Internet-Draft TrackOne VTL March 2026
El Khatabi Expires 15 September 2026 [Page]
Workgroup:
Independent Submission
Internet-Draft:
draft-elkhatabi-verifiable-telemetry-ledgers-01
Published:
Intended Status:
Informational
Expires:
Author:
B. El Khatabi
TrackOne Project

Verifiable Telemetry Ledgers for Resource-Constrained Environments

Abstract

This document specifies a verifiable telemetry ledger profile for resource-constrained sensing environments. The profile defines how a gateway accepts framed telemetry, applies anti-replay policy, projects accepted frames into canonical facts, builds deterministic daily Merkle commitments, and anchors daily artifacts with external timestamp proofs. OpenTimestamps (OTS) is the default anchoring mechanism; optional parallel attestation methods (RFC 3161 timestamp protocol and peer signatures) are also described.

The goal is interoperability and independent auditability, not new cryptographic primitives. Successful verification is limited to the disclosed artifacts and claimed disclosure class; it does not by itself establish completeness of all observed frames, physical truth of measurements, or safety for autonomous actuation.

Status of This Memo

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 15 September 2026.

▲

Table of Contents

1. Introduction

Long-lived telemetry deployments such as environmental monitoring, heritage conservation, and infrastructure health need evidence that measurements were not silently altered after collection. A coastal sensor array that reports temperature every six hours over a five-year deployment produces tens of thousands of records; stakeholders such as regulators, researchers, and insurers may need to verify, years later, that the data they rely on is the same data the sensors produced.

Existing standards provide important building blocks:

However, none of these building blocks defines a complete operational profile for the constrained case: devices with intermittent uplinks, limited compute budgets, and no persistent Internet connectivity at the collection point. SCITT assumes a transparency service is reachable for receipt-oriented workflows. COSE Merkle proofs define proof encodings, but not batching policy or anchoring lifecycle. RATS addresses device identity attestation, but not telemetry commitment.

The discussion below compares this profile directly with [SCITT] and [COSE-MERKLE] because those documents define adjacent, but not identical, transparency and proof disclosure models.

This document fills that gap. It specifies a practical profile that combines these building blocks for low-power telemetry systems:

  1. Emit encrypted framed telemetry.
  2. Ingest and validate frames with anti-replay.
  3. Project accepted frames into canonical facts.
  4. Build deterministic daily Merkle commitments.
  5. Chain days with previous-day root linkage.
  6. Anchor the day artifact using external timestamp proofs.
  7. Verify independently of disclosed artifacts.

The profile is informed by TrackOne pre-production validation, including constrained uplink simulations, hardware-in-loop testing, and daily batching and verification workflows.

Verification claims in this profile are intentionally scoped. A successful result means that the disclosed artifacts are internally consistent, that the authoritative day artifact was bound correctly to the disclosed proof material, and that the checks reported by the verifier succeeded. It does not prove that every observed frame was committed, that the measurements are physically correct, or that any downstream operational decision is safe.

1.1. Relationship to Existing Work

This document is complementary to, not a replacement for, the following work:

  • SCITT ([SCITT]): SCITT defines architectures for transparent, append-only logs of signed statements with receipts. This profile differs in that it is optimized for disconnected, daily-batch operation where a transparency service is not assumed to be continuously reachable.
  • COSE Merkle Tree Proofs ([COSE-MERKLE]): COSE-MERKLE defines proof encodings. This profile defines a batching and commitment contract and may use COSE-based proof encodings in a future revision.
  • RATS ([RFC9334]): This profile explicitly defers device identity, attestation, and key lifecycle to deployment-specific mechanisms.
  • RFC 8949 ([RFC8949]): This profile defines a TrackOne deterministic CBOR commitment profile for gateway-side commitments. It does not attempt to define a general-purpose CBOR profile beyond the commitment path described here.

2. Conventions and Terminology

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.

Terms:

3. System Roles

In this document, optional parallel attestation channels are not required for baseline conformance. A conforming deployment MUST implement at least one anchoring channel, with OTS as the default channel described here. RFC 3161 timestamp responses and peer signatures MAY be absent entirely; when present, they MUST bind to the same authoritative day-artifact digest and verifiers MUST report their results separately.

