Cloud-Edge Collaborative Transcoding for Adaptive Video Streaming: Enhancing QoE in Wireless Networks

Live streaming services have experienced unprecedented growth, driven by increasing mobile device usage and advancements in wireless network technologies. However, delivering high-quality video streaming to users in dynamic network environments poses significant challenges, especially in balancing computation, latency, and quality of experience (QoE). While Adaptive Bitrate (ABR) streaming enables smooth playback by dynamically adjusting video quality, traditional ABR approaches rely heavily on either cloud-based transcoding or edge computing, both of which have inherent limitations. Cloud-based solutions suffer from high transmission delays and core network congestion, whereas edge computing faces performance bottlenecks due to limited computational resources in multi-user scenarios. To address these challenges, we propose a Cloud-Edge Collaborative Transcoding (CEC) system for live streaming. This system jointly optimizes bitrate selection and transcoding task allocation between cloud and edge servers, leveraging the complementary strengths of cloud scalability and edge proximity. We formulate the problem as a nonlinear mixed-integer programming (MINLP) problem and design a Mixed Branch-and-Bound (Mixed-B&B) algorithm to solve it effectively. Our approach integrates a novel decomposition strategy, which separates bitrate decisions and re-buffering minimization, allowing for efficient computation even in highly dynamic network conditions. Simulation results demonstrate the superiority of the proposed system over conventional cloud-only and edge-only solutions. The CEC system achieves up to 47.5% (20.2%) and 26.3% (9.78%) higher average QoE over dataset HSPDA (FCC18), respectively, while significantly reducing re-buffering events and transmission latency. This work provides a scalable and efficient framework for live streaming, offering a pathway to meet the growing demand for high-quality, low-latency video services in heterogeneous network environments.