← Research WP-09 · CONTROL PLANE Operators & Architects · CP v0.4.0

The ACSE Control Plane — Operating Surface Mutation at Estate Scale

How ACSE becomes operable: API-key + RBAC management — with ASMP MSG-005 node-granted protected actions (modelled in ProVerif; gating one protected action today; TEE binding planned) not yet integrated into the live authentication path — live estate monitoring, Auto-Discovery across 22 source types, zero-config agent self-registration, firewall and SIEM integration, policy, and a tamper-evident audit chain — all behind a single REST API.

Abstract

The Polymorphic Mutation Engine (PME) enforces the Kali Invariant on a single node; ASMP/1.0 coordinates that enforcement across an estate. Neither, on its own, is something an enterprise can operate. The ACSE Control Plane is the fourth architectural layer — the management surface through which operators discover assets, onboard them into the catalog, monitor live state, set policy, integrate with existing security infrastructure, and produce a tamper-evident record of every action.

This paper covers Control Plane v0.4.0. It specifies live authentication (API-key + RBAC) and the MSG-005 node-grant design (modelled in ProVerif; gating one protected action today, not yet integrated into authentication), the live estate dashboard, the Auto-Discovery pipeline (22 source types), the zero-config agent self-registration mechanism, the push protocol, outbound firewall and SIEM integration, the policy engine and estate REST API, and the cryptographic audit chain — followed by deployment, operations, and the v2 roadmap.

Patent & Classification

Indian patent application published · IN202641070690 · 19/06/2026 · Inventor: Arul Raj · Filed with the Indian Patent Office with expedited examination requested (Form 18A) · Classification: Public Technical Paper — Post-Patent Publication

Control Plane v0.4.0 — At a Glance

2,130 tests · 0 failures (20 October 2026)  ·  API-key + RBAC authentication (3 roles enforced per endpoint)  ·  MSG-005 node-granted protected actions (modelled in ProVerif; one protected action today; TEE binding planned across 6 TEE/TPM adapters — Mock, TPM, SGX, Nitro, SEV-SNP, ARM CCA; only TPM is hardware-verified today) — not yet wired into the live auth path  ·  22 Auto-Discovery source types (Active Directory · Azure AD · SSH · TCP Probe · AWS / Azure / GCP · Nutanix · Cisco UCS · VMware vCenter · Azure Stack HCI · Pure Storage · NetApp ONTAP · Dell EMC · Brocade FC · Cisco MDS · Ceph · Panorama · FortiManager · SNMP · CSV · Agent Self-Register)  ·  3 firewall adapters  ·  Splunk / QRadar / Syslog / stdout dispatch  ·  SHA3-256 tamper-evident audit chain  ·  ProVerif: MSG-005 management binding CORRECT (protocol design)  ·  Windows MSI + Linux DEB  ·  Zero-config agent self-registration

Control Plane v0.4.0 · Endpoint Agent v0.4.0 · System Tray v0.4.0 (pme, asmp, and alert-bridge maintain independent internal versioning — currently v0.1.0 — as they are candidates for standalone crate/RFC publication: pme for ASRK-ODS reuse, asmp as an ASMP/1.0 protocol RFC candidate)

1. From Engine to Operable System

A mutation engine that no one can deploy, monitor, or audit is a research artifact, not a product. The journey from PME to a fielded defence runs through a set of unglamorous but decisive operational requirements: how do you discover the hundreds of assets that need protecting? How do you watch their state in real time? How do you change policy across an estate without touching every node by hand? How do you feed events into the SOC's existing SIEM? And how do you prove, later, exactly what happened and who authorised it?

The Control Plane answers each of these. Crucially, it does so without weakening the security model of the layers beneath it. A management plane is the highest-value target in any system — compromise it and you inherit the estate. The Stryker–Handala wiper attack of March 2026 is the canonical example: the attackers did not break encryption, they weaponised the victim's own MDM management console to issue estate-wide wipe commands.

Design Principle

The management plane must be at least as hard to subvert as the data plane. Today, that plane is gated by API-key + RBAC. The Control Plane also specifies a stronger design — every privileged operation bound to a hardware-attested MSG-005 token, so valid administrator credentials would be necessary but not sufficient — modelled in ProVerif at the protocol level, without the hardware binding (see Section 3), but not yet integrated into the live authentication path.