4. Data and Commitment Model

This section uses the current TrackOne transport profile to explain the reusable commitment contract. The most deployment-specific part is the framed wire transport in Section 4.1; the more reusable interoperability surface is the deterministic frame-to-fact projection, commitment encoding, Merkle calculation, disclosure classes, and anchoring and verification behavior.

4.1. Frame Contract

A frame is transported as NDJSON with fields:

{
  "hdr": { "dev_id": 101, "msg_type": 1, "fc": 42, "flags": 0 },
  "nonce": "base64-24B",
  "ct": "base64-ciphertext",
  "tag": "base64-16B"
}
Figure 1

Gateways MUST validate header field presence, header ranges, and AEAD authentication before fact emission.

  • hdr.dev_id MUST be an unsigned integer in the range 0..65535.
  • hdr.msg_type MUST be an unsigned integer in the range 0..255.
  • hdr.fc MUST be an unsigned integer in the range 0..(2^32-1).
  • hdr.flags MUST be an unsigned integer in the range 0..255.
  • nonce MUST be base64 text that decodes to exactly 24 bytes.
  • tag MUST be base64 text that decodes to exactly 16 bytes.

Frames that fail parse, range, or AEAD validation MUST be rejected before fact commitment and MUST NOT produce committed facts.

The wire shape in this section is the current TrackOne transport profile. Other deployments MAY use a different transport framing if they preserve the acceptance semantics needed to produce the same canonical fact objects under Section 4.3.

4.2. Anti-Replay Admission Criterion

For framed telemetry, the replay unit is (dev_id, fc). A gateway MUST commit a fact from a frame only if all of the following hold:

  1. No fact has already been committed for the same replay unit (dev_id, fc).
  2. The frame counter fc is within the configured acceptance window for that device relative to durable replay state.
  3. No continuity-break, reset, or replay-state-loss condition requires explicit resynchronization before further acceptance.

Frames that do not satisfy these conditions MUST NOT produce committed facts.

The RECOMMENDED default acceptance window is 64 frame counter values per device. Frames with fc more than window_size behind the highest accepted counter for a device MUST be rejected. In the current TrackOne gateway profile, frames with fc more than window_size ahead of the highest accepted counter MUST also be rejected, rather than silently accepted across a large discontinuity.

Structured Rejection Evidence

Gateways SHOULD produce structured rejection evidence for rejected frames. A rejection record SHOULD include at minimum:

  • device_id,
  • fc (or null if unavailable),
  • reason,
  • observed_at_utc, and
  • frame_sha256.

Rejection evidence is an audit artifact and MUST NOT be hashed into ledger commitments.

Replay State Persistence

Gateways SHOULD persist replay state across restart. If replay state is lost, gateways SHOULD record a continuity break event and SHOULD NOT silently re-accept counters that could already have been committed.

Scope Limitation

This profile provides tamper-evidence for committed facts. It does not prove the absence of selective pre-commitment drops by a malicious gateway.

4.3. Frame-to-Fact Projection

The frame-to-fact projection transforms a validated, decrypted frame into a canonical fact suitable for commitment. The committed fact is a structured record derived from the decrypted payload, not a copy of transport bytes.

  1. The gateway MUST verify AEAD authentication.
  2. The gateway MUST decrypt the ciphertext.
  3. The gateway MUST parse the decrypted plaintext according to the frame's msg_type.
  4. The gateway MUST construct a fact object.
  5. The gateway MUST serialize that fact object under the gateway commitment profile.
  6. The canonical fact bytes are then hashed for Merkle inclusion.

A committed fact under this profile MUST contain at minimum:

  • pod_id,
  • fc,
  • ingest_time,
  • pod_time,
  • kind, and
  • payload.

Interoperability depends on two distinct layers: the projection contract that maps an accepted frame into a fact object, and the commitment profile that serializes that fact object and reduces the resulting digests. Independent implementations claiming the same result MUST agree on both layers.

A commitment profile or deployment profile claiming conformance MUST fix the type and source of each committed field so independent implementations can construct identical fact bytes.

