### Table 1: \Frequentist quot; upper 95% C.L. limits (accurate to 0.001) on a Poisson process with background, incorporating uncertainties by Bayesian integration.

### Table 2: \Frequentist quot; upper 95% C.L. limits (accurate to 0.001) on a Poisson process with background, incorporating uncertainties by Bayesian integration.

### Table 3: \Frequentist quot; upper 95% C.L. limits (accurate to 0.001) on a Poisson process with background, incorporating uncertainties by Bayesian integration.

### TABLE II SIMULATION INPUTS: POISSON PROCESS PARAMETERS Service Intensity 1/5 1/4 1/3 1/2 1 2 3 4 5

### Table 1: Parameter Estimates for the Poisson Arrival Process with a Gamma-Distributed Correlation Factor for Tuesday (The Number of Arrivals is Per Half Hour)

"... In PAGE 4: ... We get a better goodness of fit when we assume that the arrival process is time-of-the-day and day-of-the- week dependent. Table1 shows the estimated param- eters for Tuesdays. We observe the arrival rates that are time-of-the-day dependent.... ..."

### Table 2 evidently shows, that the MMPP as well as the Poisson process are clearly inferior compared with the customized BMAP and loosely capture standard deviation, skewness and kurtosis, over all considered time-scales. These observations are emphasized by the analysis of traffic burstiness, which can be expressed in terms of the Hurst parameter H. Figure 8 plots the R/S statistics [15] of the measured traffic, the customized BMAP, the MMPP as well as the Poisson process. The degree of traffic burstiness H, can easily derived by the slopes of linear regression plots of the R/S statistics. As expected, the Poisson process (H = 0.5558) fails to capture the traffic burstiness, while the MMPP (H = 0.6408) and the customized BMAP (H = 0.6418) both indicate a significant amount of traffic burstiness compared with the Hurst parameter of the measured traffic (H = 0.6785).

2002

"... In PAGE 14: ... Moreover, these sample paths show the clear advantage of the customized BMAP over the MMPP and the Poisson process, which fail to capture the original sample path over almost all time-scales. Table2 presents additionally statistical properties for the data rates of the measured traffic, the BMAP, the MMPP, and the Poisson process, on different time-scales in terms of mean, standard deviation, skewness, and kurtosis. Recall, that the mean gives the center of the distribution and the standard deviation measures the dispersion about the mean.... In PAGE 14: ... The third moment about the mean measures skewness, the lack of symmetry, while the forth moment measures kurtosis, the degree to which the distribution is peaked. In Table2 , skewness and kurtosis are standardized by an appropriate power of the standard deviation. We observe, that mean and standard deviation of the measured traffic and the customized BMAP perform quite similar over the considered time-scales, with exception of the BMAPs standard deviation on time unit in sec.... In PAGE 14: ...01 0.1 1 Table2... In PAGE 17: ...cales, i.e., 0.01 sec and 0.1 sec, while the customized BMAP overestimates the skewness on the smallest time-scale and underestimates it on the largest time-scale. Furthermore, the last column of Table2 indicates, that kurtosis, i.e.... ..."

Cited by 3

### Table 1: Coefficients of fitted models for spatial point patterns in three sample blocks. Considered models are: CSR - homogeneous Poisson process, SSI - simple sequential inhibition, Strauss - homogeneous Strauss process, StrHC - Strauss hard-core, SC - soft-core.

"... In PAGE 5: ... An example curve of the MPL of the SC model in the three blocks is displayed in Figure 2. The parameters for all final models are displayed in Table1 for the three point patterns. The parameter estimates are similar between the sample blocks.... ..."

### Table 7: Daily Performance Measures Obtained from the Simulation Where the Arrival Process is Poisson with Deterministic Rates and Exponential Service Times

"... In PAGE 7: ... Note that these distributions have the same means as their corresponding counterparts that we have chosen for our original simulation model. Table7 shows a significant increase in the QoS of the simulation under the new set of distributions compared to the original simulation model. This is not surprising, because assuming deterministic arrival rates reduces the traffic variability.... ..."

### Table 1: Simulation parameters. Each request is modeled by interarrival time, client ID, and choice of a video. Interarrival times follows a Poisson process. The client is randomly chosen from inactive stations. The selection of videos follows a Zipf- like distribution [12]. That is, the probability of choos- ing videoi is

### Table 4. Group segregation: Poisson source

2001

"... In PAGE 5: ... Tables 4 and 5 present simulated average waiting times for each class under different probability parameter settings in a 4-class single server PP system where packets have equal size. For Table4 , the arrival process for each class is a Poisson process; the load for each class is AQ BD BP AQ BE BPBCBMBEBEBH, AQ BF BP BCBMBCBL and AQ BG BP BCBMBFBI. For Table 5, each class is loaded with an ON/OFF source.... ..."

Cited by 10