# Study Guide: Designing Facebook Newsfeed

This guide synthesizes insights from **HelloInterview**, **Grokking the System Design Interview**, and **ByteByteGo**. It highlights the key concepts, requirements, challenges, and optimal solutions for designing a scalable and performant newsfeed system.

---

## Problem Understanding and Requirements

### Functional Requirements
- Users should:
  1. Be able to create posts with text, images, or videos.
  2. Follow other users, pages, or groups.
  3. View a newsfeed of posts from their connections, ordered chronologically or based on relevance.
  4. Fetch additional posts while scrolling (pagination).

### Non-Functional Requirements
- **High scalability** to support billions of users.
- **Low latency**:
  - Feed generation under 2 seconds.
  - Post propagation to followers’ feeds within 5 seconds.
- **Consistency**:
  - Eventual consistency acceptable with tolerable delays.
- **High availability**.

---

## High-Level Design

### Core Components
1. **Feed Publishing**
   - Handles post creation and storage in the database.
   - Updates followers’ feeds for new posts.

2. **Feed Generation**
   - Assembles and ranks posts for a user’s newsfeed.
   - Employs caching and precomputed feeds for efficiency.

3. **Notification Service**
   - Alerts users to new posts or activity.

4. **Media Storage**
   - Stores images and videos in a content delivery network (CDN) for fast access.

5. **Database and Cache**
   - Stores user data, posts, and feed data.
   - Uses distributed cache for high-frequency data retrieval.

---

## Database Design

### Tables
- **User**: Stores user details.
- **Follow**: Tracks relationships between users.
- **Post**: Contains post data with user and timestamp references.
- **Feed** (optional): Precomputed feeds for users.

### Indexes
- Use partition keys for efficient lookups:
  - User ID for posts.
  - Follower-followed relationships in the Follow table.

---

## Key Design Challenges and Solutions

### 1. Scaling Feed Generation
#### Challenges
- High fan-out for users with many followers.
- Efficient ranking and sorting of posts.

#### Solutions
- **Fan-Out Models**:
  - **Fan-out-on-Write** (Push):
    - Precompute feeds when a post is created.
    - Efficient for frequent users; costly for large followings.
  - **Fan-out-on-Read** (Pull):
    - Assemble feeds dynamically on request.
    - Avoids wasted computation for inactive users.
  - **Hybrid**:
    - Push for regular users.
    - Pull for high-follower accounts (e.g., celebrities).
- **Async Workers**:
  - Use message queues and workers to process fan-out asynchronously.
  - Stripe tasks by follower segments for balanced load distribution.

### 2. Handling Users with Many Followers
#### Challenges
- Writing posts to millions of feeds.

#### Solutions
- Partition followers for parallel processing.
- Limit precomputation for inactive or low-priority users.

### 3. Ensuring Fast Reads
#### Challenges
- Large data volume leads to potential delays.

#### Solutions
- Use distributed caches (e.g., Redis) for hot data.
- Implement a replicated cache for highly popular posts to reduce hotspot issues.

### 4. Live Updates and Notifications
#### Challenges
- Delivering real-time updates for new posts.

#### Solutions
- **Polling**: Periodic client requests for updates.
- **Push Notifications**: Notify active users of new content.
- **Server-Sent Events (SSE)**: Efficient server-to-client updates for active users.

---

## APIs

### Feed Publishing API
- **Endpoint**: `POST /v1/posts`
- **Payload**: `{"content": "Hello World!", "media": "image.jpg"}`

### Feed Retrieval API
- **Endpoint**: `GET /v1/feed`
- **Params**: `user_id`, `last_post_id` (for pagination).

---

## Comparative Insights from Sources

| Aspect                | HelloInterview                                | Grokking                                   | ByteByteGo                               |
|-----------------------|-----------------------------------------------|-------------------------------------------|------------------------------------------|
| **Fan-out Model**     | Hybrid (Push & Pull)                         | Focus on Fan-out-on-Write                 | Fan-out-on-Write with optimizations      |
| **Caching**           | Distributed feed cache                      | Cache precomputed feeds                   | Separate feed and post caches            |
| **Updates**           | Polling and Push for real-time updates       | Long polling for live feeds               | Hybrid of polling and server-side push   |
| **Scaling Solutions** | Partition followers for write efficiency     | Pre-generate feeds offline                | Consistent hashing to balance hot keys   |
| **Latency Focus**     | Low latency with eventual consistency         | 2-second limit for feed generation        | Push for active users, pull for inactive |

---

## Final Tips for Interview Success

1. **Clarity in Assumptions**:
   - Clarify use cases and constraints (e.g., follower limits, active user ratio).
2. **Iterative Design**:
   - Start with naive solutions, then incrementally optimize.
3. **Trade-offs Discussion**:
   - Highlight the pros/cons of design choices (e.g., Push vs. Pull).
4. **Focus on Scalability**:
   - Stress handling of fan-out and large user bases.
5. **Visualization**:
   - Use diagrams to show system components and flows.

---

This study guide combines the strengths of all three resources to provide a comprehensive overview for designing a scalable and robust newsfeed system.