For each committed field, the applicable projection contract MUST state whether the field is transport-derived, payload-derived, gateway-assigned, or deployment metadata-derived. Independent implementations MUST NOT substitute a different field source while claiming the same projection contract.

For trackone-canonical-cbor-v1, the committed fact object is a CBOR map whose required top-level fields are pod_id, fc, ingest_time, pod_time, kind, and payload. The TrackOne-specific field definitions are:

  • pod_id MUST be a deterministic pod identifier. When framed telemetry is used, the profile MUST define a deterministic mapping from hdr.dev_id or equivalent deployment alias to pod_id. The current TrackOne reference profile uses a lowercase 8-byte hexadecimal pod identifier such as 0000000000000065.
  • fc MUST be the accepted frame counter as a non-negative integer.
  • ingest_time MUST be a UTC integer timestamp fixed by the projection contract.
  • pod_time MUST be either a pod-supplied integer timestamp or null when the pod does not supply one.
  • kind MUST identify the fact family under the applicable projection contract.
  • payload MUST be the parsed plaintext object associated with the accepted frame.

A deployment profile MAY carry additional top-level fields such as ingest_time_rfc3339_utc or signature, but any such field becomes part of the committed fact object when present. Published conformance vectors for a commitment profile MUST therefore fix both the presence and exact value of any optional committed fields.

An implementation claiming parity with trackone-canonical-cbor-v1 MUST reproduce that exact post-projection fact object before applying the deterministic CBOR rules in Section 4.4.

Ciphertext, raw transport bytes, and the authentication tag MUST NOT be part of the committed fact object. The exact payload schema is deployment-specific; the deterministic projection contract is the normative requirement and MUST be published for any commitment profile that claims interoperability.

Published conformance vectors for a commitment profile MUST include the post-projection fact objects used as commitment inputs, not only transport frames.

4.4. Deterministic CBOR Commitment Encoding

This section does not define a new general-purpose CBOR variant. It records the narrow deterministic CBOR encoding used for commitment bytes in the current TrackOne gateway and ledger implementation. The identifier trackone-canonical-cbor-v1 names this commitment recipe so verifiers can tell which byte-level rules were used.

The authoritative commitment artifacts, namely CBOR fact artifacts and the canonical day artifact, use a constrained subset of deterministic encoding under Section 4.2.1 of [RFC8949]. For TrackOne commitment bytes, the following concrete choices apply:

  • All commitment-path items MUST use definite-length encoding.
  • Integers MUST use the shortest encoding width permitted by [RFC8949].
  • Map keys MUST be CBOR text strings.
  • Map keys MUST be sorted by encoded key length ascending, then by lexicographic order of the encoded key bytes.
  • Finite floating-point values MUST be encoded using the shortest of float16, float32, or float64 that exactly preserves the value.
  • NaN, positive infinity, and negative infinity MUST be rejected in commitment paths.
  • CBOR tags MUST NOT appear in commitment bytes.
  • Supported values are unsigned integers, negative integers, byte strings, text strings, arrays, maps, booleans, null, and deterministic finite floats.

Implementations MUST NOT accept generic CBOR serializers as authoritative commitment encoders. An encoder is acceptable only if it yields the same bytes as these rules.

JSON projections of fact artifacts and day artifacts are optional and non-authoritative. They MUST NOT be used as commitment inputs. When produced, such projections SHOULD follow [RFC8785].

Device-side or embedded components MAY use other internal encodings, including different deterministic CBOR layouts optimized for local constraints. Those encodings are not the authoritative commitment encoding described here unless they are explicitly identified by a distinct commitment_profile_id and verified under their own rules.

4.5. Deterministic Commitment Tree Calculation

For a given day D, the current commitment profile computes a daily root from the canonical fact commitment bytes produced under Section 4.4. The following steps describe that calculation.

Leaf Digests

  • Each canonical fact byte string is hashed with SHA-256, yielding a 32-byte leaf digest.

