EdgeVR

In this project, we developed an edge offloading-based approach to minimize the on-device computation and the power consumption in the live VR streaming system. Please read and refer to our recently accepted NOSSDAV '22 paper for more details about this project:

Zichen Zhu, Xianglong Feng, Zhongze Tang, Nan Jiang, Tian Guo, Lisong Xu, and Sheng Wei, “Power-Efficient Live Virtual Reality Streaming Using Edge,” Workshop on Network and Operating System Support for Digital Audio and Video (NOSSDAV), Jun. 2022.

And our paper is placed under the paper folder.

System Setup

Edge Server

Renderer and Viewport Predictor

1. Clone the repo from FFmpeg360 git clone https://github.com/bingsyslab/ffmpeg360.git
2. Open the folder cd ffmpeg360
3. Check out the commit git checkout 70d1a945213a298c588486270cac46196bd338d3
4. Apply changes from our work git apply ../renderer_and_viewport_predictor.diff

RTMP Server

1. Install Dokcer with instructions from Docker: https://docs.docker.com/engine/install/
2. Run the RTMP server docker run --name nginx-rtmp -p 1935:1935 -p 8080:8080 -d jasonrivers/nginx-rtmp

Client - Video Player

1. Clone git repo from ijkplayer git clone https://github.com/bilibili/ijkplayer.git
2. Open the folder cd ijkplayer
3. Run ./init-android.sh
4. Check out the commit git checkout cced91e3ae3730f5c63f3605b00d25eafcf5b97b
5. Apply changes from our work git apply ../client_player.patch
6. Open ffmpeg folder cd android/contrib/ffmpeg-armv7a
7. Apply changes from our work git apply ../../../../client_sei.diff
8. Follow instructions in ijkplayer

Dataset

Please download from https://wuchlei-thu.github.io/.

Cite Our Work

Zichen Zhu, Xianglong Feng, Zhongze Tang, Nan Jiang, Tian Guo, Lisong Xu, and Sheng Wei, “Power-Efficient Live Virtual Reality Streaming Using Edge,” Workshop on Network and Operating System Support for Digital Audio and Video (NOSSDAV), Jun. 2022.