Go Game

October 10, 2021 7:00 PM PDT

---

This document outlines the discussion and technical requirements for a multi-player web-based Go game, including system design, API specifications, and technology choices.

---

Requirements

• Multi-player game for Go
• Each game is between 2 players
• Support for 10,000 concurrent players
• Web-based game with frontend and backend
• Matching system to pair players of similar skill levels

User Journey

1. Click "start game" -> request sent to matching system
2. System sends request to another player
3. Game starts

Inside the Game

• Player A plays with black pieces
• Players take turns
• Winning condition: system grants score to the winner for level up

API Design

• putPeace(user_id, piece_type) -> return ACK(ok/error)
• heartBeat(user_id)
• resumeGame(user_id)

System Design

Technology Choices

• Lambda: Used for serverless architecture
• Database: NonSQL options include Cassandra and DynamoDB
  • Eventual consistency model
  • Cost considerations based on QPS (Queries Per Second)
  • Max throughput: 10,000 QPS for strong reads, 3,000 QPS per partition, requiring 4 partitions

Database Design

• API:
  • PutPiece(user_id, piece_type, coordinate_x, coordinate_y)
  • IsSecure(user_id) -> return error
• Pseudo code for put piece function:
  • getLatestBoardFromDB(GameID)

Interviewer Comments

Soft Skills

• Requirement gathering and decision-making are critical
• Solutions should be described clearly
• Address vague requirements and improve critical journey

Technical Considerations

• System scalability and architecture diagrams are essential
• Pseudo code should be at the appropriate level for L4
• Lambda is an excellent solution for scaling and deployment

Areas for Improvement

• Action time: cannot trust client-side inputs due to potential cheating
• Engage constructively with interviewer suggestions
• Make decisions with clear trade-offs

Technical Implementation

• Heartbeat mechanism using JavaScript, WebSocket, and push notifications
• Define message types for client-server communication
• Handle 10,000 QPS effectively

Project Lifecycle

• Discuss methods for handling extensions and rollbacks
• Feature flags are widely used in the industry; justify their use

Audience Considerations

• Questions on server hardware settings and core handling for L5 candidates
• Discuss DynamoDB consistency models, including strong and eventual consistency
• API calls and server-client interactions should be clearly defined