论文信息 - Debugging Time Behavior of Probabilistic Real-time Systems

Debugging Time Behavior of Probabilistic Real-time Systems

In order to detect bugs in the temporal behavior of real-time systems (i.e., discrepancies between the theoretical and the actual behavior), measurements of the running system can be taken. These measurements must then be compared with the specifications and the model used to analyze the feasibility of the system. This comparison, which is straightforward when a deterministic model is used, becomes more difficult when probabilistic constraints and models are used. In this paper we discuss this problem and explore some solutions. We propose two heuristic metrics, optimism and pessimism, and show, by means of a case study, how they help in finding bugs.

Joaquín Entrialgo | J. I. Garcia | J. L. Días | D. F. and García

[1] Mark K. Gardner,et al. Probabilistic analysis and scheduling of critical soft real-time systems , 1999 .

[2] Alan Burns,et al. An extendible approach for analyzing fixed priority hard real-time tasks , 1994, Real-Time Systems.

[3] Joaquín Entrialgo,et al. Measurement-based analysis of real-time posix applications , 2003, ICSE 2003.

[4] Chang-Gun Lee,et al. Stochastic analysis of periodic real-time systems , 2002, 23rd IEEE Real-Time Systems Symposium, 2002. RTSS 2002..

[5] James W. Layland,et al. Scheduling Algorithms for Multiprogramming in a Hard-Real-Time Environment , 1989, JACM.

[6] Guillem Bernat,et al. WCET analysis of probabilistic hard real-time systems , 2002, 23rd IEEE Real-Time Systems Symposium, 2002. RTSS 2002..

[7] John P. Lehoczky,et al. Fixed priority scheduling of periodic task sets with arbitrary deadlines , 1990, [1990] Proceedings 11th Real-Time Systems Symposium.