Mapping series-parallel streaming applications on hierarchical platforms with reliability and energy constraints

Streaming applications come from various application fields such as physics, where data is continuously generated and must be processed on the fly. Typical streaming have a series-parallel dependence graph, they are hierarchical failure-prone platform, for instance in miniaturized satellites. The goal to minimize energy consumed when processing each set, while ensuring real-time constraints terms of time. Dynamic voltage frequency scaling (DVFS) used reduce consumption, we ensure reliable execution by either executing task at maximum speed, or triplicating it, so that time execute set without failure bounded. We propose structure rule partition map onto prove optimization problem NP-complete. design dynamic-programming algorithm special case linear chains, which optimal class schedules. Furthermore, this provides an interesting heuristic building block designing heuristics general case. compared baseline solution, executed speed. Simulations realistic settings demonstrate good performance proposed heuristics; particular, significant savings can obtained.