next up previous contents
Next: Contents   Contents

\includegraphics{/home/rja/TechReports/JISCcolour15.eps}

SAKAI EVALUATION EXERCISE
(A Report to JISC).

Rob Crouchley and Adrian Fish

e-Science Centre of Excellence, University of Lancaster

December 31st, 2004

Abstract:

This report summarises an evaluation of various aspects of the CHEF, Sakai and other related projects that may play a significant role in establishing a Virtual Research Environment (VRE) for the UK.

The work carried out built on e-Science activities at Lancaster and Daresbury, which include software development and hardware deployment for collaborators in the physical sciences, substantive e-Social Science research and social science training and awareness-raising. It built upon the existing collaborations that the project team have with each other and with other university groups in the UK. It is now using existing mechanisms to disseminate the outcomes of the work. Throughout the period of this evaluation we have received clear expressions of interest in working with us from the developers of Sakai (http://www.sakaiproject.org). This included discussions and presentations at the Sakai Developers Workshop, Denver, USA, 23-27/6/04 and subsequently at the JISC-CETIS workshop, Oxford 4-5/11/04.

The evaluation exercise was greatly facilitated by Lancaster deploying a CHEF implementation (http://redress.lancs.ac.uk) from day one and later the Sakai RC2 implementation to use as a collaboration tool for developers based at Lancaster and Daresbury.

This report summarises the outcomes and contains the conclusions of the evaluation which covered:

  1. Comparing Sakai/ CHEF with Alternative Frameworks for VREs;
  2. Establishing the ease of administration (EoA) of Sakai/ CHEF for a VRE;
  3. Establishing the feasibility of making existing VRE (Grid) components available via Sakai/ CHEF;
  4. Establishing the feasibility of extending the functionality of Sakai/ CHEF particularly to use Web Services for distributed development and deployment.

This report is available on line at http://redress.lancs.ac.uk:8080/portal under ``SEE/Resources'' and http://www.grids.ac.uk/Sakai/sakai_doc.pdf [PDF] or http://www.grids.ac.uk/Sakai/sakai_doc [HTML].

Executive Summary

This is a report of the JISC-funded Sakai Evaluation Exercise and is organised as follows. Firstly there are sections giving the background to portal and portlet technology and introducing the requirements and expectations for a VRE. The conclusions of our evaluation work are also explained and a way forward in using Sakai to address the needs of VRE developers is presented along with some architectural suggestions.

$\bullet$ Introduction - background to this report.
$\bullet$ Portals and Portlets 2003 - summary of the Portals and Portlets 2003 workshop held at NeSC 14-17th July 2003.
$\bullet$ Sakai SEPP Developers' Conference - trip report from the first SEPP conference.
$\bullet$ Role of Portals in a Virtual Research Environment - background material on VREs and the Open Service Architecture.
$\bullet$ Sakai Evaluation including CHEF, OGCE and GridSphere - capabilities of the chosen frameworks and summary of technical evaluations.
$\bullet$ Additional Comments on Portal Architectures.
$\bullet$ Conclusions.
$\bullet$ Acknowledgements.

Secondly there are appendices which give more detail of the work carried out and further background information.

$\bullet$ WP 1 - Comparing Sakai/ CHEF with Alternative Frameworks for a VRE.
$\bullet$ WP 2 - Ease of Administration (EoA) of Sakai/ CHEF for a VRE.
$\bullet$ WP 3 - Making existing VRE (Grid) Components available in Sakai/ CHEF.
$\bullet$ WP 4 - Issues involved in extending the Functionality of Sakai/ CHEF.
$\bullet$ Integration of UK VRE Tools into Sakai.
$\bullet$ Resources.
$\bullet$ Service Capability Set.
$\bullet$ Generic Portal Engines.
$\bullet$ Another independent Technical Report.

The internal Sakai architecture is still in flux, which means that we have carried out less tool porting than originally planned, we mainly used CHEF instead. Nevertheless the user interface is stable and was liked by the groups to whom we have given demonstrations. We were able to modify both CHEF and Sakai RC1/ RC2 and Sakai v1.0 to use a PostgreSQL database and were able to port the ReDRESS portal content from CHEF into Sakai RC1/ RC2 and Sakai v1.0. Several other tools have been ported as explained in the appendices.

Our conclusions in brief are as follows:

  1. The administration and content management systems in CHEF/ Sakai are more comprehensive than other frameworks such as GridSite as is the suite of tools collaboration available.
  2. CHEF is widely used for both Managed Learning (initially at University of Michigan, but taken up in other US institutions) and for Grid projects (e.g. OGCE, NEESGrid, alliance Portal, BioGrid, CMCS, TeraGrid, DOE Fusion, LEAD);
  3. CHEF and its tools will evolve into Sakai;
  4. Based on evaluation-tree metholodology using criteria reported in the concluding section, CHEF/ Sakai scored highest of the open-source portal content management frameworks tested.

It is our opinion that these conclusions make Sakai a clear potental winner for both VRE and VLE operations. A fully objective analysis of our evaluation is provided in Section 6.




next up previous contents
Next: Contents   Contents
Rob Allan 2005-05-09