| Internet-Draft | MICA-NEX Measurement Ingestion | October 2026 |
| Watts | Expires 25 April 2027 | [Page] |
This document describes a narrow, testable next step in the MICA-NEX proposal: cryptographically binding nanoscale instrument measurement claims to externally provisioned acquisition and calibration signer roles while denying model planners the ability to self-certify measurement status. The reference implementation uses Ed25519 signatures, SHA-256 content binding, typed units, time windows, source-kind separation, and a persistent local replay ledger. In a hardware-disconnected test harness, 60/60 signed synthetic test records from three simulated instrument classes were admitted and 30/30 adversarial cases were rejected for the expected reason; 28/28 Python tests passed. No instrument, laboratory calibration, external reviewer, clinical context or independent physical replication was involved. Signature validity is deliberately distinguished from scientific validity.¶
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶
This Internet-Draft will expire on 25 April 2027.¶
Copyright (c) 2026 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶
The prior MICA-NEX v0.1 gate [MICANEX01] accepts only SIMULATED records, so it cannot truthfully ingest MEASURED claims without a new interface. This conflict was logged as EM-005 in EXP-01 v0.2. Our question is whether one strictly typed, model-neutral acquisition-ingestion contract can accept records across scanning-probe, microfluidic-fluorescence and electrochemical categories while rejecting spoofed measurement authority and preserving evidence classification.¶
Do not confuse a signature authenticating a claim with a verified physical measurement or traceable calibration. NIST defines metrological traceability via a documented calibration chain contributing to uncertainty [NIST-TRACEABILITY]. This implementation checks a declaration and its file digest, not the validity of any real-world calibration chain.¶
This Internet-Draft is a faithful conversion of the author's research report "MICA-NEX v0.3: Authenticated Measurement Ingestion for Model-Agnostic Nano-Instrument Agent Swarms -- A hardware-disconnected, falsifiable software research experiment", Version 0.3, dated 20 October 2026. Mathematical notation and the table have been rendered in plain text. It is published for discussion; it is not a standard.¶
The agent coordinator issues an unprivileged read-only request; separate laboratory holders would sign acquisition and calibration data outside the agent runtime. The central verifier binds the signed claim to bytes and prescribed instrument semantics.¶
Untrusted planner -> deterministic policy
-> independent operator-record source
-> signed/typed ingestion -> evidence receipt
-> independent scientific review
No actuation path in this prototype; no AI-owned signing keys.
¶
Three unit profiles are fixed: scanning-probe length in nm, microfluidic concentration in nM, and electrochemical potential in V. These units are illustrative for the narrow test, not a universal dimensional ontology.¶
Let R be raw acquisition bytes, C be the calibration document, A the signed acquisition claim, and K_A, K_C separately provisioned public keys. Admission requires:¶
Verify_{K_A}(A, sigma_A) = 1
Verify_{K_C}(C_claim, sigma_C) = 1
SHA256(R) = A.raw_sha256
SHA256(C) = C_claim.evidence_sha256
unit(A) = unit(C_claim) = U_i
t_cal,start <= t_A <= t_cal,end
¶
Ed25519 is specified by RFC 8032 [RFC8032]. The software uses project-specific sorted-key JSON with restrictive value types, labeled MICA-NEX-CJ/1. It does not claim RFC 8785 compliance [RFC8785]; cross-language canonicalization is deferred. Duplicate input JSON member keys are refused.¶
The post-signature state is one of:¶
Neither positive case can establish independent physical validation. The distinction resembles the separation between provenance activities, agents, and entities in the W3C PROV model [W3C-PROV], without claiming implementation of PROV-O.¶
Ephemeral Ed25519 keys are created by a test harness in memory and never packaged as laboratory credentials. A deterministic trial runner constructs 20 synthetic acquisition records per substrate plus ten negative-control cases per substrate. It calls a model-neutral policy coordinator and a common ingestion verifier. The test harness does not access an actual microscope, optical sensor or potentiostat. Units, hashes, state transitions, and SQLite replay protection are exercised on synthetic bytes only.¶
| Test class | Passed | Total | Interpretation |
|---|---|---|---|
| Signed TEST_FIXTURE records admitted | 60 | 60 | Software mock only |
| Adversarial cases correctly rejected | 30 | 30 | Synthetic negatives |
| Local unit tests | 28 | 28 | Python unittest |
| Physical equipment connections | 0 | 0 | NOT RUN |
| Independent third-party reproductions | 0 | 0 | NOT RUN |
Negative scenarios include an agent-held forged key, false evidence-state promotion, changed raw or calibration bytes, expired calibration, inappropriate unit, stale record, forbidden physical effect, replay, and undeclared fields. All accepted records were labeled as test fixtures; no physical observation was synthesized into a scientifically verified one.¶
Software falsifier: Show any unauthorized or altered record accepted by exactly this frozen gate and verified test conditions, or a test fixture classified as physically validated. Physical falsifier: When actually acquiring from three distinct calibrated instruments, show an irreducible schema/semantics change or a safety boundary failure needed for one class. These are prospective experiments, not completed results.¶
The current code does not isolate a verifier OS process or protect a SQLite ledger from a privileged actor, provide HSM-backed keys or revocation, audit calibration traceability, verify vendor instrument identity, or enforce independent hardware interlocks. It does not establish that its components are novel compared with existing middleware, digital signatures, or provenance standards. There is no clinical or nano-robotic test and no evidence of energy improvement. Signed external review remains NOT RUN; the separate reviewer-signature checker cannot establish social or procedural reviewer independence.¶
Freeze the v0.3 contract, obtain a separately authorized laboratory and reviewer, pre-register independently produced expected results, then collect signed records from genuinely calibrated, independently operated scanning-probe, fluorescence and electrochemical measurement instruments. Preserve original raw bytes, calibration chain, reviewer-controlled key metadata, environment and signed replication report. Test real-world precision, uncertainty, source authentication, and instrument-operation safety independently of the software gate.¶
Status: SOFTWARE SYNTHETIC PASS; PHYSICAL INTEROPERABILITY NOT RUN; INDEPENDENT REPLICATION NOT RUN; FIRST-IN-WORLD NOVELTY NOT ESTABLISHED.¶
Project PI and attribution: Deonte O'dell Watts, ORCID 0009-0005-8586-3650, GoodShyt Group Inc. AI-assisted design, implementation, literature checking and drafting; test authorship is not independent. Attribution here is not a legal declaration of inventorship or patent priority. Project concept first documented October 19, 2026.¶
This document has no IANA actions.¶
A valid signature proves only control of a signing key and integrity of the signed bytes under the stated trust assumptions; it does not prove that a measurement is correct, calibrated, traceable or physically valid. Implementations derived from this design need to keep signing keys out of reach of planning agents and of any process that can alter the trust store, the replay ledger or the trusted clock, because any of these would let an agent self-certify measurement status. The prototype's replay ledger is a local SQLite file that a privileged actor can modify; it offers no HSM-backed key custody and no key revocation. The project-specific canonical JSON (MICA-NEX-CJ/1) is not RFC 8785; cross-implementation canonicalization differences can cause verification failures or, if handled loosely, signature confusion, which is why duplicate JSON member names are rejected. The prototype has no actuation path; any future connection to physical instruments requires independent hardware interlocks and safety review outside the software gate.¶