Data Fusion Services

Data Fusion Services, or DFS, is the operational data layer for FactVerse. It connects source systems, maps source data to operational targets, checks data quality, and prepares governed datasets for Inspector, AI Agent, Twin Engine, Designer workflows, and reporting.

Use this section when you need to make real operational data usable in a digital twin workflow.

Data path

Role paths

Role	Start with	Confirm before handoff
Source owner	Getting Started with DFS	Connector reachability, source meaning, units, and update cadence.
Data engineer	DFS Lite	Mapping identity, sync health, failed rows, and data quality state.
Data steward	DFS Pro	Dataset owner, lifecycle, profile, validation, lineage, and review queues.
Solution implementer	DFS Workflows	Downstream evidence contract, acceptance criteria, and handoff record.
Developer	DFS Reference	Connector type, mapping fields, permissions, and API surface.

When to use DFS Lite

Use DFS Lite when the job is to connect and operate data feeds.

Typical tasks:

• connect OT, IoT, enterprise, file, API, and database sources;
• test a connector before saving it;
• start, pause, or sync a connector;
• browse a source hierarchy;
• preview sample values;
• map source fields to assets, points, and target fields;
• check sync history, throughput, quota, and data quality.

Start with Getting Started with DFS if this is the first source in the tenant.

When to use DFS Pro

Use DFS Pro when the job requires governed data assets, repeatable data processing, review, or multi-source fusion.

Typical tasks:

• create datasets from connectors, imports, extractions, or fusion outputs;
• validate datasets and assign a data steward;
• review schema versions, lineage, profile, and change impact;
• create reusable processing methods;
• run fusion tasks across multiple datasets;
• review conflicts, low-confidence outputs, and rejected rows;
• use reviewed outputs in AI Agent workflows or BI dashboards.

Start with DFS Pro Datasets when a connector feed needs to become a reusable data asset.

Product boundary

DFS prepares and governs the data foundation. Customer source systems remain the systems of record unless a project explicitly defines a different ownership model.

AI-assisted mapping and LLM-assisted fusion are reviewable suggestions. The reviewed workflow, assigned owners, and audit trail decide whether a mapping, fusion result, or rejected-row correction is accepted.

Common DFS workflow

Phase	DFS area	Result
Connect	DFS Lite connectors	A source is reachable and can be tested.
Inspect	Browse and preview	Candidate fields, tags, rows, or topics are visible.
Map	DFS Lite mappings	Source values are bound to target entities and fields.
Sync	Connector sync	New values are ingested with run history.
Check	Data quality	Completeness, timeliness, accuracy, and quota are visible.
Govern	DFS Pro datasets	Data is packaged as a stewarded asset with lifecycle and versioning.
Fuse	DFS Pro fusion tasks	Multiple datasets are merged with conflict handling.
Review	Review and rejection queues	Uncertain outputs and rejected rows are resolved by a reviewer.

Read next

Page	Use
Getting Started with DFS	Create the first connector, test it, map one field, sync data, and check quality.
DFS Workflows	Choose an end-to-end sequence when DFS work spans source connection, governance, review, and consuming FactVerse applications.
Prepare DFS Data for AI Agent Workflows	Turn DFS Lite sources and DFS Pro datasets into reviewed Agent evidence.
DFS Lite Connectors	Operate connector lifecycle.
Mapping Source Fields	Bind source data to operational targets.
DFS Pro Datasets	Create governed datasets from connector or imported data.
Dataset Lifecycle	Steward, validate, version, and deprecate reusable datasets.
Fusion Tasks	Combine multiple datasets with reviewable fusion logic.
Governance Studio	Build repeatable dataset and method pipelines.
DFS Pro BI Reports	Build reports from governed datasets.
DFS Recipes	Choose a task-oriented path for facility data, predictive maintenance preparation, AI Agent datasets, fusion, or rejected-row recovery.
Prepare Signal History for Predictive Maintenance	Prepare clean time-series data for predictive maintenance workflows.
Create an AI Agent-Ready Dataset	Package reviewed data for Agent workflows.
DFS Reference	Check connector types, mapping fields, permissions, and API surfaces.
DFS API Surface	Plan API-level integrations around DFS Lite, DFS Pro, and BI.