P029 Passive optical taps at high lane rates

Can a passive optical tap and monitor receiver recover a high-speed optical lane within its link budget while keeping the live link error-free?

Passive optical taps and monitor receivers at 53-112 GBaud.

Terms: GBaud (gigabaud): billions of signal symbols per second. PAM4: a scheme sending two bits per symbol using four levels. FEC (forward error correction): extra data that lets a receiver fix errors. DR4/DR8: standard multi-lane optical link types.

Status

Analyzed Desk link-budget analysis for the live leg: baseline optical margins against tap insertion loss, with feasibility thresholds of 1.5 dB and 0.5 dB, for 400G-class multimode links inside a scalable unit and single-mode links between units; the monitor-leg recovery at 53-112 GBaud is left as the harder problem and the author says amplification is probably indispensable at 112 GBaud. No tap built or measured at this problem's lane rates. Milton & Reynolds (Amodo, 2026) reach a similar desk result ('even a 1 dB loss may be unacceptable') and say they plan testing; commercial passive taps exist for 26 GBaud lanes (400G-DR8), below this problem's lane rates, which are the headline's stated difficulty, so that is a cross-reference and not evidence; vendors' statements that 53 GBaud should work are unverified hearsay. Cankaya's fleet-margin argument cites Meta production telemetry (John et al., DTS 2025), not held.

Intra-SU multi-mode: Passive taps feasible only at 95/5 with direct cabling. Infeasible with standard structured cabling.
Cankaya (2026), Research Note: The Fundamentals and Feasibility of Secure Network Taps for Verifying AI Datacenter Use (Substack), appendix B.5 link-budget feasibility (2026-04-30 (live page read 2026-10-21)), L288

Posed by

Also raised in: Cankaya (2026), Research Note: The Fundamentals and Feasibility of Secure Network Taps for Verifying AI Datacenter Use; Canary Institute research; AI 2040 (2026), Get Involved in Verification; Amodo (2026), AI 2040 Plan A — Verification SITREP; Milton & Reynolds (2026), Network Tapping for AI Verification: A Technical Assessment.

Sub-questions

IDSub-questionWhat the sources report
P029.1Does a passive optical splitter on 53 GBaud lanes (400G-DR4 / 800G-DR8) keep post-FEC error-free operation on both the live and the monitor leg?partly answered
P029.2Can a monitor-leg receiver (APD, preamp or BDFA amplifier) recover a 53 GBaud PAM4 lane from a 5–10% tap port at the pre-FEC threshold?open
P029.3At 106–112 GBaud lanes (800G-DR4 / 1.6T-DR8), does any tap plus (amplified) monitor receiver work within link budget, or is a new ASIC needed?open
P029.4What added insertion loss (dB) can a live single-mode leg tolerate before post-FEC errors or a measurable rise in Annual Interruption Rate across a fleet?open
P029.5Is a passive tap feasible on intra-scalable-unit multimode links (e.g. 400G-SR8) with standard structured cabling, i.e. what margin remains after a 95/5 split with 4 matings?partly answered
P029.6Can photonic samplers (MEMS-mirror packet sampling, fibre striping) replace per-link splitters or 1x256 optical circuit switches at lower cost per sampled port?open
P029.7Can shipping 800G transceivers be firmware-configured to run their lanes at a tappable lower baud rate (e.g. 26 GBd)?open

Sub-questions not yet stated precisely enough to check

IDSub-question
P029.8If sensitive monitor receivers are developed, will operators adopt them to raise lane rates and re-open the tap gap?

Possibly relevant work found by automated search

Found by AI agents searching for this problem's sub-questions, mostly by keyword and citation in the OpenAlex scholarly database and on the web; a second AI pass reviewed each one, and we spot-checked some by hand. They seemed reasonably related, as far as one can tell from outside each paper's field. Some come from neighbouring fields, such as arms control and financial audit, and bear on the problem by analogy. Not reviewed against the status above; a paper listed here does not change that status. 928 of 1052 quotes checked word for word against the paper's text.

For P029.1: Does a passive optical splitter on 53 GBaud lanes (400G-DR4 / 800G-DR8) keep post-FEC error-free operation on both the live and the monitor leg?

For P029.2: Can a monitor-leg receiver (APD, preamp or BDFA amplifier) recover a 53 GBaud PAM4 lane from a 5–10% tap port at the pre-FEC threshold?

For P029.6: Can photonic samplers (MEMS-mirror packet sampling, fibre striping) replace per-link splitters or 1x256 optical circuit switches at lower cost per sampled port?

* Full text not reviewed; more likely to contain errors.

Data