J.-P. Goux, J. Linderoth, and M. Yoder,, "Metacomputing and the Master-Worker Paradigm," Preprint ANL/MCS-P792-0200, February 2000. [pdf]
The goal of our work is to create a tool that easily allows users to distribute large scientific computations in metacomputing environments. To achieve this goal, a number of difficult implementation issues must be addressed, which may explain the relative lack of complete tools addressing this purpose. Our tool relies on the simple master worker paradigm, and we show that this paradigm is nicely suited for performing many of the requisite tasks of our metacomputing tool. We describe an implementation and present a case study showing the paradigm's effectiveness in solving large scientific computing problems.