ABSTRACT
It is known that the performance of queueing mechanisms in computer network nodes deteriorates when traffic is strongly autocorrelated, or has the batch structure, or when the service time is of high variability. It is not obvious however, which of those three factors plays the dominant role in this deterioration. In this paper, using realistic traffic parameterizations with different system loads and buffer sizes, we are trying to determine, which of the three prevails in making queues longer and packet losses higher.
|
REFERENCES
[1] Leland W., Taqqu M., Willinger W. and Wilson D. On the self-similar nature of ethernet traffic (extended version). IEEE/ACM Transactions on Networking 2(1): pp. 1–15, (1994).
[2] Willinger W., Taqqu M. S., Sherman R. and Wilson D. V. Self-Similarity through High Variability: Statistical Analysis of Ethernet LAN Traffic at the Source Level. Proceedings of ACM SIGCOMM95, pp. 100–113, (1995).
[3] Crovella M. and Bestavros A. Self-similarity in WorldWideWeb traffic: Evidence and possible causes. IEEE/ACM Transactions on Networking 5(6), pp. 835-846, (1997).
[4] Garetto M. and Towsley D. An efficient technique to analyze the impact of bursty TCP traffic in wide-area networks. Performance Evaluation 65, pp. 181-202, (2008).
[5] Takagi, H. Queueing analysis - Vacation and Priority System. North-Holland, Amsterdam, (1991).
[6] Casale, G., Zhang, E. Z. and Smirni, E. Trace data characterization and fitting for Markov modeling. Performance Evaluation, Volume 67, Issue 2, pp. 61–79, (2010).
[7] Lucantoni D. M. New results on the single server queue with a batch Markovian arrival process. Commun. Stat., Stochastic Models 7, no. 1, pp. 1–46, (1991).
[8] Chydzinski A. Queue Size in a BMAP Queue with Finite Buffer. Lecture Notes in Computer Science, vol. 4003, pp. 200–210, (2006).
[9] Takagi H. Queueing analysis - Finite Systems. North-Holland Amsterdam, (1993).
[10] Chydzinski A. and Mrozowski P. Queues with dropping functions and general arrival processes, PLoS ONE, vol. 11, issue 3, e0150702. doi:10.1371/journal.pone.0150702, (2016)
[11] http://www.omnetpp.org/
[12] Chydzinski A. and Winiarczyk R. On the Blocking Probability in Batch Markovian Arrival Queues. Microprocessors and Microsystems, 32/1, pp. 45–52, (2008).
[13] Chydzinski, A. On the infuence of traffic autocorrelation on the queueing performance. Proc. of World Multi-Conference on Systemics, Cybernetics and Informatics (WMSCI 2016), pp. 37-42, Orlando, July (2016).
|