A schedulability analysis for weakly hard real-time tasks in partitioning scheduling on multiprocessor systems

Abstract

The importance of missing a deadline is one of the classifications of a real-time tasks or systems. A weakly hard real-time system is a new generation for a real-time system in which some degree of missed or losses deadlines are tolerated occasionally but the missing of deadlines has to be stated precisely. For a traditional real-time system, timing requirements in hard real-time are very restrictive, must meet all the tasks deadlines. Meanwhile, in soft real-time timing requirements, the requirements are too relaxed because there are no guarantees can be given for the deadline, whether it is met or missed. As a systems demand complex and significantly increased functionality, multiprocessor scheduling has been given attention and taken into consideration. In fact, within the multiprocessor, the predictability problems seem even harder. Thus, in order to deal with the problem, the partitioned multiprocessor scheduling techniques with solutions for the bin-packing problem, named R- BOUND-MP-NFRNS (R-BOUND-MP with next-fit-ring noscaling) combining with the exact analysis, called hyperperiod analysis and deadline models; weakly hard constraints and µ- pattern are present to provide weakly hard real-time guarantees under static priority scheduling algorithm. We prove that the proposed approach provides predictable weakly hard real-time tasks through validation and simulation results.