Must all source data be copied into one PrecisionX database?

No. High-cardinality telemetry, video, logs, DCIM, CMMS, VMS/PACS, carrier and customer data may remain in qualified domain stores. PrecisionDCOS establishes identity, quality, lineage, context and governed interfaces across them.

Can a customer use its own data lake or AI platform?

Yes. Tenant-scoped APIs, streams, object delivery and model-ready datasets can feed customer cloud, lakehouse, on-premises, BI, digital-twin or AI/ML environments.

Can an AI model directly control the facility?

Not by default. AI is advisory unless an independently engineered action follows the same command, approval, local-interlock, feedback and audit requirements as any other equipment-changing request.

Is customer workload content collected?

Not by default. Models, prompts, training data, files and application payloads require a separate explicit processing and security agreement.

DCOS Data FabricDCOS-DAT

DCOS Data Fabric

Governed operational data for analytics, AI & twins

Governed operational data for operators, customers, analytics, AI and digital twins — normalized and contextualized without forcing every source into one database.

Request an Architecture Review Configure a Reference Architecture

Shape: Normalized & contextual
Access: APIs · Streams · Exports
Governance: Authority & residency
Consumers: Analytics · AI/ML · Twins

DCOS Data Fabric converts facility, network, connectivity, security, compute, storage, cluster, customer and workflow records into contextualized and reusable operational information without forcing every source into one database. Values arrive already carrying identity, time and quality; the fabric contextualizes them with dependencies and events, and exposes tenant-scoped APIs, streams and exports so analytics, AI/ML and digital-twin systems consume operational reality without eroding authority or residency.

Outcomes

What this component is designed to make true when it is in place.

Data you can trust

Every value carries identity, time and canonical quality, so downstream models reason on reliable inputs.

Governed access

APIs, streams and exports are scoped by authority and residency, not open firehoses.

AI-ready context

Dependencies and events contextualize data so AI/ML and twins see relationships, not just points.

Data model

Normalization

Vendor signal is mapped to the shared model with stable identity, units and quality states.

Contextualization

Values are related to assets, services, paths and events so context travels with data.

Governed APIs

Authority-scoped REST and query APIs with rate control and full audit trail.

Streaming

Real-time streams for events and telemetry with durable delivery semantics, backpressure and replay where configured.

Exports

Governed batch exports for warehouses, lakes and reporting under residency policy.

Digital-twin feeds

Structured feeds that keep external digital twins synchronized with operating reality.

Data classes and delivery

Operational information is organized into governed classes and published through many methods with durable delivery semantics.

Data classes

Operational state
High-cardinality performance
Events and incidents
Logs and traces
Asset/configuration/dependency
Work/change/maintenance
Evidence and reports
Video/media
Customer/workload content excluded by default

Publication methods

REST
GraphQL where approved
Webhooks
MQTT/Sparkplug
OPC UA
Batched HTTPS/NDJSON
CSV / JSON / Parquet
Kafka-compatible event bus
Object storage / lakehouse
Digital-twin interface
Reports and evidence packages
Durable delivery semantics, backpressure and replay where configured

AI operating principle

Use AI to increase operator context, speed, consistency and predictive insight. Do not use AI to erase source quality, bypass local authority, replace required qualified judgment or turn an analytics endpoint into a plant-control path.

Data rights

Customer operational data rights follow the agreement
No customer-identifiable cross-customer training without express authorization
Workload content excluded by default
Tenant-scoped external endpoints
No direct operational-database access
Open handback in practical formats
Secrets transferred separately
Retention and deletion controlled by contract

Operating contract

How this component upholds the shared platform contract.

What this component upholds

Data carries stable identity, source/receipt time and canonical quality end to end.
Access authority is declared and enforced per API, stream and export.
Residency and retention policy govern where operational data lives and how long.
Every access is logged as evidence supporting assurance and troubleshooting.

Access surfaces

Access surfaces are indicative; exact shapes, latency and governance are set per data agreement.
Surface	Shape	Latency	Governance
REST / query API	Normalized, contextual	Request/response	Authority-scoped + audited
Event stream	Lifecycle events	Real-time	Scoped subscriptions
Telemetry stream	Quality-aware series	Near real-time	Backpressure-controlled
Batch export	Warehouse/lake	Scheduled	Residency-governed
Twin feed	Structured model	Continuous	Authority + residency

Access surfaces are indicative; exact shapes, latency and governance are set per data agreement.

Scope & boundaries

What this component is — and deliberately is not — responsible for.

Data Fabric governs operational data; it does not become the authoritative system of record for federated domains.
Access is always authority- and residency-scoped rather than an unrestricted export.
It provides trustworthy inputs for AI/ML; it does not itself own model training or business logic.

Technical FAQs

Frequently asked questions

Explore the family

Bring DCOS Data Fabric into your reference architecture

We map this component to your sites, existing systems and assurance obligations, then agree a delivery path.

Request an Architecture Review Review Configuration Options