Abstract
This study introduces a new analytic queuing model, the N-Burst/G/1 model with heavy-tailed service-time distribution, which captures many of the issues that affect Web servers as observed by empirical studies. An asymptotic calculation of the model?s waiting-time distribution is presented; this relies on calculating the waiting-time distribution in the M/G/1 model with heavy-tailed service-time distribution. Finally, using real data and simulation we verify the model?s assumptions and demonstrate its accuracy.