Accelerating VR Latency Optimization in Computing-Aware Networks via Graph Extension

Fully immersive Virtual Reality (VR) experiences demand extensive data delivery and processing within ultra-low end-to-end latency, along with fast decision-making to maintain a smooth user experience. However, minimizing the end-to-end latency of VR data presents a complex optimization challenge, which is impractical to solve quickly in large networks with multiple users. In this paper, we introduce a novel graph extension method called Graph Copying and Bridging (GCB) to achieve unified scheduling of communication and computing resources in computing-aware networks (CANs). Considering the non-conservation of flows at computing nodes, we formulate the routing problem in CAN as a nonlinear integer programming (NIP) problem. Then we leverage the GCB method to extend the CAN and develop an efficient algorithm based on the extended network to approximate the optimal solution while significantly reducing solving time. Finally, numerical experiments are conducted to illustrate that the proposed algorithm can reduce average running time by up to 103 times compared to other algorithms.