OPTIMIZING CONGESTION CONTROL FOR QUALITY OF SERVICE (QOS) IN BANDWIDTH-CONSTRAINED WIRELESS NETWORKS

Abstract

Modern wireless networks typically operate on a best-effort service model, which, while able to support both real-time and non-real-time traffic, often falls short in ensuring the required Quality of Service (QoS) for real-time applications. Real-time applications, such as video streaming, voice over IP (VoIP), and online gaming, are highly sensitive to network conditions and require a predictable, low-latency environment to maintain performance. However, the best-effort model does not prioritize traffic effectively, leading to poor performance under high network load, with issues such as high jitter, excessive delay, and increased packet loss. QoS in wireless networks is traditionally assessed through performance metrics such as throughput, jitter, delay, and packet loss, all of which are crucial in determining the user experience in real-time applications. These metrics directly impact overall network efficiency and user satisfaction, with high delay or packet loss leading to degraded service quality, particularly for latency-sensitive applications. In this context, this study introduces a novel QoS framework tailored specifically for bandwidth-constrained networks, where managing limited resources is crucial. Instead of relying on the traditional approach of over-provisioning bandwidth, which can be inefficient and costly, the proposed model employs differentiated services combined with dynamic scheduling based on real-time measurements of incoming data rates and packet classification. By dynamically adapting the network's resource allocation to the changing traffic demands, the framework ensures that real-time applications receive the necessary priority, while non-real-time traffic is handled more flexibly. This results in a more efficient use of available resources, as bandwidth is allocated based on real-time traffic characteristics rather than fixed allocations. The framework incorporates an optimized queuing mechanism that prioritizes packets based on their type and current queue length, allowing for more accurate traffic management. This mechanism helps minimize delays for high-priority packets, such as those associated with real-time applications, while ensuring that lower-priority packets are processed appropriately without congesting the network. By reducing packet waiting times and minimizing the chances of packet loss, the approach significantly improves the QoS for real-time traffic, even in environments where bandwidth is limited. Furthermore, the model aims to minimize resource over-provisioning, which is a common issue in traditional network designs that often result in underutilized resources or excessive costs for provisioning higher bandwidth than necessary.