The task allocation problem consists in defining a space and time sharing of system resources that can meet all logical and operational constraints of the tasks being considered. For real-time tasks with predictable behavior (periodic), static allocation provides a reliable way to meet temporal scheduling requirements, together with other operational constraints, such as task replication for fault-tolerance. Unfortunately, finding an optimal feasible allocation is known to be NP-hard, and thus heuristic approaches have been proposed. In this paper, the use of a new heuristic to generate static allocations for critical periodic tasks in real-time distributed control systems is described.