Skip to content

MULTI_WORKER_SETUP.md

Latest commit

 

History

History
151 lines (110 loc) · 3.87 KB

MULTI_WORKER_SETUP.md

File metadata and controls

151 lines (110 loc) · 3.87 KB

Multi-Worker MediaSoup Implementation

Overview

This implementation creates multiple MediaSoup workers to distribute room load across CPU cores, improving performance and scalability.

Key Features

1. Multiple Workers

  • Configurable worker count via MEDIASOUP_WORKER_COUNT (default: 4)
  • Each worker gets isolated RTC port ranges (1000 ports per worker)
  • Worker 0: 10000-10999, Worker 1: 11000-11999, etc.

2. Intelligent Load Balancing

  • Room Count Priority: Distributes rooms evenly across workers
  • CPU Usage Monitoring: Considers worker CPU usage when selecting
  • Last Used Time: Round-robin for workers with similar load

3. Resource Monitoring

  • Real-time CPU usage tracking per worker
  • Room count tracking per worker
  • Worker health monitoring

4. API Endpoints

  • GET /stats - Complete system overview with room distribution
  • GET /workers - Worker-specific statistics

Configuration

Environment Variables

# Number of MediaSoup workers (default: 4)
MEDIASOUP_WORKER_COUNT=4

# Base IP for WebRTC (default: 127.0.0.1)
WEBRTC_LISTEN_IP=127.0.0.1

# Announced IP for external connections
ANNOUNCED_IP=your.public.ip

Port Allocation

Each worker gets a dedicated port range:

  • Worker 0: 10000-10999
  • Worker 1: 11000-11999
  • Worker 2: 12000-12999
  • Worker 3: 13000-13999

Load Balancing Algorithm

When creating a new room, the system:

  1. Checks room count - Selects worker with fewest rooms
  2. Evaluates CPU usage - If room counts are equal, chooses worker with lower CPU
  3. Uses round-robin - If CPU usage is similar, uses least recently used worker

Benefits

Performance

  • CPU Distribution: Spreads load across multiple cores
  • Memory Isolation: Each worker has isolated memory space
  • Better Fault Tolerance: One worker failure doesn't affect others

Scalability

  • Horizontal Scaling: Can easily increase worker count
  • Resource Monitoring: Real-time visibility into system load
  • Dynamic Load Balancing: Automatically adapts to changing conditions

Monitoring

Real-time Statistics

# Get complete system overview
curl http://localhost:3000/stats

# Get worker-specific information
curl http://localhost:3000/workers

Example Response

{
  "workers": [
    {
      "workerId": 0,
      "roomCount": 3,
      "cpuUsage": 0.25,
      "pid": 12345,
      "closed": false
    }
  ],
  "distribution": {
    "0": ["room1", "room2", "room3"],
    "1": ["room4", "room5"]
  },
  "totalRooms": 5
}

Implementation Details

Files Modified

  1. services/mediasoup.ts

    • Multi-worker initialization
    • Load balancing logic
    • Resource monitoring

  2. services/room.ts

    • Worker-aware room creation
    • Room-to-worker mapping
    • Cleanup handling

  3. index.ts

    • Monitoring endpoints
    • Statistics API

Key Functions

  • initMediasoup() - Initializes all workers
  • getLeastLoadedWorker() - Selects optimal worker
  • getWorkerStats() - Returns worker statistics
  • getRoomDistribution() - Shows room distribution across workers

Best Practices

Production Deployment

  1. Set worker count to match CPU cores: MEDIASOUP_WORKER_COUNT=8
  2. Configure firewall for port ranges: 10000-18000
  3. Monitor worker health using /stats endpoint
  4. Use process manager (PM2) for automatic restarts

Scaling Recommendations

  • Small servers (2-4 cores): 2-4 workers
  • Medium servers (8-16 cores): 4-8 workers
  • Large servers (32+ cores): 8-16 workers

Network Configuration

Ensure firewall allows:

  • TCP ports: Base port range (10000 + worker_count * 1000)
  • UDP ports: Same range as TCP
  • HTTP/WebSocket: 3000 (or configured PORT)

This implementation provides a robust foundation for scaling MediaSoup applications across multiple CPU cores while maintaining optimal performance and reliability.
⚡