Digest Ordering

  • To make the daily root independent of file order or ingest order, leaf digests MUST be sorted in ascending byte order before reduction.
  • Lowercase hexadecimal is a representation format for artifacts and examples only; internal Merkle computation operates on raw hash bytes.
  • Sorting by lowercase hexadecimal is equivalent to bytewise ascending order over the raw digests.

Pairwise Reduction

  • The sorted digests are reduced pairwise by computing SHA-256(left_child_bytes || right_child_bytes), where both operands are raw 32-byte digests.
  • If a layer has an odd number of digests, the final digest is duplicated to form the last pair.
  • The current commitment profile does not prepend domain-separation bytes to leaf or parent hashes.

Empty Day

  • If no facts are committed for the day, the daily root is the SHA-256 digest of zero bytes: e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855.

The resulting daily root is deterministic for the set of committed facts. Because the leaf digests are sorted before reduction, the result depends on the committed fact set rather than on ingestion order.

Any future change to this calculation that alters commitment bytes (for example, adding domain separation) MUST use a new commitment_profile_id.

4.6. Day Artifact Schema

The authoritative day artifact is a CBOR-encoded day record produced under Section 4.4. The day record contains the following fields:

  • version (uint): day-record schema version, currently 1.
  • site_id (tstr): site identifier.
  • date (tstr): UTC day label in YYYY-MM-DD form.
  • prev_day_root (tstr): previous day root as 64 lowercase hexadecimal characters.
  • batches (array): array of batch objects.
  • day_root (tstr): deterministic day root as 64 lowercase hexadecimal characters.

Each batch object contains:

  • version (uint): batch-record schema version, currently 1.
  • site_id (tstr): site identifier.
  • day (tstr): UTC day label in YYYY-MM-DD form.
  • batch_id (tstr): batch identifier.
  • merkle_root (tstr): batch Merkle root as 64 lowercase hexadecimal characters.
  • count (uint): number of committed facts in the batch.
  • leaf_hashes (array of tstr): sorted leaf hashes as lowercase hexadecimal strings.

The batch objects embedded in the day artifact are the authoritative batch metadata for verification.

  • For each batch object, count MUST equal the length of leaf_hashes.
  • For each batch object, merkle_root MUST equal the Merkle reduction of that batch's leaf_hashes under Section 4.5.
  • The multiset union of all batch leaf_hashes for the day MUST equal the day's leaf digest multiset from which day_root is computed.

day/YYYY-MM-DD.cbor is authoritative. The corresponding day/YYYY-MM-DD.json file is a projection only.

This document uses normative field tables rather than CDDL. A future revision may add a formal CDDL appendix if broader independent implementations require it.

4.7. Day Chaining

Day records include prev_day_root.

  • The genesis day for a site MUST set prev_day_root to 64 ASCII zero characters, representing 32 zero bytes.
  • Non-genesis days MUST set prev_day_root to the previous committed day's day_root.

Because day labels are UTC-based, chaining semantics are also defined on UTC day boundaries.

5. Artifacts and Verification Bundles

Illustrative artifact layout:

Deployments MAY store artifacts differently and MAY export them as bundles. The path shapes above are illustrative.

Standalone block metadata files are convenience projections of batch objects already carried in the authoritative day artifact. They are not additional commitment inputs. Implementations that claim standalone block-metadata validation MUST compare disclosed projections with the corresponding batch object in the authoritative day artifact and reject mismatches. Current TrackOne verifier behavior does not yet enforce full block-metadata parity on every manifest-absent path, so successful verification on those paths MUST NOT be read as proving complete block projection equality beyond the checks the verifier actually performs.

Every verification bundle MUST disclose the commitment_profile_id. A bundle SHOULD disclose it in a machine-readable verification manifest. On paths where no standalone manifest is present, the current TrackOne verifier still depends on verifier configuration or equivalent out-of-band context to determine the expected profile identifier.

A standalone verification manifest is RECOMMENDED, but not every supported producer path emits one yet. Manifest-absent bundles remain verifiable on current tooling paths when the artifact set is otherwise sufficient for the claimed disclosure class and the verifier is given the expected commitment profile by configuration or equivalent local policy. Deployments and verifier-facing reporting SHOULD make that transitional condition visible, for example with manifest-absent or equivalent wording.

