System requirements
Functional:
- User can post messages
- Assume no video and image
- User can follow other users
- User can track the posts from followees and him/herself in reverse chronological order
Non-Functional:
- Service needs to scale in peak time to support 500M daily active users
- Service needs to be available
Capacity estimation
Estimate the scale of the system you are going to design...
- Assume 3B users in total
- Assume it does not increase for simplicity
- Assume they have 500M DAU
- Assume it does not increase for simplicity
- An average user posts 5 times a day
- An average user refresh the page 20 times a day
- Posts will be saved for 10 years
- An average user follows 50 users
- A post is 300 byte
- 300 * 500M * 5 * 365 * 10 = 2.74 PB -> 3 PB to store posts
- User metadata is 200 byte
- 200 * 3 billion = 600 GB
- One follow is 16 byte
- 16 * 3B * 50 = 2.4 TB
API design
Define what APIs are expected from the system...
- post(String message)
- follow(String user)
- unfollow(String user)
- refresh()
- Return a list of posts
Database design
Defining the system data model early on will clarify how data will flow among different components of the system. Also you could draw an ER diagram using the diagramming tool to enhance your design...
- User
- UserID (primary key) 8B
- userEmail 100B
- name 50 B
- creationDate 8B
- Post
- PostID (primary key) 8B
- UserID 8B
- message: 280B
- Follow
- FolloweeID 8B
- FollowerID 8B
High-level design
You should identify enough components that are needed to solve the actual problem from end to end. Also remember to draw a block diagram using the diagramming tool to augment your design. If you are unfamiliar with the tool, you can simply describe your design to the chat bot and ask it to generate a starter diagram for you to modify...
3 services and 2 main DB
- PostService
- Post messages
- FollowService
- Handle follow/unfollow users
- RefreshService
- Provide latest timeline
- Metadata DB
- Store posts
- Graph DB
- Store who follows who
Request flows
Explain how the request flows from end to end in your high level design. Also you could draw a sequence diagram using the diagramming tool to enhance your explanation...
Post
- Client -> LB -> API gateway -> PostService -> MetadataDB -> RefreshService -> MetadataDB,GraphDB -> Client
Follow
- Client -> LB -> API gateway -> FollowService -> GraphDB, cache
Refresh
- Client -> LB -> API gateway -> RefreshService -> MetadataDB,GraphDB -> Client
Detailed component design
Dig deeper into 2-3 components and explain in detail how they work. For example, how well does each component scale? Any relevant algorithm or data structure you like to use for a component? Also you could draw a diagram using the diagramming tool to enhance your design...
All the services & DB are hosted in multiple regions and instances for scalability and fault tolerance.
- RefreshService
- User pagination (with like 100 latest posts) to create timeline so that return values will not be too large
- Search the most recent posts of followees and the user from metadata and return
Trade offs/Tech choices
Explain any trade offs you have made and why you made certain tech choices...
- Graph DB is used to store "follow" info
- Metadata could be SQL
Failure scenarios/bottlenecks
Try to discuss as many failure scenarios/bottlenecks as possible.
- Always active services are tracked with service monitors like Zookeeper and queries goes to active service
- If request fails, impose retry strategy
Future improvements
What are some future improvements you would make? How would you mitigate the failure scenario(s) you described above?
- Support uploading media
- Display trends