A Comprehensive Analytical Framework for Contention Resolution in Optical Burst Switching Networks: Performance Evaluation and Future Directions
Keywords:
Optical burst switching, contention resolution, burst loss probability, quality of service, performance evaluation, network architecture, adaptive algorithmsAbstract
The increasing exponential growth of internet traffic across the world has put the further pressure on the need of having efficient optical network architecture that is able to exploit the vast bandwidth that the wavelength division multiplexing (WDM) systems promises. Optical Burst Switching (OBS) has come to be a promising paradigm that will strike a balance between the granularity of optical packet switching and the practical limits of available optical technology. Nevertheless, the basic problem of burst contention is the main performance bottleneck of OBS networks. This general review is a detailed analysis of three major schemes of contention resolution: burst preemption, segmentation and deflection routing. Our comprehensive simulation and analytical modeling approach gives us a strict performance evaluation framework which proves that segmentation decreases the overall burst loss probability by 70-90 percent without sacrificing the fairness index of Jain which is over 0.85. Preemption can ensure almost zero loss in high-quality traffic categories but at the cost of severe losses of fairness in the case of heavy loads. The performance threshold of deflection routing is 0.5 Erlang where instability in the network leads to degradation of performance. In our study, a new adaptive hybrid model has been proposed, which selects resolution strategies according to the dynamic network conditions, and this offered 95 per cent improvements in quality of service (QoS) compliance. The paper ends with implementation guidelines and specifying promising research directions to be pursued with intelligent contention management in next-generation optical networks.
Published
How to Cite
Issue
Section

This work is licensed under a Creative Commons Attribution 4.0 International License.