At minimum, an OTS sidecar MUST bind:

Verifiers MUST recompute the day artifact digest and compare it with the sidecar before accepting any proof validation result.

6. Anchoring and Verification

6.1. Anchoring Contract

The generic anchoring contract is simple: a gateway computes the SHA-256 digest of the authoritative day artifact and submits that digest to one or more external timestamping channels. Verifiers MUST first recompute the day artifact digest locally; proof validation occurs only after digest binding validation succeeds.

A deployment conforming to this profile MUST use at least one anchoring channel. OTS is the default channel described by this document; RFC 3161 and peer signatures are optional parallel channels.

6.2. OTS Anchoring Profile

When [OTS] is used, the gateway stamps SHA-256(day/YYYY-MM-DD.cbor) and stores an OTS proof plus an OTS sidecar.

OpenTimestamps is referenced here as a deployed public timestamping ecosystem rather than an IETF-standardized proof format. Implementations claiming OTS support depend on the interoperable behavior of the public OTS project, its calendar servers, and compatible client tooling. OTS proof-format interoperability is therefore defined operationally by that ecosystem and its reference implementations.

6.2.1. OTS Anchoring Lifecycle

  1. Submission: the gateway submits the day artifact digest to one or more OTS calendars.
  2. Pending: a calendar may return an incomplete proof while awaiting Bitcoin commitment.
  3. Upgrade: the gateway or a background process may later upgrade the proof to include completed attestations.
  4. Verification: a verifier recomputes the artifact digest and validates the proof.

Gateways SHOULD submit to multiple independent calendars to reduce single-calendar unavailability risk.

6.2.2. Handling Delayed or Failed Anchoring

If OTS submission fails, times out, or yields only an incomplete proof, the gateway MUST still write the authoritative day artifact and MUST treat OTS as a separate channel whose state is not yet complete. The gateway MAY retain a placeholder or incomplete proof artifact and MAY later replace or upgrade it as additional OTS evidence becomes available. Until a valid proof is disclosed and verified, verifiers MUST report the OTS channel as pending, missing, or failed according to the disclosed artifacts and local policy, rather than treating the day artifact itself as invalid. Any later replacement or upgrade of the OTS proof MUST continue to bind to the same authoritative day-artifact digest.

Gateways MUST retain disclosed OTS proof artifacts for at least the operational retention period of the corresponding day artifacts.

6.2.3. Proof Status Vocabulary

Verifiers and bundle manifests SHOULD use a consistent status vocabulary for OTS and optional parallel attestation channels:

  • verified: proof validation succeeded for the disclosed artifact binding.
  • pending: a proof exists but is incomplete or awaiting upgrade; this is not equivalent to invalid.
  • missing: the expected proof or channel artifact is absent.
  • failed: validation was attempted and did not succeed.
  • skipped: validation was not attempted because of disclosure class, verifier configuration, or local policy.

Whether missing, pending, or skipped is verifier-fatal depends on the disclosure class, local verifier policy, and whether the relevant channel is required.

6.3. Optional Parallel Attestation

Deployments MAY also produce:

  • An [RFC3161] timestamp response over the same day-artifact digest.
  • A peer signature quorum over (site_id, day, day_root, context).

When multiple channels are present, verifiers SHOULD validate all available channels independently and report per-channel results.

If a verifier is configured in strict mode for optional channels, failure of those channels MUST cause overall verification failure.

6.4. Verification

Verifiers MUST first determine the applicable commitment_profile_id from disclosed bundle metadata. When the authoritative day artifact does not carry that identifier in-band, the verifier MUST obtain it from the verification manifest or equivalent local verifier configuration and MUST reject absent or unsupported values.

Verifiers MUST determine the applicable verification scope from the disclosed artifacts, the claimed disclosure class, and local verifier policy. Reported outcomes MUST NOT claim checks or assurances outside that scope.