2. Control Plane Architecture

The Control Plane sits atop the three runtime layers — PME, ASMP/1.0, and the profile layer (ten bio-inspired domain profiles plus TorpedoRay) — as the operator-facing management layer. It is a single Rust binary (acse-cp) backed by PostgreSQL, exposing a REST API on port 9000 and the pme-console live dashboard, with four families of connectors: inbound discovery, inbound agent push, outbound firewall orchestration, and outbound SIEM dispatch.

OPERATOR Admin · Operator · RO RBAC CHECK Per-endpoint role TEE ATTESTATION MSG-005 · TTL 1 cycle SPECIFIED — NOT YET LIVE CONTROL PLANE API Axum · REST · port 9000 PME ESTATE Profiles + ASMP/1.0 REJECTED No valid TEE quote (once MSG-005 is live) SHA3-256 Append-only audit chain Design: once MSG-005 is integrated, every privileged operation would be gated by a TEE-attested token (TTL = 1 cycle). Today, RBAC-cleared credentials execute directly — the dashed gate is not yet live.
Figure 1: Designed Control Operation Lifecycle (MSG-005 dashed = specified and modelled in ProVerif, not yet in this path) — credentials pass RBAC and execute today without a TEE-attestation gate; once MSG-005 is integrated, the operation would only execute after a valid TEE quote produces a one-cycle token. Every committed action is appended to the SHA3-256 audit chain regardless.
ComponentRoleImplementation
Estate REST APISingle point of authority for all management operationsAxum; API-key + RBAC on every privileged route today (MSG-005 TEE attestation specified, not yet integrated)
pme-consoleLive operator dashboard — estate status, Kali InvariantServer-side rendered; auto-refreshes every 5 seconds
Discovery engineEnumerate and classify the estate from 22 source types22 connectors → asset catalog → profile suggestion
Self-register endpointZero-config agent onboardingPOST /v1/discovery/self-register · Registration token auth
Firewall orchestrationPropagate surface mutations to perimeter devicesPalo Alto · Cisco ASA · Fortinet adapters
SIEM dispatchStream mutation and Torpedo events to the SOCSplunk HEC · IBM QRadar · RFC 5424 Syslog · stdout
Policy engineMutation schedules, EWMA thresholds, active profilesPer-estate policy assigned across nodes in one call
Audit aggregatorEstate-wide tamper-evident recordSHA3-256-chained, PostgreSQL-backed, paginated API
Kali Invariant monitorGlobal health/invariant status across all nodesContinuous; surfaced live on dashboard and GET /v1/invariant

3. Authentication & RBAC (MSG-005 — Specified, Not Yet Live)

Authentication in the Control Plane today is single-factor at the transport layer: an operation requires an authenticated identity (API-key) with sufficient role (RBAC). ASMP message MSG-005 specifies a stronger, two-factor design — identity-plus-role and a fresh hardware attestation, binding a management trust token to a Trusted Execution Environment quote with a time-to-live of exactly one mutation cycle. MSG-005's node-grant property is proven in ProVerif (see WP-07; no TEE is modelled), but it is not currently wired into any live API authentication path — production authentication is API-key + RBAC only, and MSG-005 is planned for a future release.

3.1 Role-Based Access Control

  • Admin — full estate authority: policy changes, node registration, profile assignment, connector configuration, user management.
  • Operator — day-to-day operations: trigger defensive leaps, acknowledge anomalies, run discovery, configure SIEM and firewall adapters, read all state.
  • ReadOnly — observation only: dashboard, audit chain, and reports, with no mutating operations.

