Our primary aim is to evaluate the appropriateness of the Sakai/ CHEF frameworks for the development of Virtual Research Environments (VREs) within the UK education, digital information and research communities. This report should help JISC to formulate the long-term requirements of VREs. The relationship between, and capabilities of, the CHEF/ OGCE and Sakai frameworks is explained in the full report. Another JSR-168 compliant framework, GridSphere, is also described for completeness as it has a certain popularity for Grid projects in Europe.
A broad definition of a VRE was provided following some debate by the JISC JCSR VRE Working Group . It encompasses some, but not all functionality of:
CHEF, the underlying technology which was the initial focus of this review, is an open source, freely available framework designed to provide useful collaboration tools, such as a chat facility, discussion boards, shared calendars and file sharing in the context of a course or tutorial 'worksite'. It currently provides analogues of the most popular features of commercial groupware offerings such as IBM's Lotus Notes. The Sakai project aims to bring CHEF and uPortal together and then add more VLE facilities. It is in fact designed to encourage open collaboration and the sharing and contribution of new tools by a large and growing community of users, and this appears to be appropriate to the needs of the UK research community. After some discussion it was felt appropriate to include GridSphere in the evaluation of open source tools as it is being used in some UK e-Science projects. Bodington was included for completeness because of its prominence as a UK Virtual Learning Environment.
Jetspeed and uPortal are open source, freely available portal servers that allow portal content to be imported, or linked to, by an administrator and then flexibly configured by users, thus empowering the users in the development of their own collaborative environment. They are hosting containers supporting portlet standards, but do not offer higher-level content management or tools and therefore are not included in the final assessment. Other generic portal engines are listed in an Appendix.
Virtual Research Environments, VREs, by their very nature will continue to evolve. It is important to make sure that our existing resources, services and applications can be made accessible in emerging standards-compliant frameworks and can be supported in the long term. A part of the funding requested for this evaluation was for a UK subscription as early adopters to the Sakai educational Partners Programme, SEPP. There are obviously good financial reasons for moving to an open source platform for VRE delivery and collaboration, such as Sakai/ CHEF. The software is free and works on various platforms, being 100% Java code on the server side. There is thus no tie in to specific hardware, and the system will scale well financially with no licensing costs for extra server nodes etc. There are good logical and technical reasons also. Sakai will be completely open source and open architecture, so institutions will be able to customise existing, or add new, Java JSR-168 compliant portlet codes to connect to a legacy system running on their campus, e.g. to use existing project services such as databases, evaluations and timetables. Sakai will be configurable to use various different database management products, both commercial and open source, thus avoiding RDBMS lock in. At Daresbury an Oracle 9i RAC meta-data server and SRB are being used, but not all projects use Oracle so we tested PostgreSQL too. By exposing existing VRE services using the standard portlet API through a customisable portal framework, and re-using much existing code and sharing additional collaboration tools, we can achieve a large reduction in software development outlay and encourage closer community integration. Of course, this reduction has to be offset against the potentially increased need for software developers to throw away their bespoke solutions and adapt or interface to legacy systems so that they can co-exist with a VRE framework. If they do so however we can achieve a portable and maintainable solution.
This evaluation has addressed the following areas:
The deliverables of the project summarised in the remainder of the full report report are:
Our work has been written up in this report and also included in a couple of papers to the 2004 e-Science All Hands workshop, see References [30,31]. An interim report was submitted to Nicole Harris of JISC on 3/9/04.
Separate summaries, project deliverables and the full report are available from the Sakai Evaluation worksite of the ReDRESS Portal. It is accessible from http://redress.lancs.ac.uk:8080/portal by logging in with (username=guest, passwd=eResearch). The full report is also available from http://www.grids.ac.uk/Sakai/sakai_doc.pdf.
It is clear that, whilst we have attempted to be inclusive in our
survey and review, such work is time-bound and never complete. We
apologise if we have omitted to reference the full set of appropriate
tools and technology and beg their authors to contact us with more