Abstract:

In weakly hard real-time applications, the （p,k） constraint can provide finer QoS guarantee. To improve the scalability of the streams with the （p,  k） constraint, this paper analyzes the （p,k） constraint and presents the evidence of classification. Then, the corresponding static and dynamic class s selection algorithms are brought up. Experimental results show that the presented algorithms can provide a proper trade-off between the performance an   d the overhead.

Key words: weakly hard real-time, QoS, scalability, dynamic failure