3.2 The Six TEE/TPM Adapters

AdapterPlatformAttestationStatus
MockTEEAdapterAny (software)Deterministic mock quotesActive (dev/test)
TpmAdapterTPM 2.0 (commodity hardware)TPM2_Certify (binds the signing key's name; not a PCR quote)Live — key custody hardware-verified; no CP-side quote verification yet (#143)
SGXTEEAdapterIntel SGX (x86_64)EPID / DCAP remote attestationBuilt, unit-tested — pending real-hardware validation
NitroTEEAdapterAWS Nitro EnclavesPCR-based attestationBuilt, unit-tested — pending real-hardware validation
SEVSNPTEEAdapterAMD SEV-SNPVM-level hardware attestationBuilt, unit-tested — pending real-hardware validation
ARMCCATEEAdapterARM CCA (Realm)Realm attestation tokenBuilt, unit-tested — no commercially available ARM CCA hardware exists to validate against

Of these, only TpmAdapter has been run against real hardware and produces a verified audit-signing key today (KeyCustody::Tpm). SGX, Nitro, and SEV-SNP have real, unit-tested implementations but have not been validated against physical hardware of their respective kind — that is a funded-hardware-access gap, not further engineering. ARM CCA is a harder limit: no commercially purchasable ARM CCA hardware exists to test against at all, so the adapter is retained for interface completeness (see Patent Claim 2's detection-order enumeration) but cannot be verified on anything. Selecting an adapter is a one-line change at construction; all Control Plane logic is adapter-agnostic.

Formal Verification — MSG-005 (Protocol Design)

Under the Dolev-Yao adversary model (full network control), ProVerif confirms the management binding is CORRECT across all queries: as specified, an adversary holding valid administrator credentials but no TEE quote could not execute a privileged operation. This is the property that would turn the Stryker–Handala class of attack from catastrophic into inert once MSG-005 is integrated — it is not the mechanism protecting the Control Plane today. Production authentication is API-key + RBAC; MSG-005 is planned for a future release. See WP-07 for the full model.

4. The Estate Dashboard

The pme-console dashboard presents a live organ-state view across every registered node, refreshing every five seconds. It is the operator's primary situational-awareness surface, and the same data is available programmatically through the REST API (GET /v1/estates).

SignalMeaning
Organ stateSachs (steady) → Hunter (elevated) → Main Organ (full response) per node
EWMA anomaly scoreExponentially-weighted moving anomaly score driving state transitions (0.35 → Hunter, 0.70 → Torpedo)
Torpedo countForced-Twitch / Defensive-Leap firings — confirmed scanner self-identifications
Mutation countTotal mutation cycles completed by the node since registration
Channel fingerprintCurrent mutation-cycle fingerprint of the protected surface
JA3 / p0f signaturesPassive client and OS fingerprints observed against the surface
Kali Invariant statusPer-node and estate-wide invariant health (green = all surfaces rotating correctly)
Last push timestampTime of most recent agent push — alerts if > 2× push interval

Asset rows are auto-classified — firewall, server, or network device — so an operator can tell at a glance which surfaces are under which profile, and which discovered nodes are still awaiting onboarding.

5. Auto-Discovery — 22 Source Types

Most deployments stall at inventory: nobody has a complete, current list of the assets that need protecting. Auto-Discovery removes that obstacle by enumerating the estate from the identity and network systems an organisation already runs — then cataloguing, classifying, and suggesting the appropriate mutation profile for each asset.

v0.2.2 ships with 22 source types across six categories, up from 6 in v0.1.0. The expansion covers HCI management planes (Nutanix, Cisco UCS, VMware vCenter, Azure Stack HCI) and storage fabric controllers (Pure Storage, NetApp ONTAP, Dell EMC, Brocade FC, Cisco MDS, Ceph) — surfaces that carry catastrophic blast radii and have historically been protected by static, never-rotated credentials.

