[Solution] End-to-end Scheduling of Real-time Task Pipelines on Multiprocessors

Task pipelines are common in today’s embedded systems, as data moves from source to sink sensing-processing-actuation task chains. A real-time pipeline is constructed by connecting a series of periodic tasks with buffers. In time-critical system, end-to-end timing and data-transfer properties must be guaranteed. guarantee could mathematically expressed assigning constraints the pipeline. However, deriving scheduling parameters meet guarantees an NP-hard constraint optimization problem. Hence, traditional solver not suitable runtime solution.In this paper, we present heuristic algorithm, CoPi, derive execution times periods pipelined that schedulability requirements. We consider two upper bound on pipeline: delay loss-rate. After satisfying these constraints, CoPi schedules set asynchronous independent tasks, under rate-monotonic algorithm. Simulations show has comparable acceptance ratio significantly better than open-source MINLPsolvers. Furthermore, use map multiple multiprocessor system. demonstrate partitioned algorithm coupled accommodates dynamically appearing pipelines, while attempting minimize migrations.