Functional partitioning of multi-processor architectures

Many real-time computations such as process control and robotic applications may be naturally distributed in a functional manner. One way of ensuring good performance, reliability and security of operation is to map or distribute such tasks onto a distributed, multi-processor system. The time-critical task is thus functionally partitioned into a set of cooperating sub-tasks. These sub-tasks run concurrently and asynchronously on different nodes (stations) of the system. The software design and support of such a functional distribution of sub-tasks (processes) depends on the degree of interaction of these processes among the different nodes. [Continues.