A Java framework for massively distributed symbolic computing Original Research Article

We present a framework for implementing massively distributed applications in symbolic computing. Using this framework, computations with massive resource requirements can be distributed and processed in parallel on a network of workstations or on a large scale network such as the Internet. For each concrete application only minimal code is needed to complement the generic framework in order to enable large scale distributed processing of the application. Our framework introduces a “divide and be conquered” model for massively distributed computations. We compare this model to a more traditional one, in a symbolic computing setting. We stress the major problems and propose solutions for some of them.