P033 Reproducibility and canonical definition of tapped traffic
Can tapped traffic be defined precisely enough (hash boundaries, payload depth) that both parties reproduce identical hashes from replay plus declared state?
Defining tapped traffic so both parties can reproduce its hashes.
Status
Started The author's own early simulations of network-traffic reproducibility, reported and not published, excluding ML non-determinism; no hash-boundary or payload definition has been specified (the note calls an implementable protocol 'a natural next step'); Amodo's SITREP tags network reproducibility 'Not started' and Reynolds & Foster call the payload definition the most important open question.
Can network traffic (excluding ML non-determinism) be reliably reproduced? Early simulation work looks promising but needs further validation.
Posed by
- Cankaya (2026), Research Note: The Fundamentals and Feasibility of Secure Network Taps for Verifying AI Datacenter Use:
meaning that a tap hashing a raw bitstream is likely insufficient, as any “lost” nonce in a preimage makes the hash non-reproducible.
(§3.2; L141; live page, read 2026-10-21)
Also raised in: Cankaya (2026), Catching Misreporting About ML Hardware Use by Turning Noise into Signal, Part II; AI 2040 (2026), Get Involved in Verification; Amodo (2026), AI 2040 Plan A — Verification SITREP; Reynolds & Foster (2026), Network Traffic Hashing.
Sub-questions
| ID | Sub-question | What the sources report |
|---|---|---|
| P033.1 | Has an implementable verification protocol been specified (hash boundaries, sampling strategy, challenge-response, metadata for identification)? | open |
| P033.2 | What verifier storage overhead results from parsing/storing protocol headers, given raw-bitstream hashes are non-reproducible (nonces)? | partly answered |
| P033.3 | What fraction of datacenter-edge bytes (SSH, telemetry, JSON/HTTPS packaging, diagnostics) can be reproduced byte-exactly from replay plus declared state? | open |
| P033.4 | Can a modified stack make repeated runs of the same inference request produce byte-identical frontend packets (or can schemes avoid needing exact packet replication)? | open |
| P033.5 | What fraction of sampled compute-fabric (east-west) tensor hashes can prover re-execution reproduce? | partly answered |
Sub-questions not yet stated precisely enough to check
| ID | Sub-question |
|---|---|
| P033.6 | What exactly is hashed (reconstruction depth / payload definition), defined so both parties can reproduce it? |
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 P033.1: Has an implementable verification protocol been specified (hash boundaries, sampling strategy, challenge-response, metadata for identification)?
For P033.4: Can a modified stack make repeated runs of the same inference request produce byte-identical frontend packets (or can schemes avoid needing exact packet replication)?
Data
- problems.csv — one row per problem
- problem_status.csv — status, scope and evidence per problem
- problem_sources.csv — the posed-by quote and other documents per problem
- leaves.csv — all sub-questions, including those without a checkable form
- leaf_sources.csv — every source row behind each sub-question