- name
- latency
- description
- Network latency fundamentals
- license
- MIT
- compatibility
- opencode
- metadata
- {"audience":"web-developers","category":"networking"}
## What I do
- Measure and analyze network latency
- Identify latency sources in systems
- Design low-latency architectures
- Optimize data transfer protocols
- Implement caching strategies
- Reduce round-trip times
- Profile application performance
## When to use me
Use me when:
- Diagnosing slow network performance
- Building real-time applications
- Designing distributed systems
- Optimizing user experience
- Understanding network bottlenecks
- Planning capacity and scaling
## Key Concepts
### Latency vs Bandwidth
- **Latency**: Time for single packet (ms)
- **Bandwidth**: Data per second (Mbps)
- They affect performance differently:
- Latency critical for interactive apps
- Bandwidth critical for bulk transfers
### Speed of Light Limits
- Fiber: ~200,000 km/s (2/3 c)
- Theoretical minimum latency = distance / speed
- Example: NYC to London ~70ms minimum
### Latency Components
```
Total Latency =
┌─────────────────┐
│ Processing │ (routers, switches)
│ + Serialization │ (bits on wire)
│ + Propagation │ (distance/speed of light)
│ + Transmission │ (bandwidth limits)
│ + Queuing │ (congestion)
└─────────────────┘
```
### Measurement Tools
- **ping**: RTT to host
- **traceroute**: Path and per-hop latency
- **WebPageTest**: Full page load analysis
- **Chrome DevTools**: Network timing
- **APM tools**: Application latency tracking
### Common Latency Targets
| Application | Target Latency |
|-------------|----------------|
| Web (initial) | < 2 seconds |
| API calls | < 200ms |
| Real-time (gaming) | < 50ms |
| Financial trading | < 1ms |
| VoIP | < 150ms |
Voir sur GitHub