Main Concept

concept
In MMM@HPC we employ a technology that provides highly extendable and easily reusable components for simulations protocols and individual application packages. Complex simulation protocols including multiple steps to treat different size- and time-scales will be described in the form of extendable generic workflows. Reusable application interfaces are implemented as reusable components as GridBeans [1] in the UNICORE client. For the simulation workflows the UNICORE workflow services are deployed. Our concept ensures transferability and general applicability of the developed simulation services. The middleware UNICORE is broadly and productively deployed in different grid infrastructures.
In order to implement a particular workflow the user selects those GridBeans that are required to perform the calculations for the particular project intended using the graphical interface of the UNICORE Rich Client (URC) and adjusting the parameters within the GUI to fit their computing requirements. In order to incorporate a wider range of modelling tasks within materials sciences, users may use the framework of existing GridBeans to tailor novel workflows for materials science applications using the URC. Using this approach, all methods implemented in MMM@HPC as GridBeans can be used to generate novel multiscale algorithms in materials science. Since the partners cover a wide range of multiscale materials modelling tools, many applications that are not within the demonstration applications as examples within the MMM@HPC can be realized with little effort using this framework. The developed GridBeans will be made freely available on this web site. In addition, a repository will be implemented as a GridBean deployment service with which available GridBeans can be discovered and downloaded by the user with the URC.


[1] R. Ratering, A. Lukichev, M. Riedel, D. Mallmann, A. Vanni, C. Cacciari, S. Lanzarini, K. Benedyczak, M. Borcz, R. Kluszcynski, P. Bala, G. Ohme, "GridBeans: Support e-Science and Grid Applications," Proceedings of the Second IEEE International Conference on e-Science and Grid Computing (e-Science'06), p. 45, IEEE 2006.

A summary for the MMM@HPC Project is also available for download: fact sheet.