Name
Abstract | ||
|---|---|---|
This paper presents the Poisson Pareto burst pro- cess (PPBP) as a simple but accurate model for Internet traffic. It presents formulae relating the parameters of the PPBP to measurable traffic statistics, and describes a technique for fitting the PPBP to a given traffic stream. The PPBP is shown to accurately predict the queueing performance of a sample trace of aggregated Internet traffic. We predict that in few years, natural growth and statistical multiplexing will lead to an efficient optical Internet. I. INTRODUCTION For over a quarter of a century researchers have been looking for a stochastic process which could be used as an accurate and simple model for traffic in packet switched networks. The criteria for such a stochastic process are: (i) It is defined by a small number of parameters. (ii) If these parameters are fitted using measurable statistics of an actual traffic stream the following will be achieved: a) the first and second order statistics including the autocovariance function of the stochastic process (the model) will match those of the actual traffic stream, and b) if fed through a single server queue (SSQ), perfor- mance results for the model will accurately predict those of the real traffic stream fed into an identical SSQ. This must be true for a wide range of buffer sizes as well as for a wide range of service rates. (iii) It is amenable to analysis. If the process also parallels the nature of the traffic that is being modeled, this will give maximum confidence in its usefulness. In this paper we examine the Poisson Pareto burst process (PPBP) and demonstrate that this model meets these challeng- ing criteria. To the best of our knowledge, this makes the PPBP the first model which has been demonstrated to meet all of these criteria. The PPBP is a process based on multiple overlapping bursts, where the burst lengths follow a heavy-tailed distribution. It has been shown that the burst lengths of WAN file transfers |
Year | DOI | Venue |
|---|---|---|
2003 | 10.1109/INFCOM.2003.1208709 | INFOCOM 2003. Twenty-Second Annual Joint Conference of the IEEE Computer and Communications. IEEE Societies |
Keywords | Field | DocType |
Internet,Pareto distribution,queueing theory,stochastic processes,telecommunication traffic,Internet traffic modeling,Poisson Pareto burst process,optical Internet,sample trace queueing performance,statistical multiplexing,traffic statistics | Pareto distribution,Computer science,Computer network,Stochastic process,Queueing theory,Poisson distribution,Statistical time division multiplexing,Pareto principle,Internet traffic,The Internet | Conference |
Volume | ISSN | ISBN |
1 | 0743-166X | 0-7803-7752-4 |
Citations | PageRank | References |
45 | 2.58 | 23 |
Authors | ||
3 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Moshe Zukerman | 1 | 1660 | 175.61 |
| Timothy D. Neame | 2 | 128 | 12.67 |
| Ron Addie | 3 | 168 | 21.39 |