A Reinforcement Learning-Based Routing for Real-Time Multimedia Traffic Transmission over Software-Defined Networking

The resulted data were obtained by implementing the proposed architecture on Ubuntu 16.04 installed in an HP Z230 tower workstation with an Intel Xeon processor and 16 GB RAM. Mininet emulator is used to run the infrastructure plane, which includes the SDN-enabled video streaming network. RYU Controller utilised to emulate the control plane, which collects information about network topology to obtain the environment states. Three realistic network topologies are used for the experimental evaluation of our approach; a modified Abilene topology, Geant, and Cernet. The topologies have been built and implemented in Mininet using a Python script; SDN-Openflow switches replaced the nodes for each network topology. Each switch has a host that forwards and receives different types of traffic. Multimedia providers are deployed in a number of Openflow switches. The provider is able to stream real-time Dynamic Adaptive Streaming over HTTP (DASH) based video flows. DASH video is divided into 4s chunks encoded into five discrete bit rates ranging from 260 Kbps to 2998 Kbps using FFmpeg version 4.3.2 with the H.264 codec, and segmented based on GPAC MP4Box in order to create the DASH manifest and associated files. The video content streamed by multimedia providers is the “Big Buck Bunny” animation with a 1920 × 1080 pixels resolution and was cut into 5 min long. The selection of hosts that partake in the experiment has been constructed to enable the traffic flow to pass through the whole network topology. In the meantime, the study has utilised Wireshark as video traffic monitoring software in the end user's device in order to capture the received video segments during video streaming. The OSPF-based approach is compared with our solution to evaluate customers' desired satisfaction across real-time DASH video streaming. The provided files present the performance comparison of the proposed RL-based solution with the OSPF protocol in terms of the customers’ satisfaction represented by Multimethod Assessment Fusion (VMAF) and Structural Similarity Index Metric (SSIM), and network throughput.

本数据集的生成数据通过将所提架构部署于搭载Intel Xeon处理器、16GB运行内存的HP Z230塔式工作站且预装Ubuntu 16.04操作系统的环境中获得。实验采用Mininet仿真器搭建基础设施平面，该平面包含支持软件定义网络（Software Defined Network, SDN）的视频流网络；采用RYU控制器模拟控制平面，该控制器可采集网络拓扑信息以获取环境状态。本实验采用三种真实网络拓扑开展方法的性能评估，分别为改进型Abilene拓扑、GEANT网络与中国教育和科研计算机网（CERNET）。所有拓扑均通过Python脚本在Mininet中构建实现，且将各网络拓扑中的原始节点替换为支持SDN的OpenFlow协议交换机。每台交换机均挂载一台主机，用于转发和接收各类网络流量。部分OpenFlow交换机上部署了多媒体内容提供商节点，该节点可流式传输基于HTTP的动态自适应流媒体（Dynamic Adaptive Streaming over HTTP, DASH）实时视频流。DASH视频被切割为4秒时长的分片，并使用FFmpeg 4.3.2版本结合H.264编解码器编码为五种固定码率，码率范围为260 Kbps至2998 Kbps；随后通过GPAC MP4Box工具进行分片处理，以生成DASH清单文件及相关附属文件。多媒体内容提供商所流式传输的视频素材为分辨率1920×1080像素、时长5分钟的动画《大雄兔》（Big Buck Bunny）。实验所选参与流量传输的主机路径需覆盖整个网络拓扑，以确保流量完整流经所有网络节点。与此同时，本研究在终端设备中部署Wireshark视频流量监控软件，捕获视频流传输过程中接收的视频分片。为评估实时DASH视频流场景下的用户满意度，本研究将基于开放最短路径优先（Open Shortest Path First, OSPF）的方案与所提方案进行性能对比。本数据集附带的文件展示了所提强化学习（Reinforcement Learning, RL）方案与OSPF协议的性能对比结果，对比指标包括以多方法评估融合（Multimethod Assessment Fusion, VMAF）与结构相似性指数（Structural Similarity Index Metric, SSIM）表征的用户满意度，以及网络吞吐量。