CategorySource TypeDiscoversAuto-Suggested Profile
Identity / DirectoryldapAll enabled computer objects from AD/LDAPKrakenNet for DCs; GlassFrog for DBs
azure_adAzure AD / Entra ID cloud identities and devicesKrakenNet (Windows); KaliCoreTarget (Linux)
Active ScansshLinux servers in no-AD environments (SSH into CIDR ranges)By hostname pattern + service detection
active_probeAny live host by TCP port signature (no credentials)By port fingerprint (RDP→KrakenNet, Postgres→NautilusVault)
CloudawsAWS EC2 instances via aws-cliKrakenNet (Windows); KaliCoreTarget (Linux)
azureAzure VMs via az-cliKrakenNet (Windows); KaliCoreTarget (Linux)
gcpGCP Compute Engine instances via gcloudKrakenNet (Windows); KaliCoreTarget (Linux)
Network DevicespanoramaPalo Alto firewall fleet via Panorama XML APIN/A — routes to Firewall Adapters
forti_managerFortinet FortiGate fleet via FortiManager JSON-RPCN/A — routes to Firewall Adapters
snmpAny network device via SNMP v2c sweepN/A — routes to Firewall Adapters
HCI Management PlanesnutanixNutanix Prism Central — clusters, hosts via REST v3KrakenNet — credential surface
cisco_ucsCisco UCS Manager — blades via XML APIKrakenNet — credential surface
vmware_vcenterVMware vCenter — ESXi hosts and vSAN clusters via RESTKrakenNet — credential surface
azure_stack_hciAzure Stack HCI clusters via Azure Arc REST APIKrakenNet — credential surface
Storage Fabricpure_storagePure Storage FlashArray — arrays and volumes via REST v2NautilusVault — data vault surface
netapp_ontapNetApp ONTAP — cluster and SVMs via REST APINautilusVault — data vault surface
dell_powerstoreDell EMC PowerStore appliances via REST APINautilusVault — data vault surface
brocade_fcBrocade FC switches via FOS REST APILeviathanGrid — topology surface
cisco_mdsCisco MDS FC fabric via NX-APILeviathanGrid — topology surface
cephCeph cluster — OSDs, monitors, pools via Dashboard APIKaliCoreTarget — maximum protection
ImportcsvBulk import via CSV file uploadBy node_type column or admin override
agent_self_registerCreated automatically on agent self-registrationKaliCoreTarget (default; overridable)
Human-in-the-loop by design

Discovery never auto-onboards. Classification and profile suggestion are advisory; an Admin or Operator must approve before a surface enters mutation. The catalog is a decision aid, not an autonomous actuator. The one exception is agent self-registration — where a pre-shared registration token constitutes advance human authorisation.

6. Estate Catalog — Discover → Decide → Protect

Every node discovered by any source lands in the discovered_nodes catalog table. The catalog is the intermediate state between "found" and "protected." The admin reviews, approves, and clicks Protect — the Control Plane creates an estate and api_key automatically.

StatusMeaningNext Action
discoveredFound by a scan — not yet reviewedAdmin reviews: approve or exclude
approvedMarked for protection — awaiting agent installPOST /v1/discovery/catalog/:id/protect → get api_key
excludedIntentionally unprotected (DMZ, test nodes)Stays in catalog for audit; not counted in Invariant
protectedEstate created, agent active and pushingMonitor via dashboard; counted in Kali Invariant

The POST /v1/discovery/catalog/:id/protect endpoint returns the estate record, api_key, and platform-specific install instructions in a single response — the operator copies the key to the node, installs the agent, and that node is under the Kali Invariant within minutes.

7. Estate Lifecycle

An estate is the Control Plane's representation of a protected node. It is the join point between the PME agent (which executes mutations) and the Control Plane (which aggregates, dispatches, and monitors).

  1. Registration: POST /v1/estates (manual) or via catalog Protect (auto). Returns a per-estate api_key — shown once, SHA3-256 hash stored only.
  2. Agent push: The PME agent calls POST /v1/estates/:id/push every 30 seconds with mutation count, EWMA score, organ state, and audit records.
  3. SIEM dispatch: On every push, the bridge module dispatches CEF-formatted events to all configured SIEM sinks.
  4. Firewall sync: When a TorpedoRay surface mutates, the CP pushes the new channel fingerprint to connected firewall adapters automatically.
  5. Audit chain: Every mutation record is appended to the estate's SHA3 audit chain in PostgreSQL. Each record contains the SHA3-256 hash of its predecessor — the chain is tamper-evident.
  6. Kali Invariant: The CP continuously checks that every active estate is in a valid mutation state. A violation (estate offline, chain gap) flags the global invariant.

8. Agent Self-Registration — New in v0.3.0

Self-registration is the zero-config deployment model introduced in v0.2.1. It eliminates the manual register-then-install workflow for large fleets: set a pre-shared registration token on the CP and on each agent. On first boot, the agent registers itself and receives its api_key.

Zero-Config for Enterprise Scale

Set ACSE_REGISTRATION_TOKEN on the Control Plane and on each agent node. On first boot, the agent calls POST /v1/discovery/self-register with its hostname, IP, and OS fingerprint. The CP creates the catalog entry, creates the estate, generates the api_key, and returns it. The agent saves it in memory and begins pushing immediately. No admin action required per node.

