Abstract | ||
---|---|---|
The tightness of stochastic performance bounds has been a lingering issue in the theory and practical applications of stochastic network calculus (SNC). More often than not, inappropriate stochastic traffic arrival models and/or service models lead to loose stochastic bounds. In practice, loose bounds occur due to inaccurate a-prior assumption on traffic arrivals and/or the obliviousness to possible correlation between arrivals. To alleviate this problem, this paper uses copula analysis to capture the correlation in traffic flows and introduces a statistical method to infer traffic arrival models. Using copula theory, we show the range of performance bounds that SNC can achieve. With concrete numerical examples and real-world experiments, it is demonstrated that copula analysis offers a new opportunity for extending SNC research and augmenting its impact in practice. |
Year | DOI | Venue |
---|---|---|
2015 | 10.1109/INFOCOM.2015.7218532 | IEEE International Conference on Computer Communications |
Keywords | Field | DocType |
Network Calculus, Copulas Analysis | Mathematical optimization,Multivariable calculus,Copula (linguistics),Computer science,Network calculus,Stochastic modelling,Copula theory | Conference |
ISSN | Citations | PageRank |
0743-166X | 5 | 0.41 |
References | Authors | |
9 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Fang Dong | 1 | 5 | 1.09 |
Kui Wu | 2 | 30 | 5.06 |
S. Venkatesh | 3 | 53 | 7.11 |