On the calculation of sample-path backlog bounds in queueing systems over finite time horizons

作者: Michael Beck , Jens Schmitt

DOI: 10.4108/VALUETOOLS.2012.250238

关键词: Bounding overwatchStochastic processSimple extensionStatistical time division multiplexingQueueing theoryNetwork calculusComputer scienceNetwork analysisExtreme value theoryMathematical optimization

摘要: The ability to calculate backlog bounds is of key importance for buffer sizing in packet-switched networks. In particular, it critical capture the statistical multiplexing gains which, turn, calls stochastic bounds. network calculus (SNC) a promising methodology compute such So far literature SNC-based apply only an arbitrary, but fixed single point time. Yet, from engineering perspective, one would rather like have sample path bound, i.e., bound that applies (with certain violation probability) all While, general, are hard obtain we investigate this paper how can be computed over finite time horizons. show simple extension known SNC results lead suboptimal by deriving alternative (based on extreme value theory) bounding Interestingly, none two methods completely dominates other. For new method also discuss evolved into corresponding analysis analogous existing SNC.

参考文章(18)
Laurens de Haan, Ana Ferreira, Extreme value theory : an introduction Springer. ,(2006)
C Embrechts, P., Klüppelberg, T Mikosch, Modelling Extremal Events: for Insurance and Finance ,(1997)
Yuming Jiang, Yong Liu, Stochastic Network Calculus ,(2008)
Nick McKeown, Isaac Keslassy, Guido Appenzeller, Sizing router buffers ,(2005)
Florin Ciucu, Exponential supermartingales for evaluating end-to-end backlog bounds measurement and modeling of computer systems. ,vol. 35, pp. 21- 23 ,(2007) , 10.1145/1330555.1330565
Valentin V. Petrov, Sums of Independent Random Variables ,(1975)
Nikolay Likhanov, Ravi R. Mazumdar, Ozcan Ozturk, Large Buffer Asymptotics for Fluid Queues with Heterogeneous M / G /∞ Weibullian Inputs Queueing Systems. ,vol. 45, pp. 333- 356 ,(2003) , 10.1023/B:QUES.0000018026.49168.96
Adam Shwartz, Alan Weiss, LARGE DEVIATIONS FOR PERFORMANCE ANALYSIS ,(2007)
Markus Fidler, An End-to-End Probabilistic Network Calculus with Moment Generating Functions international workshop on quality of service. pp. 261- 270 ,(2006) , 10.1109/IWQOS.2006.250477