Probabilistic QoS Guarantees with FP/EDF Scheduling and Packet Discard in a Real Time Context: A Comparative Study of Local Deadline Assignment Techniques
In this paper, we are interested in comparing local deadline assignment techniques in a multi-hop network supporting real time traffic with end-to-end delay constraints, when the FP/EDF scheduling is used, assuming that packets which don’t respect their local delay constraints are discarded. In each node, packets are scheduled according to their Fixed Priorities (FP), and within the same priority, packets are scheduled according to the Earliest Deadline First (EDF) policy, using local deadlines, which correspond to the sojourn times not to be exceeded in that node. Consequently, an accurate choice of these local deadlines must be done in order to respect the flows’ end-to-end delay constraints and minimize the packet discard rate. As we are interested in giving probabilistic QoS guarantees, we develop a mathematical model to compare the performances of five existing deadline assignment techniques. We show that all these techniques give very high packet discard rates. So, we propose to use another packet discard policy and we show that it gives better results.
KeywordsSojourn Time Laplace Transform Delay Constraint Earliest Deadline First Expedite Forwarding
Unable to display preview. Download preview PDF.
- 2.Knightly, E.D.: Enforceable quality of service guarantees for bursty traffic streams. In: Proc. of INFOCOM 1998, San Francisco, CA, pp. 635–642 (March 1998)Google Scholar
- 3.Liu, J.: Real-time systems. Prentice Hall, New Jersey (2000); Sivaraman, V., Chiussi, F.: Statistical analysis of delay bound violations at an earliest deadline first (EDF) scheduler. Perform. Eval. 36-37(1), 457–470 (1999) Google Scholar
- 4.Sivaraman, V., Chiussi, F.M., Gerla, M.: End-to-end statistical delay service under GPS and EDF scheduling: a comparison study. In: Proc. of IEEE INFOCOM 2001 (April 2001)Google Scholar
- 5.Martin, S., Minet, P., George, L.: FP/EDF, a non-preemptive scheduling combining fixed priorities and deadlines: uniprocessor and distributed cases. INRIA (April 2004)Google Scholar
- 7.Sivaraman, V., Chiussi, F.: Providing end-to-end statistical delay guarantees with earliest deadline first scheduling and per-hop traffic shaping. In: Proc. of INFOCOM 2000, Tel Aviv, Israel (March 2000)Google Scholar
- 8.Marinca, D., Minet, P.: Analysis of deadline assignment in distributed real-time systems. INRIA (March 2004)Google Scholar
- 9.Di Natale, M., Stankovic, J.A.: Dynamic end-to-end guarantees in distributed real time systems. In: Proc. of Real-Time Systems Symposium, San Juan, Puerto Rico, December 7-9 (1994)Google Scholar
- 10.Kao, B., Garcia-Molina, H.: Deadline assignment in a distributed soft realtime systems. In: Proc.of the 13th International Conference in Distributed Computing Systems, pp. 428–437 (1993)Google Scholar
- 11.Sahoo, A., Zaho, W.: Partition-based admission control in heterogeneous networks for hard real time connections. In: Proc. of the 10th International Conference on Parallel and Distributed Computing (October 1997)Google Scholar
- 12.Bonald, T., Poutière, A., Roberts, J.: Statistical performance guarantees for streaming flows using expedited forwarding. In: Proc. of IEEE INFOCOM 2001 (March 2001)Google Scholar
- 13.Takagi, H.: Queueing analysis, vacation and priority systems, vol. 1. North Holland, Amsterdam (1991)Google Scholar