Evaluation and Comparison of Queues in Classical and Fractal Queuing Systems
Abstract
Introduction: Buffer storages of network nodes are the most important traffic management resource. Research over the past decade and a half showed that network traffic is inherently self-similar and therefore the traditional queue analysis based on the Poisson flow assumption cannot accurately predict the performance of a node and the network in general. The lack of general analytical results in studying queues with self-similar traffic poses the necessity to find a way of its correct description and modeling. Purpose: The goal is to develop a revenue model in self-similar traffic which would allow you to evaluate the queueing characteristics of classical and fractal queuing systems, and to compare them in the first approximation. Results: A model of revenues in self-similar traffic is proposed, characterized by the presence of long-term dependencies in the traffic routes and the presence of a pulsating structure on many time scales. This can adequately reflect the characteristics of the network traffic. Models are proposed for calculating the buffer storages of network nodes which are more realistic for self-similar traffic. Practical relevance: The revenue model for self-similar traffic and the model of calculating buffer storages can be applied in network traffic management and resource planning.Published
2016-04-21
How to Cite
Kutuzov, O., & Tatarnikova, T. (2016). Evaluation and Comparison of Queues in Classical and Fractal Queuing Systems. Information and Control Systems, (2), 48-55. https://doi.org/10.15217/issn1684-8853.2016.2.48
Issue
Section
Information coding and transmission