800 G - OSFP
OSFP — 800G. The 800G workhorse — eight lanes of 100G PAM4 for AI/ML and hyperscale.
Overview
Octal Small Form-factor Pluggable, OSFP MSA. The primary 800G transceiver form factor — eight electrical lanes at 100G PAM4 each. Physically wider than QSFP / QSFP112; not mechanically interchangeable. Used in AI/ML training clusters (NVIDIA Quantum-2 InfiniBand, Spectrum-X Ethernet), hyperscale spines, and 800G coherent DCI. CMIS 5.x management. QSFP-DD800 is an alternative 800G form factor on select hyperscale platforms.
Key specifications
| Speed | 800G |
|---|---|
| Electrical interface | 8 lanes × 53.125 Gbaud PAM4 (100G per lane) |
| Signalling | PAM4 with KP4 / RS-FEC |
| FEC | RS-FEC (544,514) required; coherent ZR uses OpenZR+ MSA FEC layer |
| Typical power draw | 15 W – 20 W typical; up to 23 W for coherent ZR variants |
| Connectors | MPO-16 (SR8, DR8), LC duplex (LR8, ER8, ZR), OSFP DAC / AOC |
| MSA / standards | OSFP MSA rev 5.x (mechanical); CMIS 5.0 / 5.1 (management); IEEE 802.3df / 802.3dj (800G PHY) |
| Introduced | 2022 (OSFP MSA 5.0), 2024 (IEEE 802.3df 800G) |
Available variants
| Part number | Standard | Wavelength | Fiber | Reach | Notes |
|---|---|---|---|---|---|
NAP-OSFP-SR8 |
800GBASE-SR8 | 850 nm | MM OM4 / OM5 | 100 m (OM4) / 150 m (OM5) | MPO-16, 8 parallel |
NAP-OSFP-DR8 |
800GBASE-DR8 | 1310 nm | SM parallel | 500 m | MPO-16 |
NAP-OSFP-LR8 |
800GBASE-LR8 | LAN-WDM 8λ | SM | 10 km | LC duplex |
NAP-OSFP-ER8 |
800GBASE-ER8 | LAN-WDM 8λ | SM | 40 km | Extended reach single-mode |
NAP-OSFP-800ZR |
800G coherent ZR | C-band | SM | 80–120 km | DSP-based, no amplifier |
NAP-OSFP-800ZR-PLUS |
800G ZR+ (OpenZR+) | C-band | SM | 500 km+ with amps | Carrier-grade DCI |
Coded for these platforms
Every NetAPI OSFP module is OEM-coded for the target switch and validated against current production firmware. Common targets:
- NVIDIA Quantum-2 InfiniBand (NDR 400G / NDR 800G)
- NVIDIA Spectrum-4 SN5000 series
- Arista 7060X6, 7800R3 800G line cards
- Cisco Silicon One 800G platforms
- Juniper PTX10000 series
- Hyperscale custom silicon (Meta, Google, ByteDance)
Recommended for
- NVIDIA Quantum-2 / Spectrum-X AI training fabric
- Hyperscale 800G spines
- Disaggregated NVMe-oF storage at 800G
- Long-haul DCI with coherent 800ZR / ZR+
Engineering notes
- OSFP is not backward-compatible with QSFP-DD800 — different mechanical envelope. Verify your switch chassis datasheet before ordering.
- OSFP modules can draw 15–20 W; coherent ZR up to 23 W. Switch per-cage power and chassis thermal budget are real constraints.
- 800G ZR coherent modules implement OpenZR+ MSA or 800ZR standards. Per-lane BER and pre-FEC error counters are exposed through CMIS for telemetry pipelines.
- AI/ML training clusters increasingly bias toward AOC and DR8 at 800G — link reach within a row, low BER, predictable latency.
Need a coding or reach you don't see? NetAPI custom-codes against the full Cisco, Juniper, Arista, Dell, HPE, Aruba, Palo Alto, Huawei, and NVIDIA OUI sets. Contact us with your switch model and we'll quote.