Roadmap & Latency Benchmarks

1. End-to-End Latency Breakdown (1920×1080 @ 60 FPS)

Measured on Intel Core Ultra 5 135U (1920×1080, zero-copy DXGI + Media Foundation Hardware H.264):

Pipeline Stage	Measured Time	Notes
1. Capture + GPU NV12 Conversion	4.0 ms	DXGI Desktop Duplication + ID3D11VideoProcessor
2. Hardware H.264 Encoding	11.0 ms	MFCreateDXGISurfaceBuffer (QuickSync / NVENC / AMF)
3. QUIC TLS 1.3 Loopback / LAN	0.2 ms	Multiplexed QUIC media stream
4. Client H.264 Decode	6.0 ms	Low-delay frame decode
5. Frame Presentation (winit)	3.3 ms	Direct window surface draw
Total End-to-End Latency	24.5 ms (60 FPS)	Sub-30ms target achieved in Phase 1

---

2. Complete 8-Phase Engineering Matrix

See PROJECTS.md and docs/plan.md for the live tracker:

• Phase 0 (Foundation & Crypto Identity): ✅ Done
• Phase 1 (Sub-30ms Zero-Copy LAN Desktop): ✅ Done
• Phase 2 (Mesh Network, Coordinator & Relay): 🔶 Wired End-to-End
• Phase 3 (QuickSupport, Service, Clipboard & Files): 🔶 Wired End-to-End (#27)
• Phase 4 (Browser WebCodecs Client & Gateway): 🔶 Wired End-to-End (#28)
• Phase 5 (Zero-Trust HuJSON Policy, MFA & .odrec): 🔶 Wired End-to-End
• Phase 6 (Fleet RMM — Inventory, Checks & Scripts): 🔶 Wired End-to-End
• Phase 7 (Virtual Display IddCx, Cross-Platform & Protocols): 🔶 Active R&D (#31, #33)