StepWhat Happens
1Agent starts — ACSE_API_KEY absent, ACSE_REGISTRATION_TOKEN present
2Agent POSTs to /v1/discovery/self-register with hostname, OS, IP, node_id
3CP validates the Registration token against ACSE_REGISTRATION_TOKEN env var
4CP creates a discovered_node catalog entry (status: protected)
5CP creates estate + api_key via store::create_estate()
6CP returns {registered: true, api_key, estate_id, profile} to agent
7Agent stores api_key in memory — begins 30s push loop immediately
8Admin sees node in dashboard — already active under Kali Invariant

The self-register endpoint is the only /v1/ path that is public (not behind the Bearer auth middleware). It is disabled by default — only enabled when ACSE_REGISTRATION_TOKEN is set on the CP. This makes it opt-in and auditable.

9. Agent Push Protocol

The PME Agent pushes its status to the Control Plane every ACSE_PUSH_INTERVAL_S seconds (default: 30). The push is a single authenticated POST with the estate's api_key as a Bearer token.

POST /v1/estates/:id/push
Authorization: Bearer acse_XXXXXXXXXXXXXXXXXXXX
Content-Type: application/json

{
  "organ_state":    "sachs",
  "ewma_score":     0.08,
  "torpedo_count":  0,
  "mutation_count": 2880,
  "transport_mode": "asmp",
  "audit_records": [
    {
      "cycle_seq":           2880,
      "surface_fingerprint": "sha3hex...",
      "record_hash":         "sha3hex...",
      "tee_signature":       "base64...",
      "trigger_reason":      "timer",
      "mutation_result":     "success"
    }
  ]
}

On receiving a push, the CP: (1) updates estate_live_status in PostgreSQL, (2) appends audit records to the SHA3 chain, (3) dispatches CEF events to all enabled SIEM sinks, (4) checks the Kali Invariant. Response: {"status":"ok"}.

Organ StateMeaningEWMA Range
sachsNormal scanning — surface rotating at base rate0.00 – 0.34
hunterElevated mutation — attacker probe detected0.35 – 0.69
main_organStorm cycle — Torpedo mechanism armed and firing≥ 0.70
recoveringPost-storm recovery — returning to sachsfalling

10. Firewall & SIEM Integration

10.1 Firewall Orchestration

When a protected surface mutates, the perimeter must follow — otherwise a firewall rule references an identity that has changed. The Control Plane can push the updated surface identity to a connected firewall through an operator-triggered sync (POST /v1/firewalls/:id/sync); what it pushes is a fingerprint tag, and automatic per-mutation firewall sync is pilot work, not yet wired to each cycle.

AdapterMechanismWhat Gets Updated
Palo Alto NetworksPAN-OS XML APIDynamic Address Groups — IP + tag pairs per surface fingerprint
Cisco ASACisco ASA REST APIObject group membership per estate
Fortinet FortiOSFortiOS REST APIAddress object update — FQDN or IP-based per mutation cycle

10.2 SIEM Dispatch

Every mutation event and every Torpedo firing is dispatched to the SOC's SIEM, enriched with node, profile, organ state, and audit-chain references so analysts have full context without leaving their console.

TargetFormatCEF Severity
SplunkHTTP Event Collector (HEC)Info (1) → Critical (10) per event type
IBM QRadarCEF over RFC 5424 SyslogInfo → Critical per event type
Generic SIEMRFC 5424 Syslog UDP/TCPStandard syslog severity mapping
stdoutJSON linesLocal capture / container log pipelines
CEF Event ClassTriggerSeverity
acse.estate.status_pushEvery agent push (30s)Info — CEF 1
acse.torpedo.firedEWMA ≥ 0.70 — Torpedo firesCritical — CEF 10
acse.asmp.anomaly_signalASMP MSG-003 received from peerHigh — CEF 8
acse.asmp.defensive_leapASMP MSG-004 estate cascade firesCritical — CEF 10
acse.audit.chain_violationSHA3 hash gap or mismatch detectedCritical — CEF 10

11. Policy Management & the Estate REST API

Policy is defined per estate and applied across nodes in a single call: mutation schedules, EWMA thresholds that govern organ-state transitions, and the active profile for each class of asset. The REST API is the substrate for everything — the dashboard, automation, and third-party integrations all speak to it, and every privileged route carries the API-key + RBAC gate described in Section 3 (MSG-005 is specified and modelled in ProVerif, not yet integrated into this gate).

