Archived: 30 January 2006
University of Adelaide Masters by coursework thesis, November 2004.
Supervisors: Andrew Wendelborn and Paul Coddington
The aim of grid computing is to provide an environment which allows easy, ubiquitous access to geographically distributed, heterogeneous computational resources. Grid computing can therefore provide a much wider range of computational resource than the single computer or super computer.
However, due to the complex concurrent environment presented by a grid and the uncontrollable, unpredictable nature of real grid environments, it is much harder to design and implement grid-based applications than programming on a single computer. Programming for grid-based applications requires programmers to have adequate knowledge and experience in grid technologies such as resource scheduling, data synchronization.
PAGIS is a high-level programming model for grid applications, which is based on the theory of process networks and uses Globus Toolkit. With PAGIS, a programmer can define grid application using an intuitive, high-level notation that uses simple rules of composition to express an application as a process network, comprising processing nodes connected by data channels, without understanding in detail the technology involved.
This project investigated a method to map a process network to a set of services interfaces, which can be described using the Web Services Definition Language (WSDL). How to implement these services and deploy the process network using service-based Globus Toolkit V3.2 in distributed environments are also discussed.