GROWL: Grid Resources on Workstation Library

GROWL is a JISC funded project involving Daresbury and Lancaster with University of Cambridge. The aim is to create a lightweight Grid toolkit which can be easily installed by end users. It makes use of Web services technology in C to access a range of services, such as job submission to NW-GRID.

Grid computing promises great things for Computational Scientists and Social Scientists. The reality is that accessing computational jobs or data on clusters to the Grid from a desktop environment involves overcoming a variety of obstacles.

As Grid technology matures and emerges from e-Science research to deliver production systems like the National Grid Service (NGS), the North West Grid (NW-GRID) and Campus Grid, its potential user-base needs to expand. Several recent surveys have identified a need to reach out to user communities less familiar with the traditional supercomputing and large data centres. Effort at NeSC, the National e-Science Center in Edinburgh, and NCeSS, the National Centre for e-Social Science with its supporting hub in Manchester are investigating how to do this. Some of the "new" user communities such as Bioinformatics and the Social Sciences rely on well-established research procedures and software solutions such as R, Stata, SPSS and Matlab. These applications do not easily fit into the typical Grid environment, since they run on the user's desktop computer rather than on a Grid resource. There is therefore a need for integrating "heritage" applications written in Fortran, C and C++ into the Grid in as seamless a way as possible to access the power of remote computing systems and access large datasets in a seamless way.

Our aims in this project were to encourage the uptake of Grid-based computing and distributed data management, focusing on the principal issues which could either hinder or facilitate application developmentto meet the requirements of end users. We refer to the difficulties identified as the "client problem" and suggested a solution to build upon the early prototype GROWL library to produce a truly lightweight extensible toolkit which complements other solutions.

From our initial requirements-gathering exercise and maybe even more so from our close inter-working with practising quantitative and computational scientists, it has become clear that user's needs vary considerably. Contributing factors include whether they are coming from a background where cluster or parallel computing is the norm (in which case many aspects of the Grid paradigm are fairly accessible) sych as in physics and chemistry. Other factors include which environments they are comfortable (or indeed uncomfortable) using, including the above-mentioned desktop applications but also bespoke GUI interfaces.

This led the development towards enhancing the original toolkit with a larger set of more simplistic GROWL Scripts for direct access to Grid resources and the development of the new GROWL server to support the SABRE-R project and making its functionality available as a demonstrator.

The GROWL Web site is at: http://www.growl.org.uk.

Final Report of the GROWL project: http://www.grids.ac.uk/GROWL/Doc/GROWL_FinalReport.pdf.

Objectives

Publications

J. Kewley, A.L. Braimah, R.J. Allan, M. Hayes, P. Brorsson, R. Crouchley, et al. GROWL Scripts: Lightweight Access to Grid Resources Proc. UK e-Science All Hands Meeting (2007)

GROWL (last edited 2009-02-12 16:30:14 by RobAllan)

This website maintained by Research Computing Services, University of Manchester