Verifiers SHOULD apply checks in the following fail-fast order, subject to the claimed disclosure class:

  1. Validate that disclosed artifacts are sufficient for the claimed disclosure class and disclose a commitment_profile_id.
  2. For Class A bundles, recompute canonical fact leaf digests from disclosed fact CBOR artifacts, validate the batch metadata contract, and recompute day_root. Compare the recomputed result to the authoritative day_root.
  3. For Class B bundles, validate the authoritative day artifact, validate any disclosed batch metadata, and report that public fact-level recomputation was not attempted for withheld material. If the bundle additionally discloses commitments covering withheld material, validate those commitments under the applicable bundle policy.
  4. For Class C bundles, skip fact-level and batch-level recomputation and treat the result as anchor-only evidence.
  5. Recompute SHA-256(day/YYYY-MM-DD.cbor) and compare it to the sidecar artifact_sha256.
  6. Validate the OTS proof when OTS is required or present.
  7. Validate optional RFC 3161 and peer attestations as configured.

When batch metadata is available, verifiers SHOULD apply the following checks before accepting a recompute result:

  • each batch count equals the length of its leaf_hashes;
  • each batch merkle_root equals the Merkle reduction of its leaf_hashes;
  • the union multiset of batch leaf_hashes equals the leaf digest multiset derived from disclosed facts when fact artifacts are available; and
  • any standalone block-metadata projection matches the corresponding batch object in the authoritative day artifact.

On current TrackOne tooling paths, these batch-metadata checks are enforced most strongly when a standalone verification manifest is present. Manifest-absent verification MUST NOT be interpreted as proving every disclosed standalone block-metadata field was validated unless the implementation explicitly reports that those checks were executed.

Verifier implementations SHOULD expose machine-usable failure categories:

  • malformed or missing artifacts,
  • missing or unsupported commitment_profile_id,
  • Merkle mismatch,
  • batch metadata mismatch,
  • missing or invalid OTS proof,
  • sidecar mismatch or digest mismatch,
  • insufficient disclosure for the claimed verification level, and
  • optional-channel failure.

Verifier output SHOULD state the claimed disclosure class, the verification scope actually exercised, the per-channel proof status, which checks were executed, which checks were skipped, and whether the resulting claim is public recompute, partial verification, or anchor-only evidence.

Verifier output MUST NOT be represented as proving more than the exercised verification scope. In particular, a successful result does not by itself establish dataset completeness, physical truth of measurements, or suitability for autonomous actuation or sanctions.

7. Disclosure Classes

Verification claims depend on what artifacts are disclosed. This profile defines three disclosure classes.

A verifier claim for any disclosure class MUST be limited to the verification scope supported by the disclosed artifacts.

A Class A bundle MUST include:

A Class A bundle SHOULD also include a machine-readable verification manifest and SHOULD record the commitment_profile_id.

If a Class A-capable producer path omits the standalone manifest, deployments and verifier-facing reporting SHOULD make that omission visible, for example with manifest-absent or equivalent transitional wording, rather than silently presenting it as a fully packaged bundle.

Class A permits fact-level recomputation, batch-metadata validation, day-root recomputation, and anchor validation when the corresponding artifacts are disclosed and checked.

At the current 0.1.0-alpha.10 implementation boundary, Class B is primarily a disclosure label and verifier reporting mode rather than a universally emitted controlled-disclosure bundle format. The current verifier skips public fact-level recomputation for Class B and reports a non-public result, but it does not yet require a standardized withheld-material commitment artifact set on every tooling path.

A fuller Class B controlled-disclosure bundle SHOULD include:

Class B outputs MUST NOT be represented as publicly recomputable.

Class B validation is limited to the authoritative day artifact, disclosed batch metadata, any withheld-material commitments actually disclosed in the bundle, and any anchor channels present in the bundle. Verifier output SHOULD explicitly state when fact-level recomputation was partial or was not attempted for withheld material.

A Class C disclosure MUST be labeled as existence and timestamp evidence only and MUST NOT claim fact-level reproducibility.

A Class C bundle MUST include:

Class C verification validates artifact-digest binding and external timestamp evidence only. It does not establish fact-level reproducibility.

Class C verifier output SHOULD be reported as anchor-only evidence.

