Glossary

---

A

Accurate

Data integrity principle (ALCOA+): records are correct and verified. In Open Industrial, e-signature hash verification (SHA-256) ensures every audit event is mathematically verifiable.

See also: Governance → Audit Event Log, ALCOA+

ALCOA+

Data integrity framework for regulated industries. ALCOA = Attributable, Legible, Contemporaneous, Original, Accurate. The "+" adds Complete, Consistent, Enduring, Available. Required by 21 CFR Part 11 (FDA) and EU Annex 11 for electronic records in pharma, life sciences, and manufacturing.

See also: Governance Overview

Amendment

A documented correction to an existing record, preserving the original and recording the reason for change. Amendments form a linked chain from the original through each correction.

See also: Amendment Log, Amendment Chain

Amendment Chain

The linked-list history of all amendments to a record, from the original through each correction. Each amendment references the previous, forming a cryptographically verifiable chain.

See also: Amendment, Governance Overview

API Key

A credential used to authenticate API requests. Generated from workspace settings and associated with your user permissions.

See also: REST API Reference, Authentication

Attributable

Data integrity principle (ALCOA+): who performed the action. In Open Industrial, every audit event records the user identity of the person who triggered the workspace change.

See also: Governance → Audit Event Log, ALCOA+

Audit Event

An append-only record of a workspace configuration change, including server timestamp, user identity, action type (create/update/delete/etc.), entity type, and before/after state snapshots. Each event carries a SHA-256 digest and is linked to the previous event via hash chain.

See also: Governance → Audit Event Log, E-signature

Available

Data integrity principle (ALCOA+): authorized users can access records when needed. In Open Industrial, audit events are queryable via REST API and exportable as CSV/JSON for compliance review.

See also: REST API → Audit Logs, ALCOA+

Azi

The AI collaborator in OpenIndustrial. Azi can propose queries, configurations, and changes—but only you can approve them. Named for Azure IoT, representing the bridge between human expertise and AI assistance.

See also: Meet Azi, Governed AI

---

B

Bring Any AI

The principle that OpenIndustrial works with any MCP-native AI, not just Azi. You can connect Claude, Copilot, or any compatible AI to your governed data.

See also: Connect Any AI, MCP

---

C

Complete

Data integrity principle (ALCOA+): the standard asks that actions be captured without gaps. In Open Industrial, configuration commits across the tracked workspace collections are captured automatically in the commit path — the web app does not write audit events.

See also: Governance → Audit Event Log, ALCOA+

Connection

An entry point for telemetry data into OpenIndustrial. Connections link to external data sources like Azure IoT Hub or Event Hub.

Types:

• Azure IoT Hub
• Azure Event Hub

Related: Simulators are separate nodes in the Node Bank that attach to connections for testing. See Simulator Guide.

See also: Connecting Data Guide, Connections Troubleshooting

Consistent

Data integrity principle (ALCOA+): the format is standardized across all records. In Open Industrial, all audit events use the same schema regardless of action type or entity type.

See also: ALCOA+

Consumer Group

An Azure concept for parallel reading from Event Hub or IoT Hub. OpenIndustrial uses dedicated consumer groups to avoid contention with other readers.

Contemporaneous

Data integrity principle (ALCOA+): recorded when it happened, not after the fact. In Open Industrial, audit event timestamps are generated server-side — client-provided timestamps are rejected.

See also: Governance → Audit Event Log, ALCOA+

Control Plane

The governance layer that manages how AI interacts with your industrial data. The Control Plane is where proposals are reviewed, approved, and tracked.

---

E

E-signature

A SHA-256 hash computed from the full audit event payload (timestamp, user, action, entity, before/after state, plus the previous event's hash). Used to verify that an event has not been tampered with. Part of the append-only hash chain that provides tamper-evidence for the entire audit trail.

See also: Governance → Audit Event Log, Accurate

Enduring

Data integrity principle (ALCOA+): records persist for the required retention period. In Open Industrial, each audit log carries a retention setting (0 = indefinite, or a number of days). Storage-level enforcement — a locked Azure Blob immutability policy that prevents deletion for the period — ships with the Provenance blob store.

See also: Governance → Audit Event Log, ALCOA+

---

F

Field-Level Diff

The exact before/after comparison of changed fields in an amendment. When a record is corrected, the field-level diff shows precisely which values were modified, added, or removed — providing auditors with granular change visibility.

See also: Amendment Log, Amendment

---

G

Governance Loop

The core workflow pattern in OpenIndustrial: Propose → Review → Approve → Execute → Audit. Every AI interaction follows this loop, ensuring human oversight at every step.

See also: You Stay in Control, Control Plane

Governed AI

AI that operates under human oversight and approval. In OpenIndustrial, this means:

• AI proposes changes
• Humans review and approve
• Nothing happens without explicit consent
• All actions are audited

The Three Principles:

1. "Bring any AI" (MCP-native)
2. "Nothing without approval"
3. "Fire us, keep running"

See also: Vision, Execution Model

---

I

Impulse

A single telemetry event or data point entering the system. Impulses flow through connections and become available for querying in surfaces.

Example: A temperature reading from a sensor is one impulse.

