Design a Notification Service

Question

Design a scalable notification system supporting push, email, and SMS. Cover event ingestion, user preference management, template rendering, delivery tracking, and rate limiting to prevent notification fatigue.

Codemia · Accepted Answer

- **Functional Requirements:**
  - Support for push notifications, emails, and SMS.
  - User preference management for notification channels (opt-in/opt-out).
  - Template rendering for customizable messages.
  - Delivery tracking for monitoring sent notifications and user engagement.
  - Rate limiting to control the frequency of notifications per user to avoid fatigue.

- **Non-Functional Requirements:**
  - Scalability to handle millions of notifications per minute.
  - High availability with a target uptime of 99.99%.
  - Low latency in message delivery under 2 seconds.
  - Strong security measures to protect user data and preferences.

- **Assumptions:**
  - Assume 1 billion active users with an average of 3 notifications per user per day.
  - Total notifications per day = 1 billion * 3 = 3 billion notifications.
  - Peak load estimated at 10% of daily traffic, thus 300 million notifications per hour, or 83,333 notifications per second.

- **Storage Needs:**
  - Each notification takes approximately 1 KB of storage. Therefore, daily storage = 3 billion KB = ~2.8 TB.
  - Consider a retention period of 30 days for delivery tracking, leading to a total storage requirement of ~84 TB.

The architecture consists of the following components:

1. **API Gateway:**
   - **Technology:** Google Cloud Endpoints
   - Manages incoming requests for notifications, user preferences, and delivery status.

2. **Notification Service:**
   - **Technology:** Google Cloud Functions
   - Handles business logic for notification creation, template rendering, and sending to preferred channels.

3. **User Preference Store:**
   - **Technology:** Google Cloud Firestore
   - Stores user preferences for notification channels.

4. **Delivery Service:**
   - **Technology:** Google Pub/Sub
   - Manages event ingestion for notifications and ensures they are sent to the correct delivery channels.

5. **Tracking Service:**
   - **Technology:** Google BigQuery
   - Used for analytics and tracking the delivery status and engagement metrics.

6. **Rate Limiter:**
   - **Technology:** Redis
   - Enforces rate limits based on user preferences to prevent notification fatigue.

The schema design includes the following tables/collections:

1. **Users:**
   - `user_id` (String, Primary Key)
   - `email` (String)
   - `phone_number` (String)
   - `preferences` (Map of notification types with boolean values)

2. **Notifications:**
   - `notification_id` (String, Primary Key)
   - `user_id` (String, Foreign Key)
   - `message` (String)
   - `channel` (Enum: Push, Email, SMS)
   - `status` (Enum: Sent, Delivered, Failed)
   - `timestamp` (Datetime)

3. **Templates:**
   - `template_id` (String, Primary Key)
   - `template_body` (String)
   - `variables` (Map of variable names to types)

4. **Delivery Tracking:**
   - `tracking_id` (String, Primary Key)
   - `notification_id` (String, Foreign Key)
   - `user_id` (String)
   - `delivered_at` (Datetime)
   - `opened_at` (Datetime, Optional)

- **Choosing Google Pub/Sub for Delivery Service:**
  - Pros: Highly scalable and decouples the notification ingestion from delivery, allowing for asynchronous processing.
  - Cons: Slightly increased complexity in handling message guarantees (at-least-once delivery).

- **User Preference Management in Firestore vs. SQL Database:**
  - Firestore allows for easier scalability and real-time updates, suitable for user preference changes.
  - Trade-off: Might lack complex querying capabilities compared to SQL.

- **Rate Limiting with Redis:**
  - Pros: Fast in-memory store ideal for managing state and rate limits.
  - Cons: Requires additional infrastructure management and potential data loss on instance failure (mitigated with persistence).

Design a Notification Service

Google

How to Approach This

Sharpen Your Skills on Codemia

Sample Answer

Requirements

Capacity Estimation

Submit Your Answer

Google Software Engineer Interview Guide

Related Questions

Design a high-throughput Inventory Management System

Design a low-latency Rate Limiting System

Design a fault-tolerant Payment System

Design a fault-tolerant Messaging System

Design Walmart Product Search