If a verification manifest is present, it SHOULD include:

8. Versioning

This profile has several independent version surfaces:

The commitment profile defined in this document is trackone-canonical-cbor-v1. If a verifier encounters an unsupported commitment_profile_id, it MUST reject the verification claim rather than silently using a fallback interpretation.

Bundles disclose the applicable commitment_profile_id via a verification manifest or, when no standalone manifest is present, via equivalent local verifier configuration or other out-of-band bundle context. The required OTS sidecar metadata does not currently carry that identifier.

9. Conformance Vectors

Determinism claims in this profile are testable. When conformance vectors are published for a commitment profile, implementations that claim conformance to that profile MUST be able to reproduce them. For trackone-canonical-cbor-v1, publication of the full machine-readable vector corpus remains planned work at the current alpha.10 implementation boundary.

Published vector sets SHOULD include coverage for:

Cross-implementation checks MUST verify byte-for-byte parity across independent implementations. Any mismatch in canonical bytes or roots is a conformance failure.

Published vector bundles MUST include the commitment_profile_id.

10. Security Considerations

This profile does not introduce new cryptographic primitives. Security depends on correct composition of existing primitives and on operational discipline.

Replay and Duplicate Suppression

The (dev_id, fc) replay unit enforces single-consumption: at most one committed fact per unique replay unit.

Tamper Evidence

Once a day artifact is anchored, mutation of that artifact changes its digest and invalidates the proof. Day chaining extends this property across days.

Proof Substitution

Sidecar metadata binds an artifact digest to a proof path. Verifiers MUST recompute the artifact digest independently.

Calendar Trust and Withholding

A compromised or unavailable calendar can delay or withhold proofs. Multi-calendar submission reduces single-calendar dependency but does not eliminate coordinated compromise risk.

Operational Misuse

Test or placeholder proofs MUST NOT be treated as production attestations.

Decision Boundary

This profile is not a standalone safety case. A successful verification result MUST NOT be used as the sole basis for autonomous actuation, safety-critical control, or regulatory sanction without deployment-specific policy, corroborating evidence, and human review.

Confidentiality Boundary

This profile addresses integrity and timing provenance only. Payload confidentiality remains the responsibility of the deployment's AEAD and key-management layers.

Pre-Commitment Censorship

This profile proves inclusion and timestamping of committed facts. It does not prove completeness of all observed or emitted frames.

Hash-Domain Separation

The current commitment profile does not prepend domain-separation bytes to leaf or parent hashes. This is acceptable for trackone-canonical-cbor-v1 because the profile is frozen around the published implementation and conformance vectors, but implementers MUST treat that choice as part of the profile contract rather than as an accidental omission. Any future profile that introduces explicit leaf or parent domain separation MUST use a new commitment_profile_id.

11. Privacy Considerations

Telemetry payloads may include sensitive operational data. Operators SHOULD:

Privacy-preserving disclosures remain valid, but they MUST NOT be described as publicly recomputable unless Class A conditions are met.

12. IANA Considerations

This document has no IANA actions.

In particular, this revision does not request:

The current profile is identified by in-band version and profile fields, not by IANA allocation.

13. Interoperability Notes and Open Questions

14. References

14.1. Normative References

[RFC2119]
Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, DOI 10.17487/RFC2119, , <https://www.rfc-editor.org/rfc/rfc2119>.
[RFC8174]
Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, , <https://www.rfc-editor.org/info/rfc8174>.
[RFC8949]
Bormann, C. and P. Hoffman, "Concise Binary Object Representation (CBOR)", STD 94, RFC 8949, DOI 10.17487/RFC8949, , <https://www.rfc-editor.org/rfc/rfc8949>.

