Design Facebook Newsfeed - Scale and Constraints

User Scale Analysis

User Base Metrics

Total Registered Users: 3 billion
Daily Active Users (DAU): 2 billion (67%)
Monthly Active Users (MAU): 2.9 billion (97%)
Peak Concurrent Users: 200 million
New User Signups: 2 million per day
User Growth Rate: 10% year-over-year

User Distribution

Geographic Distribution:
- Asia-Pacific: 45% (900M users)
- Europe: 20% (400M users)
- North America: 15% (300M users)
- Latin America: 12% (240M users)
- Rest of World: 8% (160M users)
Platform Distribution:
- Mobile (iOS/Android): 85%
- Web: 15%
User Activity Levels:
- Power users (>10 posts/week): 5% (100M)
- Active users (1-10 posts/week): 25% (500M)
- Casual users (<1 post/week): 70% (1.4B)

Social Graph Metrics

Average Friends: 338 friends per user
Average Page Follows: 40 pages per user
Average Group Memberships: 5 groups per user
Celebrity Users: 100,000 users with >1M friends
Top User: 150M friends (Facebook's own page)

Content Volume Analysis

Post Generation

Total Posts per Day: 500 million
Posts per Second:
- Average: 5,787 posts/second
- Peak: 17,000 posts/second
- Off-peak: 2,500 posts/second
Post Distribution:
- Text only: 30% (150M/day)
- Photos: 50% (250M/day)
- Videos: 15% (75M/day)
- Links: 5% (25M/day)

Engagement Volume

Reactions per Day: 10 billion
Comments per Day: 5 billion
Shares per Day: 2 billion
Reactions per Second: 115,740 RPS average
Comments per Second: 57,870 CPS average

Media Content

Photos per Day: 350 million
Videos per Day: 100 million
Average Photo Size: 2MB (original), 200KB (compressed)
Average Video Size: 20MB (original), 5MB (compressed)
Video Duration: Average 2 minutes

Daily Storage Requirements

Text Posts:
500M posts × 500 bytes = 250GB/day

Photos:
350M photos × 2MB = 700TB/day (original)
350M photos × 200KB = 70TB/day (compressed)

Videos:
100M videos × 20MB = 2PB/day (original)
100M videos × 5MB = 500TB/day (compressed)

Total Daily Storage: ~3.3PB/day
Monthly Storage: ~99PB/month
Yearly Storage: ~1.2EB/year

Traffic Patterns and Load

Read Traffic

Feed Requests: 100 billion per day
- Average: 1,157,407 requests/second
- Peak: 3.5 million requests/second
Post Views: 200 billion per day
Comment Views: 10 billion per day
Profile Views: 5 billion per day

Write Traffic

Post Creation: 500 million per day (5,787 TPS)
Reactions: 10 billion per day (115,740 TPS)
Comments: 5 billion per day (57,870 TPS)
Shares: 2 billion per day (23,148 TPS)

Read:Write Ratio

Overall Ratio: 200:1 (extremely read-heavy)
Feed Reads: 100B vs 500M posts = 200:1
Cache Hit Rate: Target >95% for feed requests

Peak Load Scenarios

Major Events: Elections, Super Bowl, New Year
Traffic Spike: 3-5x normal load
Duration: 2-6 hours sustained peak
Example: New Year generates 5M posts/minute

Performance Targets

Latency Requirements

Feed Load: <1 second p95 (20 posts)
Post Creation: <500ms p95
Reaction: <100ms p95
Comment Load: <300ms p95
Video Playback Start: <2 seconds p95
Image Load: <500ms p95

Throughput Requirements

API Gateway: 4 million requests/second peak
Post Ingestion: 17,000 posts/second peak
Feed Generation: 3.5M feeds/second peak
Database Reads: 10 million queries/second
Database Writes: 200,000 writes/second
Cache Operations: 20 million ops/second

Availability Targets

System Uptime: 99.99% (52 minutes downtime/year)
API Availability: 99.99% for critical endpoints
Data Durability: 99.999999999% (11 9's)
Regional Failover: <10 minutes
Disaster Recovery: RTO <2 hours, RPO <15 minutes

Data Storage Constraints

Database Storage

User Data:
3B users × 5KB per user = 15TB

Post Metadata (5 years):
500M posts/day × 365 × 5 × 500 bytes = 456TB

Social Graph:
3B users × 338 friends × 16 bytes = 16TB

Engagement Data (1 year):
10B reactions/day × 365 × 24 bytes = 88TB/year

Total Database Storage: ~575TB

Media Storage

Photos (5 years):
350M/day × 365 × 5 × 2MB = 1,277PB (original)
350M/day × 365 × 5 × 200KB = 128PB (compressed)

