This is a middleware solution built with Python and FastAPI, which is optimized for video streaming.
It serves as a middleware from the browser request to the requested object, which is the video file. The implementation is based on the Range Requests Specification in RFC 7233.
The code has comments in the individual functions to provide further explanations in regard to the different features.
The sequence diagram below shows how the middleware acts for a browser request from a client.
sequenceDiagram
Client->>FastAPI: GET /watch?v=<video_id>
opt Authentication
FastAPI->>Client: 401 Unauthorized
end
Note over FastAPI: Authentication is successful
FastAPI-->>Lyve Cloud: Request for Video
Note over FastAPI, Lyve Cloud: Default is 8 MB per request, 1 MB chunks
Lyve Cloud->>FastAPI: 1 MB Video Data
FastAPI-->>Client: 200 OK [1 MB Video Data]
Note over Client: Time-to-first-byte
Note over Client: Video can start playing
Lyve Cloud->>FastAPI: 1 MB Video Data
FastAPI-->>Client: 206 Partial Content [1 MB Video Data]
Lyve Cloud->>FastAPI: 1 MB Video Data
FastAPI-->>Client: 206 Partial Content [1 MB Video Data]
Note over Lyve Cloud, Client: Repeat until 8 MB transferred
- Lyve Cloud access and secret keys. These can be obtained from the Lyve Cloud Console by creating a new Service Account to be used by this solution.
Python3.
This solution does not do any caching like Redis. It is merely a transfer from the Lyve Cloud endpoint currently. Caching is a future implementation which either be done on both the middleware or CDN level.
Step 0: Change your directory to /code if you are not already in /code. All following steps are done in /code.
Step 1: Get your Lyve Cloud bucket credentials.
Here's what you'll need:
- Access Key
- Secret key
- Region Name
- Endpoint URL
Step 2: Enter your credentials in config.json file:
"access_key": "",
"secret_key": "",
"region_name": "",
"endpoint_url": ""
Step 3: Install all required packages: pip3 install -r requirements.txt.
Step 4: Run the python file (development mode) using
uvicorn main:app --host 127.0.0.1 --port 8080 --reload
Step 5: Stream media.
Any video from your Lyve Cloud storage is accessible via the following link:
http://127.0.0.1:8080/watch?v=<video-name>&bucket=<bucket-name>
e.g. http://127.0.0.1:8080/watch?v=my-video.mp4&bucket=my-media-bucket
This link can be played via browser or media player.
Additional Features: Besides streaming media, the following requests are supported by the solution:
- List Bucket
- List Objects
- Create Bucket
- Upload Object
Visit http://127.0.0.1:8080/docs for full documentation.
From Chrome's DevTools, the TTFB for a 206 Partial Content is on average roughly ~30-70 ms. This is using the default of 1 MB Byte Range requests.
However, this is done on a local environment with 0 distance latency from Client to FastAPI.
sequenceDiagram
Client->>FastAPI: Latency: 0 ms
Note over FastAPI: Code Execution Delay
FastAPI-->>Lyve Cloud: Latency: non-zero [Request]
Note over FastAPI, Lyve Cloud: Lyve Cloud location is ap-southeast-1
Lyve Cloud->>FastAPI: Latency: non-zero [1MB Data]
FastAPI-->>Client: 200 OK [Latency: 0 ms]
Note over Client: Time-to-first-byte
Note over Client: Video can start playing
This means that the ~30-70ms is roughly equivalent to the latency of the Lyve Cloud endpoint and the code implementation execution runtime.
Additional latency on production would be the latency of the CDN / FastAPI endpoint from the client. (Which may include SSL latency delay, authentication delay, etc)
- September 18, 2022: Sasha Chernin (alexander.chernin@seagate.com) on Ubuntu 20.04.
.
├── README.md
├── code
│ └── main.py
│ └── config.json
│ └── requirements.txt
└── images
└── ttfb-result-1.png
└── streaming-video.png
This folder contains all the code files. It also includes the credentials file under config.json
This folder contains all the images used in this README.