# Assign a profile across an estate (Admin role required; TEE token not yet enforced — see §3)
POST /v1/estates/{estate_id}/policies/{policy_id}

# Create a policy
POST /v1/policies
{
  "name":              "production-default",
  "schedule_secs":     30,
  "ewma_threshold":    0.70,
  "hunter_threshold":  0.35,
  "active_profiles":   [11]
}

# Trigger an estate-wide defensive leap (Operator role)
POST /v1/invariant/leap

# Read live estate state (any role)
GET  /v1/estates/{estate_id}

# Browse discovery catalog
GET  /v1/discovery/catalog?status=discovered&node_type=windows_server

# Protect a discovered node
POST /v1/discovery/catalog/{id}/protect
Route GroupKey EndpointsMin Role
EstatesCRUD + push + auditOperator (write); ReadOnly (read)
PoliciesCRUD + assign to estateAdmin
DiscoverySources + scan + catalog + protectOperator
SIEM SinksCRUD + testOperator
FirewallsCRUD + syncOperator
InvariantGET global statusReadOnly
UsersCRUD + api_key generationAdmin
Self-RegisterPOST (no Bearer — Registration token)Public

Policy changes are themselves audited — a policy change is a privileged operation and appears in the audit chain alongside the operations it governs.

12. The Cryptographic Audit Chain

The audit chain is the Control Plane's answer to "what happened, and who authorised it?" Every mutation event and every management operation from every node is aggregated into an append-only log, persisted in PostgreSQL, with each entry cryptographically linked to its predecessor by a SHA3-256 hash — the same chaining property as ASMP MSG-001. Removing or altering any entry breaks every subsequent hash, making tampering detectable rather than merely discouraged.

The chain is exposed through a paginated REST API designed for compliance review, incident forensics, and regulatory reporting — the evidentiary spine for SEC, NIS2, DPDP, CERT-In, and HIPAA obligations.

Why a chain, not a log

An ordinary log can be edited by anyone who compromises the host. A SHA3-256-chained log cannot be edited silently: any change invalidates the hash of every entry that followed it. The audit chain is therefore admissible evidence of the estate's behaviour — not just an operational convenience. The CISO's proof: if exfiltration had occurred, the log would contain query events with valid surface fingerprints. Their absence is mathematical proof that no data moved.

Where a TPM 2.0 is present, the chain is signed by a non-extractable key held inside the TPM, and a TPM2_Certify assertion binds that signing key to the TPM. Entries do not each carry a hardware platform quote; there is no PCR / boot-state attestation and no manufacturer endorsement-key chain, and the Control Plane does not yet verify the assertion (#143, #146). So the chain is verifiable for integrity (the SHA3 chain) and is signed by a TPM-held key where one is present — not accompanied by a per-record hardware quote.

13. Deployment & Operations

The Control Plane ships as a single binary with platform-native installers, so it drops into existing enterprise estates without bespoke packaging.

ConcernProvision
WindowsSigned MSI installer — installs as Windows Service
LinuxDEB package — installs as systemd service
PersistencePostgreSQL v14+ (estate state, catalog, audit chain, policies)
SSH discovery depapt install -y libssh2-1-dev (required for ssh source type)
Dashboardpme-console on port 9000/dashboard — no auth, read-only live view
TEE selectionOne constructor parameter; all logic adapter-agnostic
Language / runtimeRust · Axum · zero unsafe in the engine core · 2,130 tests
Self-registrationSet ACSE_REGISTRATION_TOKEN env var on CP to enable

Operationally, the Control Plane is designed to sit alongside an existing stack — firewalls, EDR, SIEM, Zero Trust — rather than replace any of it. It makes those tools more effective by ensuring the surfaces they protect are no longer static.

