Different estimators of high quantiles, such as $x_p^c$ proposedin \cite{MarkKrieg-2002}, Weissman's estimator $x_p^w$ and the POT-method are considered. Regarding the estimators $x_p^c$ and $x_p^w$ the asymptoticnormality of the logarithms of ratios of these estimators to thetrue value of the quantile is proved. These estimators areapplied to real data of Web sessions and pages. Furthermore,bootstrap confidence intervals of $x_p^c$ and $x_p^w$ areconstructed for modelled data of different heavy-taileddistributions as well as for Web-traffic data.
