Bottlenecks on the way towards fractal characterization of network traffic estimation and interpretation of the Hurst parameter
نویسندگان
چکیده
In this paper we investigate practical problems of fractal characterization of network tra c focusing on the estimation and interpretation of the Hurst parameter. The analysis is based on our measurement study of ATM WAN tra c. We point out that in order to use the fractal characterization framework in practice we are faced with various misleading e ects that can deceive our self-similarity tests and Hurst parameter estimation methods. It is shown that the estimation of the Hurst parameter depends on several factors, e.g. the estimation technique, sample size, time scale, level shifts, correlation structure. The dependencies are illustrated in examples including the e ects of practical mechanisms like shaping or policing. We conclude that the estimated value of the Hurst parameter may be distorted in many practical cases and it may have no information for practical usage.
منابع مشابه
Adaptive Sampling Methods to Determine Network Traffic Statistics including the Hurst Parameter
Accurate traffic characterization by packet source is needed to predict network behavior and to properly allocate network resources to achieve a desired Quality of Service for all network users. As networks have become faster, the processing load required for complete packet sampling has also grown. In some cases, for example Gigabit Ethernet, the network can deliver packets faster than a netwo...
متن کاملA Novel Description of Multifractal Phenomenon of Network Traffic Based on Generalized Cauchy Process
Let D(n) and H(n) be the fractal dimension and the Hurst parameter of traffic in the nth interval, respectively. Thus, this paper gives the experimental variance analysis of D(n) and H(n) of network traffic based on the generalized Cauchy (GC) process on an interval-by-interval basis. We experimentally infer that traffic has the phenomenon Var[D(n)] > Var[H(n)]. This suggests a new way to descr...
متن کاملSelf-Similarity and Long-Range Dependence in Teletraffic
This paper revisits three important concepts in fractal type network traffic, namely, self-similarity (SS), long-range dependence (LRD), and local self-similarity (LSS). Based on those concepts, we address the reason why the local properties of fractional Gaussian noise (fGn) are contained in the global properties of fGn and vice versa, which may be a limitation of fGn in data traffic modeling....
متن کاملA Mixed-Fractal Model for Network Traffic
In this short paper, we propose a new multi-fractal flow model, aiming to provide a possible explanation for the crossover phenomena that appear in the estimation of Hurst exponent for network traffic. It is shown that crossover occurs if the network flow consists of several components with different Hurst components. Our results indicate that this model might be useful in network traffic model...
متن کاملModeling Superposition of Many Sources Generating Self Similar Traffic
This paper presents steps towards creating a general model for aggregated traffic streams. We show that fitting the mean, variance and Hurst parameter is insufficient to consistently model a long range dependent traffic stream. A fourth parameter, the “level of aggregation,” is required. We also show that with increased aggregation, the behaviour of a traffic stream tends towards that of a frac...
متن کامل