January 30, 2022 7:00 PM PST
This document summarizes the discussions from a mock system design interview focused on creating a real-time comments feature for a newsfeed application. The interview covered functional and non-functional requirements, system design considerations, and technical discussions surrounding the implementation of a low-latency comment delivery system.
Architecture patterns for milti region active-active app
Interview Details
- Interviewer: [Role]
- Interviewee: [Role]
- Level: L5 (Senior)
- Duration: 45 minutes
- Drawing Tool Used: Whimsical
Key Topics Covered
Functional Requirements
- Real-Time Comments: Users should see comments/posts in real time with low latency (target of 100 ms).
- Post Types: Support for both text and video posts, focusing on comments.
- Comment Permissions: Initially deprioritize permissions; consider allowing only friends to comment.
- Scalability Estimates:
- 100M posts with tens of millions of posts every second.
- 100k comments per second can be delivered to potential readers.
- 1M daily active users.
Non-Functional Requirements
- Latency: Top priority for the system design.
- Persistence of Comments: Necessary for scenarios where users access older posts.
System Design Considerations
External APIs
- Push vs. Pull Model:
- Push Model: Continuously sends messages to users.
- Pull Model: May introduce delays in message delivery.
Messaging System
- Message Queue: Utilization of Kafka for message queuing.
- Topic Management: Each post can be treated as a separate topic, potentially requiring 10M message queues.
Consumer Management
- Subscription Model: Users subscribe to their post’s topic rather than all topics.
- Dynamic Subscription: As users scroll, they can unsubscribe from topics that are no longer relevant.
Communication Mechanisms
- Frontend Service: Acts as a bridge between the comment receiver and the user interface.
- WebSocket Usage: Proposed for real-time communication, allowing for bidirectional communication between client and server.
Scalability Challenges
- Fan-Out Strategy: Addressing how to efficiently deliver comments to millions of users simultaneously, especially for high-traffic posts.
- Dynamic Post ID Management: Frontend services need to maintain an updated list of post IDs based on user interactions.
Technical Discussions
- Caching Strategies: Implementing a cache to store user-post relationships for quick access.
- Handling High Traffic: Strategies for managing large volumes of comments and ensuring low latency during peak times.
- Data Consistency: Challenges related to maintaining consistency across distributed systems, especially with comment ordering.
Feedback and Recommendations
Interviewer Feedback
- The interviewee demonstrated satisfactory requirement gathering skills but needs to improve on clarifying the push vs. pull model and its implications on user experience.
- Suggested further practice on designing low-latency systems and understanding the complexities of real-time data delivery.
Audience Feedback
- Evaluated on soft and hard skills, with emphasis on the need for a deeper understanding of system design principles related to real-time applications.
Conclusion
The mock interview highlighted the complexities involved in designing a real-time comment system for a newsfeed application. Key challenges include ensuring low latency, managing scalability, and maintaining data consistency across distributed systems. Further practice and exploration of these concepts are recommended for the interviewee to enhance their technical proficiency.