Conjunction-proof deep space communications infrastructure. "Starlink, but Heliocentric" — the original 2026 proposal that became a validated research program with 11 papers and 298 automated tests.
Note (March 2026): This repository contains the original architectural proposal that launched the TIN research program. For the full DTN simulation engine, sparse-law analysis, and planetary relay validation across 11 bodies (Mercury to Titan), see the canonical TIN repository.
Every 26 months, the Sun blocks all communication between Earth and Mars for 2–8 weeks. TIN eliminates that blackout with 5–6 satellites: 3 in polar heliocentric orbits and 2–3 at Sun-Earth Lagrange points. The worst-case relay adds only 6 minutes of latency — replacing weeks of silence.
Visualizations:
- 🌐 Original architecture — interactive 3D orbital visualization
- 🪐 Heliocentric orrery — animated full-system DTN architecture (2026)
📊 Supporting data — CSVs, technical audit, and coverage analysis 🔬 Full research repo — DTN simulator + percolation engine + sparse law validation
| Satellites | 5–6 (3 polar relay + 2–3 Lagrange AI hubs) |
| Availability | 99.9%+ (conjunction-proof) |
| Optical Bandwidth | 1–10 Gbps per link |
| Latency | 3–27 light-minutes (one-way) |
| Deployment | Phase 0 lunar relay ~2029, full network ~2032 |
| Cost | ~$2.5–3B (realistic mid-range) |
v0.3.1 (March 2026) — This proposal evolved into a full research program at github.com/toxic2040/TIN:
- DTN simulation engine (custody FSM, RW-CGR routing, fragment aggregation)
- Percolation engine spanning 11 planetary bodies (Mercury to Titan)
- Sparse-law validation: DR = S_T × η across 35,000+ configurations
- Mars deep-space architecture with L4/L5 relay modeling
- 11 papers on Zenodo, 298 automated tests, MIT licensed
See CHANGELOG for full version history.
Open proposal. Not affiliated with SpaceX, NASA, or any agency. Offered in the public interest.
For technical implementation, simulation results, and research papers, see TIN canonical repository and the Zenodo collection (11 papers).
Feedback, technical critique, and collaboration inquiries welcome. Open an issue or reach out.