14.2. Informative References

[RFC3161]
Adams, C., Cain, P., Pinkas, D., and R. Zuccherato, "Internet X.509 Public Key Infrastructure Timestamp Protocol (TSP)", RFC 3161, DOI 10.17487/RFC3161, , <https://www.rfc-editor.org/rfc/rfc3161>.
[RFC8785]
Rundgren, A., Jordan, B., and S. Erdtman, "JSON Canonicalization Scheme (JCS)", RFC 8785, DOI 10.17487/RFC8785, , <https://www.rfc-editor.org/rfc/rfc8785>.
[SCITT]
Birkholz, H., Delignat-Lavaud, A., Fournet, C., Deshpande, Y., and S. Lasker, "An Architecture for Trustworthy and Transparent Digital Supply Chains", Work in Progress, Internet-Draft, draft-ietf-scitt-architecture, , <https://datatracker.ietf.org/doc/draft-ietf-scitt-architecture/>.
[COSE-MERKLE]
Steele, O., Birkholz, H., Delignat-Lavaud, A., and C. Fournet, "COSE Merkle Tree Proofs", Work in Progress, Internet-Draft, draft-ietf-cose-merkle-tree-proofs, , <https://datatracker.ietf.org/doc/draft-ietf-cose-merkle-tree-proofs/>.
[OTS]
OpenTimestamps Project, "OpenTimestamps Protocol and Tooling", , <https://opentimestamps.org/>.
[RFC9334]
Birkholz, H., Thaler, D., Richardson, M., Smith, N., and W. Pan, "Remote ATtestation procedureS (RATS) Architecture", RFC 9334, DOI 10.17487/RFC9334, , <https://www.rfc-editor.org/rfc/rfc9334>.

Appendix A. Example Day Record (JSON Projection)

This appendix shows a non-authoritative JSON projection of a day artifact. The authoritative artifact is the corresponding CBOR file.

{
  "version": 1,
  "site_id": "an-001",
  "date": "2025-10-07",
  "prev_day_root": "<64 zero hex chars>",
  "batches": [
    {
      "version": 1,
      "site_id": "an-001",
      "day": "2025-10-07",
      "batch_id": "an-001-2025-10-07-00",
      "merkle_root": "9d1f...c2",
      "count": 10,
      "leaf_hashes": [
        "01ab...",
        "7fe2..."
      ]
    }
  ],
  "day_root": "9d1f...c2"
}
Figure 2

Appendix B. Illustrative Conformance Vector Bundle

A full machine-readable conformance vector corpus is planned alongside the TrackOne reference implementation, but it is not yet published as a complete authoritative bundle at the current 0.1.0-alpha.10 boundary. The figure below is illustrative of a target vector-bundle shape and naming only.

Wrapped hexadecimal values in this appendix are presentation-only; a verifier or implementer should concatenate adjacent lines without inserting whitespace.

The fixtures below are post-projection canonical fact objects, not wire-frame inputs. They use the current canonical fact shape and the active commitment profile identifier from the alpha.10 toolchain.

commitment_profile_id:
  trackone-canonical-cbor-v1

fixture fact_a:
  pod_id: "0000000000000065"
  fc: 1
  ingest_time: 1772366400
  pod_time: null
  kind: Custom
  payload.temp_c: 21.5

fixture fact_b:
  pod_id: "0000000000000066"
  fc: 2
  ingest_time: 1772367000
  pod_time: null
  kind: Custom
  payload.temp_c: 22.0

fixture fact_c:
  pod_id: "0000000000000067"
  fc: 3
  ingest_time: 1772367600
  pod_time: null
  kind: Custom
  payload.temp_c: 22.5

class-a-bundle-v1:
  disclosure_class: A
  commitment_profile_id: trackone-canonical-cbor-v1
  required_artifact_1: facts/<fact-id>.cbor
  required_artifact_2: day/YYYY-MM-DD.cbor
  required_artifact_3: day/YYYY-MM-DD.cbor.ots
  required_artifact_4: day/YYYY-MM-DD.ots.meta.json
  verifier_check_1: day_artifact_validation
  verifier_check_2: fact_level_recompute
  verifier_check_3: ots_verification
Figure 3

When the full machine-readable vector set is published, it is expected to carry exact canonical bytes, digests, expected roots, and the applicable commitment_profile_id. Until then, this appendix is illustrative only and is not an authoritative conformance corpus.

Acknowledgments

Early structural drafts of this document were prepared with AI writing assistance. All technical content, design decisions, and normative requirements were reviewed against the TrackOne implementation.

The author thanks the OpenTimestamps project for the public calendar infrastructure used during validation.

Author's Address

Bilal El Khatabi
TrackOne Project
Morocco