Scale & Constraints

📖 5 min read 📄 Part 2 of 10

Video Conferencing System - Scale Constraints

User Scale

Concurrent Users

  • Peak Concurrent Users: 100M+ users simultaneously online
  • Active Meetings: 10M+ concurrent meetings during peak hours
  • Meeting Participants: Up to 1000 participants per large meeting
  • Average Meeting Size: 4.2 participants per meeting
  • Geographic Distribution: Global user base across 6 continents

Growth Projections

  • User Growth: 50% year-over-year growth
  • Meeting Volume: 2x growth during business hours
  • Peak Usage: 10x normal load during global events
  • Seasonal Spikes: 300% increase during conference seasons
  • Regional Variations: 5x difference between peak and off-peak regions

Meeting Scale Constraints

Meeting Sizes

  • Small Meetings: 2-10 participants (80% of meetings)
  • Medium Meetings: 11-50 participants (15% of meetings)
  • Large Meetings: 51-500 participants (4% of meetings)
  • Webinars: 500-1000 participants (1% of meetings)
  • Broadcast Events: 1000+ viewers (special handling)

Meeting Duration

  • Average Duration: 45 minutes per meeting
  • Peak Meeting Hours: 9 AM - 5 PM in each timezone
  • Meeting Overlap: 60% of users in multiple meetings daily
  • Long Meetings: 10% of meetings exceed 2 hours
  • 24/7 Operations: Support for global business operations

Media Processing Scale

Video Processing

  • Video Streams: 200M+ concurrent video streams
  • Resolution Distribution:
    • 720p HD: 60% of streams
    • 1080p FHD: 30% of streams
    • 4K UHD: 5% of streams
    • 480p SD: 5% of streams (low bandwidth)
  • Frame Rates: 15fps to 60fps adaptive
  • Encoding Load: 500K+ concurrent video encoding sessions

Audio Processing

  • Audio Streams: 400M+ concurrent audio streams
  • Audio Quality: 48kHz sampling, 128kbps bitrate
  • Noise Cancellation: Real-time processing for 100M+ streams
  • Echo Cancellation: Sub-20ms processing latency
  • Audio Mixing: Dynamic mixing for group conversations

Screen Sharing

  • Screen Share Sessions: 50M+ concurrent screen shares
  • Update Frequency: 10-30fps for smooth experience
  • Resolution: Up to 4K screen sharing
  • Compression: Real-time screen content compression
  • Annotation: Collaborative markup on shared screens

Network and Bandwidth

Bandwidth Requirements

  • Per Participant:
    • Audio: 50-100 kbps
    • Video (720p): 1-2 Mbps
    • Video (1080p): 2-4 Mbps
    • Screen Share: 1-3 Mbps
  • Total Bandwidth: 500 Tbps+ aggregate bandwidth
  • CDN Traffic: 1 Exabyte+ monthly data transfer

Network Conditions

  • High Bandwidth: Fiber/Cable (>10 Mbps) - 40% of users
  • Medium Bandwidth: DSL/4G (1-10 Mbps) - 45% of users
  • Low Bandwidth: 3G/Satellite (<1 Mbps) - 15% of users
  • Packet Loss: Handle 0-10% packet loss gracefully
  • Jitter: Compensate for 0-100ms network jitter

Storage Scale

Recording Storage

  • Recording Rate: 30% of meetings recorded
  • Storage Growth: 100 PB+ new recordings monthly
  • Retention Policies:
    • Active: 1 year (hot storage)
    • Archive: 7 years (cold storage)
    • Compliance: Permanent (glacier storage)
  • Compression: 10:1 compression ratio for recordings

Metadata Storage

  • Meeting Metadata: 1B+ meeting records
  • User Profiles: 500M+ user accounts
  • Chat Messages: 100B+ chat messages
  • Analytics Data: Real-time metrics and historical data
  • Configuration Data: Meeting settings, preferences, policies

Infrastructure Scale

Server Infrastructure

  • Media Servers: 50K+ media processing servers
  • Signaling Servers: 10K+ WebRTC signaling servers
  • Web Servers: 5K+ application servers
  • Database Servers: 1K+ database instances
  • Cache Servers: 2K+ Redis/Memcached instances

Geographic Distribution

  • Data Centers: 100+ edge locations globally
  • Regions: 20+ primary regions
  • Availability Zones: 60+ zones for redundancy
  • CDN Nodes: 1000+ content delivery nodes
  • TURN Servers: 500+ relay servers for NAT traversal

Performance Constraints

Latency Requirements

  • End-to-End Latency: <150ms for real-time communication
  • Signaling Latency: <50ms for call setup
  • Media Relay Latency: <30ms additional delay
  • Screen Share Latency: <200ms for interactive sharing
  • Chat Message Latency: <100ms delivery time

Processing Constraints

  • CPU Usage:
    • Client: <30% CPU per meeting
    • Server: 80% utilization target
  • Memory Usage:
    • Client: <500MB RAM per meeting
    • Server: 64GB+ RAM per media server
  • GPU Acceleration: Hardware encoding/decoding when available
  • Battery Impact: <20% battery drain per hour on mobile

Scalability Bottlenecks

WebRTC Limitations

  • Browser Connections: Limited concurrent peer connections
  • NAT Traversal: STUN/TURN server capacity
  • Bandwidth Adaptation: Real-time quality adjustment complexity
  • Mobile Performance: Battery and processing constraints

Media Processing

  • Encoding/Decoding: CPU-intensive operations
  • Mixing: Audio/video mixing for large groups
  • Transcoding: Format conversion for different devices
  • Real-time Processing: Low-latency requirements

Network Infrastructure

  • Last Mile: User internet connection quality
  • Global Routing: Optimal path selection
  • Peak Load: Handling traffic spikes
  • Failover: Maintaining quality during outages

Capacity Planning

Growth Modeling

  • User Growth: 50% annual growth rate
  • Usage Growth: 100% growth in meeting minutes
  • Feature Adoption: New features drive 20% usage increase
  • Seasonal Patterns: Plan for 3x peak capacity
  • Regional Expansion: 25% capacity increase per new region

Resource Allocation

  • Compute: 60% media processing, 25% signaling, 15% other
  • Storage: 70% recordings, 20% metadata, 10% cache
  • Network: 80% media traffic, 15% signaling, 5% management
  • Redundancy: 2x capacity for high availability
  • Buffer: 50% headroom for unexpected growth