Performance evaluation using periodic system-state measurements

详细信息    查看全文

• 作者：W. Ellens^a ; ^{w.ellens@uva.nl" class="auth_mail" title="E-mail the corresponding author}Author Vitae ; M. Mandjes^a ; ^{m.r.h.mandjes@uva.nl" class="auth_mail" title="E-mail the corresponding author}Author Vitae ; H. van den Berg^b ; ^{j.l.vandenberg@tno.nl" class="auth_mail" title="E-mail the corresponding author}Author Vitae ; D. Worm^b ; ^{daniel.worm@tno.nl" class="auth_mail" title="E-mail the corresponding author}Author Vitae ; S. Błaszczuk^c ; ^{s.blaszczuk@gmail.com" class="auth_mail" title="E-mail the corresponding author}Author Vitae
• 关键词：Transient birth&ndash ; death process ; Measurements ; System behavior inference ; Exceedance probability ; Reversible Markov process
• 刊名：Performance Evaluation
• 出版年：2015
• 出版时间：November 2015
• 年：2015
• 卷：93
• 期：Complete
• 页码：27-46
• 全文大小：1910 K

文摘

This paper deals with the problem of inferring short time-scale fluctuations of a system’s behavior from periodic state measurements. In particular, we devise a novel, efficient procedure to compute four interesting performance metrics for a transient birth–death process on an interval of fixed length with given begin and end states: the probability to exceed a predefined (critical) level m$m$, and the expectation of the time, area, and number of arrivals above level m$m$. Moreover, our procedure allows to compute the variances and cross-correlations of the latter three metrics. The asymptotic behavior of the metrics for small and large measurement intervals is also derived.

An extensive numerical study in the context of communication networks reveals the impact of important system parameters on the considered performance metrics, and shows that the three latter metrics are very highly correlated. We also illustrate through this numerical study how our analysis can be used in practical situations to support e.g., capacity management and sla verification.