Recent traffic measurement studies from a wide range of working packet networks have convincingly established the presence of significant statistical features that are characteristic of fractal traffic processes, in the sense that these features span many time scales. Of particular interest in packet traffic modeling is a property called long-range dependence, which is marked by the presence of correlations that can extend over many time scales. In this paper, we demonstrate empirically that, beyond its statistical significance in traffic measurements, long-range dependence has considerable impact on queueing performance, and is a dominant characteristic for a number of packet traffic engineering problems. In addition, we give conditions under which the use of compact and simple traffic models that incorporate long-range dependence in a parsimonious manner (e.g., fractional Brownian motion) is justified and can lead to new insights into the traffic management of high-speed networks. 1...

Recent studies have suggested that stock markets' volatility has a type of long-range dependence that is not appropriately described by the usual Generalized Autoregressive Conditional Heteroskedastic (GARCH) and Exponential GARCH (EGARCH) models. In this paper, different models for describing this long-range dependence are examined and the properties of a Long-Memory Stochastic Volatility (LMSV) model, constructed by incorporating an Autoregressive Fractionally Integrated Moving Average (ARFIMA) process in a stochastic volatility scheme, are discussed. Strongly consistent estimators for the parameters of this LMSV model are obtained by maximizing the spectral likelihood. The distribution of the estimators is analyzed by means of a Monte Carlo study. The LMSV is applied to daily stock market returns providing an improved description of the volatility behavior. In order to assess the empirical relevance of this approach, tests for long-memory volatility are described and applied to an e...

A joint estimator is presented for the two parameters that define the long-range dependence phenomenon in the simplest case. The estimator is based on the coefficients of a discrete wavelet decomposition, improving a recently proposed wavelet-based estimator of the scaling parameter [4], as well as extending it to include the associated power parameter. An important feature is its conceptual and practical simplicity, consisting essentially in measuring the slope and the intercept of a linear fit after a discrete wavelet transform is performed, a very fast (O(n)) operation. Under well justified technical idealisations the estimator is shown to be unbiased and of minimum or close to minimum variance for the scale parameter, and asymptotically unbiased and efficient for the second parameter. Through theoretical arguments and numerical simulations it is shown that in practice, even for small data sets, the bias is very small and the variance close to optimal for both parameters. Closed for...

Abstract—RPS is a publicly available toolkit that allows a practitioner to straightforwardly create flexible online and offline resource prediction systems in which resources are represented by independent, periodically sampled, scalar-valued measurement streams. The systems predict the future values of such streams from past values and are composed at runtime out of a large and extensible set of communicating components that are in turn constructed using RPS’s extensible sensor, prediction, wavelet, and communication libraries. This paper describes the design, implementation, and performance of RPS. We have used RPS extensively to evaluate predictive models and build online prediction systems for host load, Windows performance data, and network bandwidth. The computation and communication overheads involved in such systems are quite low. Index Terms—Distributed systems, performance of systems. æ 1

It has been known for several years now that variable-bit-rate video sources are strongly auto-correlated. Recently, several studies have indicated that the resulting stochastic processes exhibit long-range dependence properties. This implies that large buffers at intermediate switching points may not provide adequate delay performance for such classes of traffic in Broadband packet-switched networks (such as ATM). In this paper, we study the effect of long-memory processes on queue length statistics of a single queue system through a controlled fractionally differenced ARIMA(1; d; 0) input process. This process has two parameters OE 1 (0 OE 1 1) and d (0 d ! 1=2) representing an auto-regressive component and a long-range dependent component, respectively. Results show that the queue length statistics studied (mean, variance and the 0:999 quantile) are proportional to e c1 OE 1 e c2 d ; where (c 1 ; c 2 ) are positive constants, and c 2 ? c 1 : The effect of the auto-correlation...

This paper evaluates linear models for predicting the Digital Unix five-second load average from 1 to 30 seconds into the future. A detailed statistical study of a large number of load traces leads to consideration of the Box-Jenkins models (AR, MA, ARMA, ARIMA), and the ARFIMA models (due to self-similarity.) These models, as well as a simple windowed-mean scheme, are evaluated by running a large number of randomized testcases on the load traces. The main conclusions are that load is consistently predictable to a useful degree, and that the simpler models such as AR are sufficient for doing this prediction.

the authors and should not be interpreted as necessarily representing the official

The burstiness of Internet traffic was established in pioneering work in the early 1990s, which demonstrated that packet arrival times are not Poisson, and packet and byte counts in fixed-length intervals are long-range dependent [17, 20]. Here we demonstrate that these results are one end of a continuum of traffic characteristics. At the other end are Poisson behavior and independence. Our study focuses on packets, what devices actually see; we study the statistical properties of packet inter-arrival times and packet sizes. As the traffic load increases --- that is, as the number of simultaneous transport connections increases --- arrivals tend to Poisson and sizes tend to independence. More specifically, long-range dependence of inter-arrivals and sizes decreases to independence, and the marginal distribution of inter-arrivals tends toward exponential; this happens (1) through time on a single link as the load increases due to daily variation, or (2) at a single point in time as the load increases going from lightly loaded links at the edges of the Internet to heavily loaded links at the core. Convergence is rapid; the packet traffic gets quite close to Poisson and independent loads far less than the maximum we observe.

We develop an ordinary least squares estimator of the long memory parameter from a fractionally integrated process that is an alternative to the Geweke Porter-Hudak estimator. Using the wavelet transform from a fractionally integrated process, we establish a log-linear relationship between the wavelet coe cients ' variance and the scaling parameter equal to the long memory parameter. This log-linear relationship yields a consistent ordinary least squares estimator of the long memory parameter when the wavelet coe cients ' population variance is replaced by their sample variance. We derive the small sample bias and variance of the ordinary least squares estimator and test it against the Geweke Porter-Hudak estimator and the McCoy Walden maximum likelihood wavelet estimator by conducting a numberofMonte Carlo experiments. Based upon the criterion of choosing the estimator which minimizes the mean squared error, the wavelet OLS approach was superior to the Geweke Porter-Hudak estimator, but inferior to the McCoy Walden wavelet estimator for the processes simulated. However, given the simplicity of programming and running the wavelet OLS estimator and its statistical inference of the long memory parameter we feel the general practitioner will be attracted to wavelet OLS estimator. Keywords

Analysis and modeling of computer network traffic is a daunting task considering the amount of available data. This is quite obvious when considering the spatial dimension of the problem, since the number of interacting computers, gateways and switches can easily reach several thousands, even in a Local Area Network (LAN) setting. This is also true for the time dimension: W. Willinger and V. Paxson in [42] cite the figures of 439 million packets and 89 gigabytes of data for a single week record of the activity of a university gateway in 1995. The complexity of the problem further increases when considering Wide Area Network (WAN) data [28]. In light of the above, it is clear that a notion of importance for modern network engineering is that of invariants, i.e. characteristics that are observed with some reproducibility and independently of the precise settings of the network under consideration. In this tutorial paper, we focus on two such invariants related to the time d...