Skip to content
Skip to navigation menu

Objectives

The computational Grid presents many challenges to application scientists. Amongst these will be the proliferation of “Grid Services" that will become available through the Open Grid Services Architecture (OGSA). The service abstraction provides a useful method to model a diverse range of possible Grid resources ranging from computational servers, data repositories, software libraries and scientific instruments. One challenge likely to face the scientist is to discover and optimise their use of these Grid services to meet their own individual requirements. It is important to note that services can include existing software libraries, and locating and utilising these effectively is an important aspect of this project. Existing emphasis has been on hardware resources, primarily, and has therefore been of restricted benefit.

The general development theme adopted in OGSA is important to enable a diverse range of resources (both hardware and software) to be more effectively deployed on the Grid. However, there are issues with extensibility and little emphasis has, so far, been placed on supporting composition of basic services to develop applications, or on enabling simple service discovery beyond that supported directly through WSDL/UDDI. Compound services must also be defined and their behaviour, in terms of inputs, outputs and events, described and recorded.

This project therefore focuses on two aspects:

  1. The development of combinations of application-level services in a complex scientific procedure using distributed resources. A problem-solving environment will be used to automatically define service relationships and dependencies. It is expected that this environment will interoperate with and coordinate services from an OGSA compatible Grid Service infrastructure;

  2. The specification and implementation of Grid Services to support logging and workflow optimisation.