Sri Lanka

This proposal is a continuation of joint research carried out over the last 4 years. The project will establish a local minigrid incorporating high performance computing resources and extend the grid over the interuniversity backbone network to at least two other local universities which will then be linked to the Swedish cluster. Initial performance evaluation would be done for specific characteristics of the local minigrid against features that are peculiar to a county like SL. Following the successful completion of the initial phase, we expect to start detailed investigations on two fronts: algorithm design for selected classes of problems, especially heuristic optimizations targeting the grid, and, secondly on middleware design for the so called OGSA. The system will then provide a worthwhile computational resource, backed by a knowledgable local staff that could form the basis for a significant enhancement in computational research in many areas. The development cluster will be moved from UU to UCSC early in 2006. Starting with software development on these local facilities, large scale computational science research in Sri Lanka could if necessary make use of even more computing power by accesing links to computational grids outside the country.

Subproject I: A prototype cluster of a type that could be established by research groups within developing countries is set up, and a computational grid is established between UU and Sri Lanka. Guidelines for the implementation of similar clusters and grids open software are derived. Research groups to study algorithms that are appropriate for a grid of such simple clusters are established, and a demonstrator application is studied.

Few people use ATM or credit cards in developing countries such as Sri Lanka given that they do not have sufficient income levels and/or the fact that few ATMs are available in rural areas. The project will provide basic ATM services to people who have mobile phones, especially in rural areas. With the proposed solution people can withdraw money from a Mobile-ATM (M-ATM) without going to a traditional ATM machine. The M-ATM system uses cheap mobiles phones, functioning as payment terminals.

This project is both a continuation and a new direction compared to the Uppsala University project in the first phase. The original proposal was to provide a high performance computing infrastructure within the university environment of a developing country and to justify the concept by developing a pilot computational project in Bioinformatics. As a result of the project, the SweLanka Grid concept was established with computational nodes in Colombo and Uppsala. The problems created by the Tsunami in the Indian Ocean in December 2004 led to logistical problems that delayed the installation and implementation of the equipment. In the aftermath of the tsunami, priorities within Sri Lanka had changed somewhat and hence the simple numerical tsunami model was added to the pilot applications. It is proposed to continue work on the tsunami modeling and to extend the bioinformatics work on advanced data analysis for generic data using the GRID resources provided in Sri Lanka.

Project Duration:
October 2005 – October 2007