Diffusive parallelism
Peter D. Stout, Brian N. Bershad · 1992
The spread of networks and powerful workstations has created an attractive source of parallel computing power. We are exploring a new parallel programming model, called diffusive parallelism, designed specifically for use with large scale, distributed computation systems. The model provides a simple, yet powerful, abstraction to the programmer, while making it possible to build a secure, robust, distributed computation system in the presence of long delays, failure, and untrusted user programs. In contrast, most existing distributed computation systems have attempted to extend programming models appropriate to a single node. Diffusive parallelism provides the programmer with a task heap and a weakly consistent, logically centralized, shared data store, or blackboard. We are implementing diffusive parallelism in a system called Wax.