination rather than via programmatic API’s.

Taking a predominately computer science perspective, researchers in the geoscience community range from deeply technical “numerical component programmers”, to “application programmers”, through to “users of applications or information resources”. Note that all are undertaking geoscientific research. This range of potential APAC grid users is what drives the importance of Conceptual Integrity and the use of Grid Services. The Grid represents the distributed and cross-organisational development of a system rather than a single program. Conceptual integrity is a key requirement in the successful use of such a system.

Key issues to be addressed in supporting the geosciences research community include:
  • Requirements formalisation, documentation and the production of a Computational Services Roadmap for APAC Grid usage by the geosciences research community
  • Easing the burden of data access and sharing via the establishment of common notions for semantics in geoscience information representation and access on a distributed grid infrastructure. The Grid infrastructure will include but not be limited to that provided by the APAC partners since there are a number of government agency, academic and industry information repositories required by the geoscience community.
  • Establishing a common Grid Services enabled network for the:
  • testing of rigorously specified hypothesis via parallel job management across a range of different (and potentially competing) scientific models created by different members of the community – a single hypothesis description should be usable across a wide variety of possible “solvers”.
  • solution of mathematical problems relevant to the geoscience community using HPC facilities
  • coarse grain coupling of geoscience grid services (information and computation)
  • Co-design & development with APAC grid infrastructure and CT&T theme projects to reduce the burden associated with integrating geoscience applications and workflows with independently developed HPC based numerical solvers and other APAC Grid services
  • Test bed demonstration of the geoscience community’s access to the APAC Grid network.

This APAC geosciences project is directed at R&D required to enable APAC Grid services to use the emerging standards for information exchange (GML, XMML, GPML) and service design patterns utilised in the SEE Grid community. The project will undertake development, experimentation and verification of the approach via a test bed running on the APAC Grid network.

The key activities are:
  • developing patterns for design of end-user applications that are (optionally) able to take advantage of APAC Grid services of the types described previously
  • Demonstrating and profiling the use of the design patterns via a test bed so that the information can be used to optimise Grid resource utilitsation and workflow and encourage use of the APAC Grid services
  • To further develop open standards for geoscience conceptual posed problems and more general mathematical posed problems so that open interfaces to computational grid services can be established
  • To see greater utilisation of APAC Grid services for solid earth and environmental sciences research on a routine basis

The project is scoped with an initial 12 month duration directed at establishing the key architectural and information standards for geoscience enabled APAC Grid services and verifying its operation via two small test bed applications:
  1. Computational Services via Snark and Finley – initially a mantle convection problem but this is only a precursor to enabling broader geoscience simulation services
  2. EarthByte 4D Data Portal - EarthByte will allow users to seamlessly connect geological and geophysical observations coded by tectonic plate and geological time to simulation software and visualisation/multimedia tools

Both test bed applications will develop and utilise SEE Grid service standards and informational models as part of the SEE Grid community effort.

This effort will build directly upon the existing foundation of SEE Grid, pmd*CRC, CSIRO and other geoscience research community efforts and leverage’s a considerable amount of previous and ongoing R&D.

Throughout the project the standards for information exchange and computational services along with any open source components and reference implementations will be published on the Solid Earth and Environmental Sciences Grid (SEE Grid) community website. APAC partners and other national and international collaborators will then be able to contribute to and utilise the R&D outcomes.
Topic revision: r5 - 15 Oct 2010, UnknownUser
 

Current license: All material on this collaboration platform is licensed under a Creative Commons Attribution 3.0 Australia Licence (CC BY 3.0).