An Optimized Hybrid CDN-P2P Framework for Efficient Live Streaming: Design, Implementation, and Evaluation

Abstract

The rapid growth of live video streaming has highlighted critical limitations in traditional Content Delivery Networks (CDNs), particularly their high costs and scalability constraints during traffic surges. While Peer-to-Peer (P2P) networks offer a more scalable and cost-efficient alternative, their reliability remains inconsistent due to variable peer performance. This study introduces an optimized hybrid CDN-P2P framework designed to balance stability and scalability for efficient live streaming. At its core is a dynamic scheduling algorithm, HQACS (Hybrid QoS-Aware Chunk Scheduler), which intelligently prioritizes P2P transmission when peers meet predefined quality thresholds, such as an upload speed exceeding 3 Mbps and uptime above 85%,  while seamlessly switching to CDN support for critical data segments or unstable conditions. The framework was implemented using open-source tools, including Python, Node.js, and Web Torrent, and evaluated through realistic simulations. Results demonstrated a peak P2P throughput of 16.5 MB/s, an average latency of 148 MS, and 88% of video chunks delivered via P2P, reducing CDN reliance to just 12%. These findings underscore the system’s adaptability through intelligent peer selection and robust fallback mechanisms. This research contributes a cost-effective, scalable, and low-latency solution for modern streaming platforms by merging CDN reliability with P2P flexibility. Future work should explore reinforcement learning for adaptive decision-making and large-scale deployments to further refine performance under diverse network conditions. The proposed framework holds significant promise for both academic research and industry applications, offering a practical approach to enhancing live streaming efficiency.