Jonathan M. Watson | CVS Cryptographic Verification Sidecar Version 1.1 | May 2026
VeritasChain Protocol (VCP), published January 13, 2026 by the VeritasChain Standards Organization, is architecturally similar to CVS. Both implement a sidecar witness pattern with hash-chained events, Merkle anchoring, and external ledger settlement. Both solve the same fundamental problem: producing independently verifiable cryptographic evidence of system execution without modifying or blocking that execution.
This document acknowledges VCP as an independent parallel development, states CVS's priority position, and identifies the architectural properties that distinguish the two systems.
Honest acknowledgment of parallel development strengthens the open commons claim — when multiple independent actors converge on the same architecture within weeks of each other, the architecture is a natural solution to the problem, not an individual invention. That is precisely CVS's philosophical position.
CVS genesis commit: December 17, 2025 (commit 32cbd9b,
github.com/JonathanMastersWatson/Evidence-Sidecar)
VCP v1.0 published: January 13, 2026
CVS was publicly committed 27 days before VCP existed. This gap is documented
in CANONICAL_COMMITMENT.md with independent verification instructions.
Apache 2.0 was formally declared on February 28, 2026; the prior art date
is the genesis commit, not the license declaration date.
VCP and CVS arrived at the same core architectural pattern independently:
- Sidecar deployment — the witness layer operates alongside the observed system without modifying or blocking it
- Hash-chained event records — each captured event cryptographically references its predecessor, making chain tampering mechanically detectable
- Merkle tree aggregation — batches of events are aggregated into Merkle roots for efficient anchoring
- External ledger anchoring — cryptographic commitments are anchored to a public settlement ledger for independent verification
- Tamper-evidence — the architecture produces evidence that any third party can verify without trusting the operator
- Non-blocking operation — the witness layer does not sit in the execution path; failure creates detectable gaps rather than execution failures
The convergence on these properties by two independent projects within 27 days is consistent with the view that the sidecar witness pattern is a discovered architecture — the natural solution to the problem of producing independently verifiable execution evidence without operational coupling.
The shared foundation aside, CVS and VCP diverge on five properties that matter in high-assurance deployment contexts.
VCP uses PTPv2 (IEEE 1588 Precision Time Protocol) for its highest precision tier, with NTP and best-effort tiers below. PTPv2 achieves sub-microsecond precision within a network but synchronizes against network time sources — sources that remain within infrastructure the operator controls or influences.
CVS uses public ledger consensus as the primary external time anchor. Settlement anchors are written to the XRP Ledger every 30–60 seconds. The timestamp on each anchor is set by ledger consensus — a distributed process no single operator controls, observable by any party with internet access, and impossible to backdate. An operator cannot alter the ledger timestamp on an anchor that has already achieved finality.
This distinction matters for any dispute, regulatory examination, or insurance claim where the opposing party questions whether evidence was fabricated after the fact. A timestamp set by a neutral public consensus ledger is structurally more defensible than a timestamp sourced from infrastructure the operator administers. The ledger anchor provides an external temporal bound that internal clock manipulation cannot override — any fabricated timestamp preceding the anchor time is exposed by the contradiction.
This property requires no special hardware. Any system with internet access — including on-premises broadcast infrastructure, legacy trading systems, and air-gapped environments with scheduled connectivity — can anchor to the public ledger. The external time anchor is available wherever CVS can be deployed.
CVS requires fail-open behavior as a non-negotiable conformance requirement. At no point may the operation, availability, or correctness of the observed system depend on sidecar availability. If the sidecar fails under any condition — power loss, network partition, ledger unavailability, software crash, clock desynchronization, key rotation failure, resource exhaustion — the observed system continues operating unchanged. The failure is observable. The resulting evidence gap is detectable and recorded upon recovery. Execution impact is always zero. This is not a design preference — it is a conformance requirement. An implementation that can block execution is not a conformant CVS implementation.
VCP operates without modifying execution logic, suggesting non-blocking design intent. Whether fail-open behavior is a formally specified conformance requirement in VCP — and whether it covers the full range of failure conditions CVS enumerates — is not stated in VCP's published specification.
