Efficient Digital Twin Data Processing for Low-Latency Multicast Short Video Streaming

TL;DR

This work addresses the challenge of achieving low-latency multicast short video streaming by leveraging an efficient Digital Twin (DT) data processing scheme. It introduces a precise DT data-processing model that links DT model size $m_i$ and user dynamics $\psi_t$ to clustering accuracy via a quadratic fit $f(m_i,\psi_t)$, and a service-latency model that decomposes latency into $\Xi(a_i)$, $\Psi_g(a_i)$, and $\Gamma_g(a_i,B_g)$ components for DT processing, transcoding, and multicast transmission. The optimization problem $\textbf{P}_0$ minimizes the long-term latency under a single active DT model and bandwidth constraints, solved by a diffusion-based TD3 variant (DFTD3) that employs a denoising diffusion process to generate high-quality actions in a high-dimensional action space. Empirical results on real-world short video datasets show that the proposed DT data processing and diffusion-based scheduling substantially reduce service latency compared with benchmark schemes, enabling more responsive DT-assisted MSVS with adaptive model size selection and bandwidth allocation. The approach offers practical impact for edge-enabled, low-latency multimedia delivery in heterogeneous network environments.

Abstract

In this paper, we propose a novel efficient digital twin (DT) data processing scheme to reduce service latency for multicast short video streaming. Particularly, DT is constructed to emulate and analyze user status for multicast group update and swipe feature abstraction. Then, a precise measurement model of DT data processing is developed to characterize the relationship among DT model size, user dynamics, and user clustering accuracy. A service latency model, consisting of DT data processing delay, video transcoding delay, and multicast transmission delay, is constructed by incorporating the impact of user clustering accuracy. Finally, a joint optimization problem of DT model size selection and bandwidth allocation is formulated to minimize the service latency. To efficiently solve this problem, a diffusion-based resource management algorithm is proposed, which utilizes the denoising technique to improve the action-generation process in the deep reinforcement learning algorithm. Simulation results based on the real-world dataset demonstrate that the proposed DT data processing scheme outperforms benchmark schemes in terms of service latency.

Figures (4)

• Figure 1: DT-assisted MSVS scenario.
• Figure 2: DT data processing procedure.
• Figure 3: Data fitting on DT data processing accuracy.
• Figure 4: Performance evaluation on convergence performance and service latency .