Practice/Meta/Design Meta Privacy

Design Meta Privacy

Product DesignMust

Problem Statement

Design a real-time sports scoreboard platform that delivers live game updates, scores, and statistics to millions of concurrent users across web and mobile applications. The system must handle multiple sports (basketball, football, soccer, baseball, etc.) happening simultaneously, with each game generating frequent score updates, player statistics, and game events. Users should receive near-instantaneous updates when scores change, and the system must scale to support major sporting events like championship games that attract tens of millions of viewers. Consider how to efficiently broadcast updates, manage different data models for various sports, and handle the spike in traffic during popular games while maintaining low latency for all users.

Key Requirements

Functional

Multi-sport support -- handle different scoring rules, game structures, and statistics for basketball, football, soccer, baseball, and other sports
Real-time score updates -- push score changes, game events, and statistics to connected clients within 1-2 seconds
Historical data access -- provide past game scores, season statistics, team records, and player performance history
Subscription management -- allow users to follow specific teams, leagues, or games and receive targeted updates
Administrative interface -- enable authorized operators to input scores and game events from venues

Non-Functional

Scalability -- support 50 million concurrent users during peak events, with 10,000+ simultaneous games
Reliability -- maintain 99.99% uptime with automatic failover, no score data loss
Latency -- deliver score updates to users within 2 seconds of ingestion, with p99 API response times under 100ms
Consistency -- ensure all users see the same final score, tolerate temporary inconsistencies during rapid updates

What Interviewers Focus On

Based on real interview experiences, these are the areas interviewers probe most deeply:

1. Real-Time Data Distribution Architecture

How you propagate score updates to millions of users efficiently without overwhelming your infrastructure. This reveals your understanding of pub/sub patterns, push vs pull mechanisms, and connection management at scale.

Hints to consider:

WebSocket connections for persistent bidirectional communication vs server-sent events (SSE) for one-way updates
Fan-out strategies: should you broadcast to all users or use topic-based subscriptions per game?
Connection management: how many WebSocket connections can a single server handle, and how do you load balance?
Fallback mechanisms when real-time connections fail (polling, exponential backoff)

2. Data Model Flexibility Across Sports

The challenge of representing diverse sports with different scoring systems, game structures, and statistics in a unified yet extensible schema. Interviewers want to see if you can balance normalization with query performance.

Hints to consider:

Schema-on-read approaches using JSON/document stores vs rigid relational schemas
Common abstractions: games, periods/quarters, teams, scores, events that work across sports
Sport-specific extensions without creating separate systems for each sport
Trade-offs between a single unified table vs sport-specific tables with shared interfaces

3. Write Path and Data Ingestion

How score data enters your system from multiple venues simultaneously, ensuring reliability and preventing duplicate or out-of-order updates. This tests your understanding of event sourcing and idempotency.

Hints to consider:

Event sourcing pattern: storing each score change as an immutable event vs updating current state
Idempotency keys to handle duplicate submissions from flaky network connections
Timestamp-based ordering and conflict resolution when events arrive out of sequence
Validation layer to prevent invalid scores (e.g., negative points, impossible progressions)

4. Caching Strategy and Read Optimization

The read-to-write ratio is extremely high (millions reading, hundreds writing), making caching critical. Interviewers probe your ability to serve stale data when acceptable and invalidate caches efficiently.

Hints to consider:

Multi-layer caching: CDN for static content, Redis for live scores, application-level caches
Cache invalidation strategies: time-based TTL vs event-driven invalidation
Handling cache stampede during major score changes when millions request simultaneously
What data should never be cached vs what can tolerate 5-10 second staleness?

5. Handling Traffic Spikes and Hot Partitions

Championship games create massive imbalances where 80% of users watch the same game. This tests your ability to identify and solve hot partition problems.

Hints to consider:

Sharding strategies: by game ID, by league, by time, and implications for hot games
Read replicas and how to route traffic to scale reads for popular games
Rate limiting and request prioritization during overload (premium users vs free users)
Pre-warming caches and scaling up infrastructure before predictable events

Practice/Meta/Design Meta Privacy

Design Meta Privacy

Product DesignMust

Problem Statement

Key Requirements

Functional

Multi-sport support -- handle different scoring rules, game structures, and statistics for basketball, football, soccer, baseball, and other sports
Real-time score updates -- push score changes, game events, and statistics to connected clients within 1-2 seconds
Historical data access -- provide past game scores, season statistics, team records, and player performance history
Subscription management -- allow users to follow specific teams, leagues, or games and receive targeted updates
Administrative interface -- enable authorized operators to input scores and game events from venues

Non-Functional

Scalability -- support 50 million concurrent users during peak events, with 10,000+ simultaneous games
Reliability -- maintain 99.99% uptime with automatic failover, no score data loss
Latency -- deliver score updates to users within 2 seconds of ingestion, with p99 API response times under 100ms
Consistency -- ensure all users see the same final score, tolerate temporary inconsistencies during rapid updates

What Interviewers Focus On

Based on real interview experiences, these are the areas interviewers probe most deeply:

1. Real-Time Data Distribution Architecture

Hints to consider:

WebSocket connections for persistent bidirectional communication vs server-sent events (SSE) for one-way updates
Fan-out strategies: should you broadcast to all users or use topic-based subscriptions per game?
Connection management: how many WebSocket connections can a single server handle, and how do you load balance?
Fallback mechanisms when real-time connections fail (polling, exponential backoff)

2. Data Model Flexibility Across Sports

Hints to consider:

Schema-on-read approaches using JSON/document stores vs rigid relational schemas
Common abstractions: games, periods/quarters, teams, scores, events that work across sports
Sport-specific extensions without creating separate systems for each sport
Trade-offs between a single unified table vs sport-specific tables with shared interfaces

3. Write Path and Data Ingestion

Hints to consider:

Event sourcing pattern: storing each score change as an immutable event vs updating current state
Idempotency keys to handle duplicate submissions from flaky network connections
Timestamp-based ordering and conflict resolution when events arrive out of sequence
Validation layer to prevent invalid scores (e.g., negative points, impossible progressions)

4. Caching Strategy and Read Optimization

Hints to consider:

Multi-layer caching: CDN for static content, Redis for live scores, application-level caches
Cache invalidation strategies: time-based TTL vs event-driven invalidation
Handling cache stampede during major score changes when millions request simultaneously
What data should never be cached vs what can tolerate 5-10 second staleness?

5. Handling Traffic Spikes and Hot Partitions

Championship games create massive imbalances where 80% of users watch the same game. This tests your ability to identify and solve hot partition problems.

Hints to consider:

Sharding strategies: by game ID, by league, by time, and implications for hot games
Read replicas and how to route traffic to scale reads for popular games
Rate limiting and request prioritization during overload (premium users vs free users)
Pre-warming caches and scaling up infrastructure before predictable events