The Organizing Committee of the International Workshop on Programmability for Cloud Networks and Applications (PROCON 2024) is pleased to announce the recipient of the PROCON 2024 Best Paper Award.
Adaptive Load Balancing for Multi-Tier CDN Edge Networks: A Reinforcement Learning Approach
Author: Yinan Qian
Award Details
About the Winning Paper
The paper addresses a critical challenge in content delivery at global scale: how to optimally distribute traffic across multi-tier CDN edge networks under dynamically changing load conditions. As CDN architectures evolve from simple caching layers to sophisticated multi-tier edge networks serving latency-sensitive applications, traditional heuristic-based load balancing approaches struggle to adapt to rapidly shifting traffic patterns.
Qian proposes a reinforcement learning framework that continuously learns and adapts load balancing policies based on real-time edge network conditions, traffic characteristics, and application-level latency requirements. The approach is validated using production-scale CDN traffic traces and demonstrates significant improvements in request distribution efficiency, reduced tail latency, and improved cache hit rates across multi-tier edge deployments.
Selection Process
The Best Paper Award was selected by the PROCON 2024 Technical Program Committee. Papers were evaluated on originality, technical quality, clarity of presentation, and potential for real-world impact on cloud network programmability and application delivery. The award was presented during the PROCON 2024 workshop, held in conjunction with the International Teletraffic Congress.
About PROCON
The International Workshop on Programmability for Cloud Networks and Applications (PROCON) is a specialized forum for research at the intersection of cloud networking, software-defined infrastructure, network programmability, and intelligent edge-to-cloud systems. PROCON is held in conjunction with the International Teletraffic Congress (ITC), the first international conference in networking science and practice, founded in 1955 and technically co-sponsored by the IEEE Communications Society.