System requirements

Functional:

1. Display view count to users.
2. Display real time view count to the users.

Non-Functional:

1. System should be highly available.
2. System can be eventually consistent.

Capacity estimation

Estimate the scale of the system you are going to design...

API design

Host - www.youtube.com

1. GET /v1/video/:videoId?viewCount=?
2. PUT /v1/video/:videoId/view

{

timestamp:

}

Database design

Video Table

1. ID (PK)
2. UserId (FK)
3. ViewCount
4. Timestamp

View table

1. Id (PK)
2. VideId (FK)
3. UserId (FK)
4. timestamp

High-level design

1. Client (Browser)
2. Load Balancer
3. Server
4. DB

Request flows

1. User renders a specific video from the server.
2. While watching the video, if the user has watched atleast 50% of the video, the client triggers a PUT request to the server to increase the video view count.
3. There is kafka queue to queue the views for increased throughput. The consumers fetch the data from Kafka queue and batch write the views to the DB for better throughput.
4. Websocket connection between client and server to get real time view count updates.

Detailed component design

Trade offs/Tech choices

1. Load balancer for routing.
2. Kafka for queueing the view requests to the DB.
3. Websocket connection between client and server to get real time updates.
4. Just clicking on a video shouldn't increase view count.

Failure scenarios/bottlenecks

1. Horizontal scaling for avoiding downtime and no single point of failure.

Future improvements

What are some future improvements you would make? How would you mitigate the failure scenario(s) you described above?