摘 要
随着高清视频流业务的快速发展，传统云计算架构在应对海量视频数据处理时面临网络带宽压力大和传输时延高的问题。为解决上述挑战，本文提出一种边缘计算与云计算协同驱动的低时延视频流处理架构。该架构通过在靠近用户侧部署轻量级边缘节点，实现视频内容的本地化预处理与分流，同时依托云端完成复杂计算任务，形成“边缘-云”协同工作机制。研究设计了基于动态权重的任务卸载决策模型，综合考虑网络状态、节点负载及任务优先级等因素，优化资源调度策略，并引入轻量化的视频编码算法以提升边缘处理效率。实验基于真实网络环境构建仿真平台，结果表明所提架构相较传统云计算模式平均延迟降低42.7%，系统吞吐量提升31.5%。本研究的主要创新在于构建了面向视频流处理的“边缘-云”协同框架，并提出了适应动态网络环境的任务调度机制，有效提升了视频处理系统的实时性与资源利用率，为未来智能视频服务提供了理论支持与实践参考。
关键词：边缘计算；云计算协同；视频流处理；任务卸载决策模型；低时延架构
Research on Low-Latency Video Streaming Processing Architecture Based on Edge Computing and Cloud Computing Collaboration
Abstract: With the rapid development of high-definition video streaming services, traditional cloud computing architectures face significant challenges in terms of network bandwidth pressure and high transmission latency when handling massive volumes of video data. To address these issues, this paper proposes a low-latency video streaming processing architecture driven by the synergy of edge computing and cloud computing. The architecture deploys lightweight edge nodes close to end users to enable localized preprocessing and traffic offloading of video content, while relying on the cloud to perform complex computational tasks, thereby establishing an "edge-cloud" collaborative working mechanism. This study designs a task offloading decision model based on dynamic weights, which comprehensively considers factors such as network conditions, node load, and task priority to optimize resource scheduling strategies. Additionally, a lightweight video encoding algorithm is introduced to enhance processing efficiency at the edge. The experiments are conducted on a simulation platform built upon a real-world network environment. Results demonstrate that compared to traditional cloud computing approaches, the proposed architecture reduces average latency by 42.7% and improves system throughput by 31.5%. The main contributions of this research lie in the establishment of an "edge-cloud" collaborative fr amework tailored for video streaming processing and the development of a task scheduling mechanism adaptive to dynamic network environments, effectively enhancing the real-time performance and resource utilization of video processing systems. This work provides both theoretical support and practical references for future intelligent video services.
Keywords: Edge Computing; Cloud Computing Collaboration; Video Stream Processing; Task Offloading Decision Model; Low Latency Architecture
目 录
1绪论 1
1.1研究背景与意义 1
1.2国内外研究现状分析 1
1.3研究内容与方法概述 1
2边缘计算与云计算协同架构设计 2
2.1协同架构的系统模型构建 2
2.2资源分配与任务调度机制 2
2.3视频流处理中的边缘-云节点协作策略 2
2.4架构性能评估指标设定 3
3低时延视频流处理关键技术分析 3
3.1视频编码与传输优化技术 3
3.2动态带宽分配与网络负载均衡 4
3.3实时性保障机制设计 4
3.4多用户并发处理能力研究 5
4基于协同架构的实验与性能评估 5
4.1实验环境与数据集构建 5
4.2系统响应时延测试分析 5
4.3资源利用率与能耗对比 6
4.4不同场景下的稳定性与扩展性验证 6
结论 7
参考文献 8
致 谢 9