14. Roadmap — Control Plane v2

v0.2.2 establishes the full operable core including Auto-Discovery, self-registration, and the HCI/Storage Wave 2 connectors. v2 extends toward multi-tenant operation, richer analytics, and executive-grade reporting. The flagship v2 deliverable is the Reporting Module — ACL-scoped historical reports covering EWMA spikes, attack/defence timeline, and audit-log extracts formatted for CISO-level review and compliance submission.

StatusCapability
● Live — v0.4.0API-key + RBAC auth · live dashboard · 22-source Auto-Discovery · agent self-registration · HCI connectors (Nutanix/UCS/vCenter/AzureStack) · Storage connectors (Pure/NetApp/Dell/Brocade/MDS/Ceph) · firewall + SIEM integration · policy & REST API · SHA3 audit chain · Windows MSI + Linux DEB · 2,130 tests
○ Roadmap — v2MSG-005 in the authentication path, bound to a TEE (today MSG-005 gates one protected action; its node-grant property is proven in ProVerif under Dolev-Yao, without the hardware binding; wiring it into the live Control Plane authentication path is planned for a future release) · Reporting module (CISO-level, ACL-scoped, CSV/JSON export) · multi-tenant SaaS · AI/ML anomaly scoring · compliance report generation (SOC 2/ISO 27001/DPDP) · licensing enforcement · auto-update / patch distribution · macOS agent support · client endpoint agent (Win/Linux/Mac) · HSM adapter (PKCS#11) · HashiCorp Vault / AWS Secrets Manager integration

15. Patent Claims Coverage

The Control Plane contributes directly to multiple claims in Patent IN202641070690:

  • ACSE Control Plane (ACSE-CP): The Control Plane itself is claimed as a core component of the system described in the application — the management layer that makes the PME estate operable.
  • Estate management and profile assignment: The mechanism by which profiles are assigned to estates and the Kali Invariant is enforced estate-wide.
  • ASMP/1.0 MSG-005 — Management Binding: The TEE-attested two-gate authentication design discussed in the Stryker MDM weaponisation case (WP-06/CS-01): admin credential (necessary) + TEE-attested token (sufficient). ProVerif confirms CORRECT under Dolev-Yao as a protocol design. MSG-005 is not yet integrated into the live Control Plane authentication path — production auth today is API-key + RBAC.
  • Tamper-evident audit chain: The SHA3-256-chained audit chain stored in PostgreSQL, signed by a key that can be held non-extractably in a TPM where one is present, is the tamper-evident forensic record described in the application. Whether it is admissible as evidence in a given regime is for that regime to determine; no legal-admissibility claim is made here.
  • Auto-Discovery with profile suggestion: The 22-source discovery engine and automated profile assignment are novel contributions. No prior art exists for cryptographic surface mutation profile assignment from enterprise discovery data.
  • Agent self-registration protocol: The zero-config deployment mechanism using a pre-shared registration token to bootstrap a TEE-bound estate key is a novel architectural contribution to the estate management domain.

16. Conclusion

The Control Plane completes the ACSE stack. The four-layer architecture — Control Plane, PME, ASMP/1.0, and the profile layer (ten bio-inspired domain profiles plus TorpedoRay) — is now operable end to end: an estate can be discovered (22 source types), catalogued, onboarded (manually or zero-config), monitored, governed by policy, integrated with existing security infrastructure, and audited, all under a management plane gated today by API-key + RBAC.

The same invariant that protects the data plane already protects reconnaissance of the management plane's data surfaces: the map an attacker builds expires before they can act on it. The stronger management-plane gate — a TEE attestation quote as a second, hardware-rooted factor on every privileged operation — is specified (MSG-005; its node-grant property is proven in ProVerif, the hardware binding is not modelled) but not yet wired into the live path; it is planned for a future release.

Further Reading

WP-00: Master Technical Whitepaper · WP-03: ASMP/1.0 Wire Protocol (MSG-001 through MSG-005 specifications) · WP-07: ProVerif Formal Verification (MSG-005 management binding proof) · WP-08: TorpedoRay — The Transport Channel Profile · WP-10: Every Surface, Every Profile, Every Claim

Patent Status

Application No. IN202641070690 · Indian Patent Office · Inventor: Arul Raj · Published 19/06/2026 · Expedited examination requested (Form 18A) · Publicly searchable on the Indian Patent Office portal.