Adaptive Burst Assembly with Dynamic Ratio Control for Improved QOS and Fairness in Optical Burst Switching Networks

Abstract

Optical Burst Switching (OBS) is a promising technology for next-generation Internet backbones. Ensuring Quality of Service (QoS) for real-time traffic within OBS networks while maintaining fairness for all traffic classes remains a significant challenge. Existing composite burst assembly schemes often rely on fixed ratios of high-priority traffic, which are often suboptimal for dynamically varying network conditions leading to increased packet loss, increased latency, and unfairness to lower-priority traffic. This paper addresses these limitations by introducing an enhanced Adaptive Burst Assembly (ABA) scheme. The ABA scheme builds upon established techniques, such as burst segmentation, to dynamically adjust the proportion of high-priority traffic within bursts. The key improvement is the addition of a dynamic ratio adjustment mechanism based on real-time network traffic load measurements. Simulation results demonstrate that the ABA scheme outperforms traditional fixed-ratio methods in reducing high-priority traffic packet loss. While this work builds on well-established concepts in OBS networks, its unique contribution lies in the adaptive nature of the burst assembly process and its focus on balancing both performance and fairness.