Videos (5 years):
100M/day × 365 × 5 × 20MB = 3,650PB (original)
100M/day × 365 × 5 × 5MB = 913PB (compressed)

Total Media Storage: ~6,000PB
With deduplication (15% savings): ~5,100PB

Cache Storage

Redis Cluster: 100TB for hot data
- Feed cache: 60TB
- User profiles: 20TB
- Post metadata: 10TB
- Session data: 10TB

Search Index Storage

Elasticsearch Cluster: 200TB
- Post index: 150TB
- User index: 30TB
- Page/Group index: 20TB

Network Bandwidth

Ingress Bandwidth

Post Text:
500M posts/day × 500 bytes = 250GB/day = 2.9MB/s

Media Upload:
450M media/day × 2MB = 900TB/day = 10.4GB/s

Total Ingress: ~10.5GB/s average, ~30GB/s peak

Egress Bandwidth

Feed Requests:
100B requests/day × 20 posts × 500 bytes = 1PB/day = 11.6GB/s

Media Downloads:
50B views/day × 2MB = 100PB/day = 1,157GB/s

Total Egress: ~1,200GB/s average, ~3,600GB/s peak

CDN Offload

CDN Hit Rate: 90% for media content
Origin Traffic: 10% = 120GB/s average
CDN Bandwidth: 1,080GB/s average, 3,240GB/s peak

Compute Resources

Application Servers

Total Servers: 200,000 application servers
Server Specs: 32 vCPU, 128GB RAM
Requests per Server: 20,000 requests/second
Auto-scaling: Up to 600,000 servers during peaks

Database Servers

Primary Databases: 5,000 shards
Read Replicas: 25,000 replicas (5 per shard)
Server Specs: 64 vCPU, 512GB RAM, 20TB NVMe SSD
Connections per Server: 2,000 connections

Cache Servers

Redis Nodes: 2,000 nodes
Node Specs: 32 vCPU, 256GB RAM
Cache Capacity: 100TB total
Operations per Node: 100,000 ops/second

Message Queue Servers

Kafka Brokers: 1,000 brokers
Broker Specs: 32 vCPU, 128GB RAM, 40TB SSD
Throughput per Broker: 200MB/s
Total Throughput: 200GB/s

Cost Estimates

Infrastructure Costs (Monthly)

Compute (Application Servers):
200,000 servers × $200/month = $40M/month

Database Servers:
30,000 servers × $500/month = $15M/month

Cache Servers:
2,000 servers × $300/month = $600K/month

Storage (Database):
575TB × $100/TB/month = $57.5K/month

Storage (Media - S3):
5,100PB × $20/TB/month = $102M/month

CDN Bandwidth:
1,080GB/s × 2.6PB/month × $50/TB = $140M/month

Total Monthly Cost: ~$298M/month
Cost per DAU: $298M / 2B = $0.15 per DAU/month

Scaling Bottlenecks

Write Bottlenecks

Post Ingestion: 17,000 TPS peak requires sharding
Fan-out Writes: Celebrity posts cause write amplification
Database Writes: 200,000 writes/second requires partitioning
Solution: Sharding, async processing, write-through caching

Read Bottlenecks

Feed Generation: 3.5M requests/second peak
Hot User Problem: Popular users cause cache stampede
Database Reads: 10M queries/second requires replicas
Solution: Multi-level caching, read replicas, CDN

Storage Bottlenecks

Media Storage Growth: 3.3PB/day requires object storage
Database Growth: 575TB requires sharding and archival
Index Size: Search index grows with content
Solution: Tiered storage, data archival, compression

Network Bottlenecks

Egress Bandwidth: 1,200GB/s requires CDN
Cross-Region Traffic: Global users need regional data centers
Media Delivery: 1,157GB/s requires CDN
Solution: Multi-region deployment, CDN, edge caching

Constraints and Limitations

Technical Constraints

Post Length: 63,206 characters maximum
Photo Limit: 50 photos per post
Video Duration: 240 minutes maximum
File Size: 10MB photos, 4GB videos
API Rate Limits: 200 requests per hour per user
Friend Limit: 5,000 friends maximum

Business Constraints

Content Moderation: <0.3% requiring manual review
Compliance: GDPR, CCPA, COPPA data retention
Uptime SLA: 99.99% availability guarantee
Data Residency: Regional data storage requirements
Cost Optimization: <$0.15 per DAU per month target

Operational Constraints

Deployment Frequency: Multiple deployments per day
Monitoring: <5 minute detection for critical issues
Incident Response: <15 minute response time
Backup Frequency: Continuous replication, hourly snapshots
Disaster Recovery: <2 hour RTO, <15 minute RPO

This comprehensive scale analysis provides the foundation for designing a system that can handle Facebook's massive scale while maintaining performance, reliability, and cost-effectiveness.