Summary
Ploss parses mtr (my traceroute) output to detect per-hop packet loss and generates ISP-ready evidence files.
Real test results on public DNS operated by Google (8.8.8.8):
- Hop 4: 81.228.84.95 shows 70.0% loss
- Final hop: 8.8.8.8 shows 0.0% loss
- Total path loss: 5.8%
Core contributions (Sole Creator)
Verified Components
- mtr output parsing for per-hop loss percentages
- Bottleneck hop identification by highest loss
- JSON structured output + CSV spreadsheet export
- Live monitoring capability using the (–watch flag)
- Zero runtime dependencies, cross-platform
The Problem
Packet loss localization requires identifying specific hops/routers or rate limits. CLI tools in this domain typically provide raw data only. Ploss adds structured analysis and ISP-ready export formats while maintaining zero dependencies.
Validation Results
Test configuration: 8.8.8.8 from Malmö, Sweden
Total hops: 12
Total loss: 5.8%
Average latency: 9.9ms
Worst hop: #4 = 70.0% loss (81.228.84.95)
Final validation: 8.8.8.8 = 0.0% loss
Files generated:
- proof.json (structured diagnostics)
- hops.csv (Excel analysis)
Platforms tested:
- macOS (primary development)
- Linux compatible
Technical Implementation
Language: Python 3 Core algorithm: mtr parsing → loss% per hop → bottleneck analysis Output formats: Console summary + JSON + CSV CLI interface: ploss [target] [–mtr] [–json] [–csv] [–watch] Binary size: 15KB, no runtime dependencies
Planned extensions:
- Cloud provider endpoint testing
- Multi-target path comparison
- Agent deployment for enterprise scale