For deployment in high-availability environments — broadcast infrastructure, financial systems, healthcare, industrial control — the distinction between "non-blocking by design intent" and "fail-open as non-negotiable conformance requirement" is operationally significant. Engineering teams reject evidence systems that carry execution risk. The conformance requirement, not just the design intent, is what enables deployment in production environments without risk negotiation.
VCP does not define a structural separation between evidence capture, evidence access, and evidence interpretation.
CVS enforces three structurally separated planes:
- Capture Plane — observes execution, constructs Evidence Objects, anchors commitments. No read access from external systems.
- Access Plane — provides read-only interfaces to the evidence chain. No write access to the Capture Plane. No interpretation logic.
- Interpretation Plane — external tools (dashboards, analytics, compliance systems) that consume evidence through the Access Plane. No modification authority over evidence.
This separation is enforced by network segmentation and IAM policy — structural controls, not administrative procedure. The planes cannot collapse through misconfiguration because the access controls are mechanically enforced.
The practical consequence: in CVS, no interpretation tool can modify captured evidence, and no evidence modification can be disguised as an access operation. In systems without structural plane separation, the distinction between capture, access, and interpretation exists only as administrative policy — inadequate for high-assurance regulatory, insurance, or legal contexts.
VCP does not specify a requirement for administrative independence between the witness layer and the observed system.
CVS makes administrative independence an architectural requirement. The witness layer cannot share operational authority, key custody, or administrative access with the system it witnesses. The operator of the observed system is not the custodian of the evidence that system generates.
This property is what makes CVS evidence admissible in contexts where the operator's own records would not be. An operator cannot suppress or alter CVS evidence without detectable action against a system they do not administratively control. The independence is structural, not procedural.
VCP does not define a selective disclosure mechanism.
CVS includes a Disclosure Kernel — a structured mechanism for selective revelation of evidence to authorized parties without overexposure. A party entitled to audit a specific event receives cryptographic proof of that event without receiving the full evidence chain. This allows CVS to serve regulators, insurers, and opposing counsel with scoped evidence packages — sufficient to answer the specific question, insufficient to constitute wholesale exposure of proprietary operational data.
| Property | VCP | CVS |
|---|---|---|
| Genesis date | January 13, 2026 | December 17, 2025 |
| Sidecar pattern | Yes | Yes |
| Hash-chained events | Yes | Yes |
| Merkle anchoring | Yes | Yes |
| External ledger anchor | Yes | Yes |
| External time anchor | PTPv2 / NTP / best-effort (operator-adjacent) | Public ledger consensus — operator-independent |
| Fail-open | Non-blocking design intent | Non-negotiable conformance requirement |
| Three-plane separation | Not defined | Structural — enforced by IAM and network segmentation |
| Administrative independence | Not specified | Architectural requirement |
| Selective disclosure | Not defined | Disclosure Kernel |
| Deployment scope | Algorithmic trading / financial markets | Any execution surface — including on-premises |
| License | Standards organization model | Apache 2.0 — fully open |
VCP's existence confirms that the sidecar witness pattern is correct. Two independent teams reached the same architecture within 27 days of each other. That is not coincidence — it is convergence on a natural solution.
CVS is open under Apache 2.0. VCP is published under a standards organization model. Both are available for implementation. The base sidecar pattern belongs to neither — it is prior art predating both, established by CVS's December 17, 2025 genesis commit.
Derivative implementations of CVS — managed witness services, certification
wrappers, interpretation platforms, domain-specific deployments — are fully
patentable and commercialisable by their creators. See CVS_ARCHITECTURE §-1.5
for the complete open commons model.
VCP specification: VeritasChain Standards Organization, VCP v1.0 (January 2026)
CVS genesis commit: 32cbd9b — December 17, 2025
CVS canonical specification commitment: commit a7762a9 — February 28, 2026
CVS Apache 2.0 declaration: commit b9d0cff — February 28, 2026
Full prior art and competitive landscape analysis: HEX/10_ip/512_CVS_PRIOR_ART_MAP.md
Canonical commitment record: CANONICAL_COMMITMENT.md
CVS and VCP — Parallel Development Acknowledgment | Version 1.1 | May 2026 Author: Jonathan M. Watson Released under Apache 2.0 consistent with Evidence-Sidecar repository licensing. This document does not constitute legal advice.