See also: Execution Model

---

K

KQL (Kusto Query Language)

The query language used in OpenIndustrial for data analysis. KQL is powerful yet readable, designed for time-series and log data.

---

L

Legible

Data integrity principle (ALCOA+): records can be read and understood by humans. In Open Industrial, audit events are stored as structured JSON with before/after state diffs, exportable as human-readable CSV.

See also: Governance → Audit Event Log, ALCOA+

LIVE STREAM

The panel at the bottom of the workspace canvas that shows real-time telemetry events (impulses) as they arrive. Each impulse displays with a cyan border, connection ID, and timestamp.

See also: Impulse, Get Data Flowing

Basic example:

TelemetryTable
| where timestamp > ago(1h)
| summarize avg(temperature) by bin(timestamp, 5m)

See also: KQL Basics, Warm Queries Troubleshooting

---

N

Node Bank

The panel on the left side of the workspace canvas that contains draggable node types. At workspace level, shows Simulator, Connection, and Surface. Inside surfaces, shows Warm Query and other surface-level nodes.

See also: Create Connection, Canvas Guide

---

M

MCP (Model Context Protocol)

An open protocol for connecting AI agents to external data and tools. OpenIndustrial implements MCP to enable "Bring Any AI" functionality.

Compatible AI:

• Claude Desktop (native)
• VS Code with Copilot (via adapter)
• Any MCP-native AI agent

See also: Connect Any AI, MCP Guides

---

O

Original

Data integrity principle (ALCOA+): this is the first-generation record, not a copy. In Open Industrial, the stored event IS the original record — written append-only, never updated in place. Storage-level immutability under a locked Azure Blob policy ships with the Provenance blob store.

See also: Governance → Audit Event Log, ALCOA+

Original Record Preservation

ALCOA+ principle that original records are never modified after creation. In Open Industrial, amendments create new linked records rather than overwriting the original. The full chain from original through every correction is always retrievable.

See also: Amendment Log, Original, Amendment Chain

---

P

Proposal

An AI-generated suggestion awaiting human approval. Proposals can include:

• New queries
• Configuration changes
• Connection settings
• Any workspace modification

See also: You Stay In Control, Proposal Buckets

---

R

Reflex

A pattern learned from repeated approvals. When you consistently approve similar actions, the system can propose a reflex—a pre-approved pattern for future actions.

Note: Reflexes are a planned feature. See Roadmap.

See also: Reflex Memory

Reason-for-Change

Required documentation when amending a record. Includes a validated reason category (from a configurable dropdown) and free-text justification. Categories include: data-entry-error, transcription-error, equipment-malfunction, protocol-deviation, recalculation, delayed-entry, and other.

See also: Amendment Log, Amendment

---

S

Schema

The structure of your data—field names, types, and relationships. Schemas are observed from incoming data and can be promoted to formal definitions.

See also: Schema Promotion

Schema Promotion

The process of formalizing observed data patterns into governed schemas. Promotion means "yes, this structure matters—treat it as official."

Note: Schema Promotion is in development. See Roadmap.

Surface

A queryable workspace containing warm queries. Surfaces connect to data connections and provide the environment for building and running queries.

Key concept: Think of a surface as a workbench where you build queries against your data.

See also: Surfaces Guide, Surfaces Troubleshooting

---

T

Telemetry

Time-series data from devices, sensors, or applications. This is the raw material that flows through connections into OpenIndustrial.

Examples:

• Temperature readings
• Pressure measurements
• Machine status events
• Production counts

---

W

Warm Query

A saved KQL query with its own API endpoint. "Warm" because:

• It's ready to run (not cold/archived)
• It returns fresh results on demand (not hot/streaming)
• It has a defined schema and output format

Creating warm queries:

1. Work with Azi to build the KQL
2. Save with a descriptive name
3. Access via REST API or MCP

See also: Warm Queries Guide, Warm Queries Troubleshooting

Workspace

A container for your OpenIndustrial resources—connections, surfaces, queries, and team members. Workspaces provide isolation and permission boundaries.

---

Quick Reference Table

Term	Short Definition
ALCOA+	Data integrity framework for regulated industries
Amendment	Documented correction with reason-for-change
Amendment Chain	Linked-list history of all amendments to a record
Audit Event	Append-only, hash-chained record of a workspace configuration change
Azi	AI collaborator that proposes, you approve
Connection	Entry point for data (Azure IoT Hub)
Control Plane	The governance layer for AI interactions
Governance Loop	Propose → Review → Approve → Execute → Audit
E-signature	SHA-256 hash verifying audit event integrity
Field-Level Diff	Before/after comparison of changed fields
Governed AI	AI that requires human approval
Impulse	Single telemetry event
KQL	Query language for data analysis
LIVE STREAM	Panel showing real-time telemetry
MCP	Protocol for connecting any AI
Node Bank	Panel with draggable node types
Proposal	AI suggestion awaiting approval
Reason-for-Change	Required documentation when correcting a record
Surface	Execution boundary for warm queries
Warm Query	Saved query with API endpoint
Workspace	Container for resources and team

---

Related

• Meet Azi - Understanding your AI collaborator
• Vision - Concepts and philosophy
• Troubleshooting - When things go wrong
• Reference - Technical details