A Scheduling Algorithm for Hybrid Distributed Real-Time Systems

Real-time systems that support only a finite set of discrete configurations can be referred to as hybrid systems. However, recent Feedback Control Scheduling (FCS) algorithms for hybrid real-time systems are only applicable to single-processor systems. And existing FCS algorithms for distributed systems often assume that the systems have continuous control inputs. This paper proposes a FCS algorithm for hybrid distributed real-time systems, which include both tasks supporting continuous configurations and tasks supporting discrete configurations. The algorithm is based on a mixed integer predictive control approach. The experimental results show that this algorithm can provide real-time performance guarantees efficiently, even in open environments.