Research On The Real-Time Fault-Tolerant Scheduling

Schedulability analysis is an important research direction in real-time systems which can be applied in the multi-core processor scheduling, fault-tolerant real-time scheduling and low-power fault-tolerant scheduling. The Schedulability analysis can base on different means like processor utilization, worst-case response time and demand bounds etc.With high performance characteristics, multi-core processors are rapidly gaining market.Real-time tasks using sequential programming model cannot be taken full advantage of multi-core processors.Parallel real-time tasks repeatedly emits a job collection, which can be utilized multiple cores at the same time.Aiming at parallel real-time scheduling, we analysis feasibility for the sporadic real-time task and each task is specified as a directed acyclic graph. We present special schedulability tests for global earliest deadline first and deadline monotonic to determine whether a set of task can meet all deadlines and give the resource augmentation bound of these schedulability tests.Dependability is a vital system requirement in real-time systems, where fault tolerance techniques play a crucial role of achieving them. In some systems, the use of spatial redundancy is limited, so it becomes essential to use time redundancies to guarantee the reliability of the real-time system. With the new fault model, we must extend the existing schedulability analysis and fault-tolerant schedulabiling algorithms to meet the new demands.Fault burst model describes phenomena which cause potentially random faults over a bounded time interval. Based on worst-case response time schedulability analysis for real-time systems under fault bursts, we propose a fault-tolerant priority assignment strategy. Comparing with traditional fault-tolerant strategy, this strategy improves system fault resilience effectively. We present a fault-tolerant priority configuration search algorithm for the proposed analysis. Moreover, the effectiveness of the proposed approach is evaluated by simulation.Checkpointing is a relatively cost effective method for achieving fault tolerance in real-time systems. Since checkpointing schemes depend on time redundancy, they could affect the correctness of the system by causing deadlines to be missed. This paper provides an exact fault-tolerant schedulability analysis for fault burst model and derive the optimal number of checkpoints.The optimal number of checkpoints can decrease the worst case response time and guarantees the timing constraints. The study shows that a fault-tolerant priority improvement approach can improve system fault resilience.Dynamic voltage and frequency scaling (DVFS) technique used in embedded real-time systems is an effective means to reduce energy consumption. However, adjusting processor frequency will impact on the execution time of real-time tasks. For discrete processor frequency model, the definition of optimal frequency is given firstly. Then, according to the energy-saving efficiency under optimal frequency, each task is assigned an energy priority. A heuristic fault-tolerant scheduling scheme which carries out DVFS on the basis of the schedulability analysis under optimal frequency is proposed.