This paper presents a detailed statistical analysis of a 2-hour empirical sample of Variable Bit Rate (VBR) video traffic, obtained by applying a simple intraframe video compression code to an action movie. The key findings include: 1. **Heavy-Tailed Distribution**: The marginal bandwidth distribution is accurately described using "heavy-tailed" distributions, such as the Pareto distribution. 2. **Long-Range Dependence**: The autocorrelation of the VBR video sequence decays hyperbolically, indicating long-range dependence, which can be modeled using self-similar processes. The authors combine these findings to develop a new non-Markovian source model for VBR video and present an algorithm for generating synthetic traffic. Trace-driven simulations show that statistical multiplexing results in significant bandwidth efficiency even when long-range dependence is present. The simulations also demonstrate that the model captures the important components of VBR video traffic that are not accounted for in current models. The paper discusses the coding method, statistical analysis of the trace, and the construction of the VBR video model. It highlights the importance of both the heavy-tailed distribution and long-range dependence in the model. The authors also explore the implications of these findings for network resource allocation and quality of service (QoS) in VBR video applications.This paper presents a detailed statistical analysis of a 2-hour empirical sample of Variable Bit Rate (VBR) video traffic, obtained by applying a simple intraframe video compression code to an action movie. The key findings include: 1. **Heavy-Tailed Distribution**: The marginal bandwidth distribution is accurately described using "heavy-tailed" distributions, such as the Pareto distribution. 2. **Long-Range Dependence**: The autocorrelation of the VBR video sequence decays hyperbolically, indicating long-range dependence, which can be modeled using self-similar processes. The authors combine these findings to develop a new non-Markovian source model for VBR video and present an algorithm for generating synthetic traffic. Trace-driven simulations show that statistical multiplexing results in significant bandwidth efficiency even when long-range dependence is present. The simulations also demonstrate that the model captures the important components of VBR video traffic that are not accounted for in current models. The paper discusses the coding method, statistical analysis of the trace, and the construction of the VBR video model. It highlights the importance of both the heavy-tailed distribution and long-range dependence in the model. The authors also explore the implications of these findings for network resource allocation and quality of service (QoS) in VBR video applications.