Name
Abstract | ||
|---|---|---|
We present a new method to estimate the Hurst parameterof the increment process in network traffic -- a process that is assumed to be self-similar. The confidenceintervals and biasedness are obtained for theestimates using the new method. This new method isthen applied to pseudo-random data and to real trafficdata. We compare the performance of the newmethod to that of the widely-used wavelet method, anddemonstrate that the former is much faster and producesmuch smaller confidence intervals of the Hurstparameter estimate. We believe that the new methodcan be used as an on-line estimation tool for H andthus be exploited in the new TCP algorithms that exploitthe known self-similar and long-range dependentnature of network traffic. |
Year | DOI | Venue |
|---|---|---|
2002 | 10.1109/LCN.2002.1181780 | LCN |
Keywords | Field | DocType |
self-similar traffic,hurst parameterof,hurstparameter estimate,new method isthen,hurst parameter,increment process,novel approach,new tcp algorithm,network traffic,new method,new methodcan,widely-used wavelet method,h andthus,local area networks,confidence interval,parameter estimation | Traffic generation model,Computer science,Hurst exponent,Algorithm,Exploit,Artificial intelligence,Local area network,Estimation theory,Confidence interval,Machine learning,Wavelet,Distributed computing | Conference |
ISSN | ISBN | Citations |
0742-1303 | 0-7695-1591-6 | 16 |
PageRank | References | Authors |
1.34 | 7 | 2 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Houssain Kettani | 1 | 34 | 7.45 |
| John A. Gubner | 2 | 119 | 11.14 |