FactoryTalk DataMosaix is Rockwell Automation's industrial DataOps software. Extractors copy data from plant systems such as historians, SCADA, batch and OPC UA servers to the cloud, where it is given context in a data model and opened to people, BI tools and applications across sites. This article follows Rockwell's public product page, the FactoryTalk DataMosaix SaaS online help and its release notes up to version 1.13.01, read on 15 November 2026.
1. What Rockwell means by industrial DataOps
Rockwell's product page describes DataMosaix as cloud-based software for multi-site, enterprise-wide data access by people and applications: data is extracted from many sources, contextualised, and access to it is managed in one place. The introduction of the online help calls it an Industrial Data Operations platform that brings IT, OT and ET data together, with standard connectors that move data from on-premise or cloud systems into a single source of truth.
The flyer
A historian sits close to the control system and records each tag's value with a time stamp (What is a process historian). In Rockwell's documentation the historian is one of the sources DataMosaix reads.
2. Deployment: SaaS in FactoryTalk Hub, extractors on the edge
The online help states that FactoryTalk DataMosaix SaaS runs as part of FactoryTalk Hub and is available only as Software as a Service in the cloud. Its prerequisites are a tenant organisation in FactoryTalk Hub and a valid subscription to the product. The product page adds that the SaaS has a fully managed option, so plant staff do not have to manage the technology or its updates.
The getting-started guide describes the architecture in four parts:
- FactoryTalk Hub is where users sign in. It holds the users, their entitlements and the basic decision whether a user may open the application.
- The Management Console, opened from the FactoryTalk Hub tile, is the Rockwell application where fine-grained access is set.
- Cognite Data Fusion (CDF) is the other part of the product, reached from the console through the Main DataMosaix link.
- Extractors are deployed on the edge and send the data to the cloud.
The flyer of March 2024 offered DataMosaix as SaaS or private cloud. Rockwell documents the private cloud product separately: the 8.03.00 release notes (December 2025) state that FactoryTalk DataMosaix Private Cloud is now rebranded as UnifyTwin. This article covers the SaaS product, released from 1.00.00 (July 2023) to 1.13.00 (August 2026) and 1.13.01.
3. Data sources and extractors
Extractors connect to source systems and push the data in its original format to CDF. The online help lists these extractor and connector types, with the versions the configuration screen supports:
| Extractor or connector | Source | What the documentation says |
|---|---|---|
| OPC UA | OPC UA servers | version 2.18 |
| FactoryTalk Historian (OSI PI) | FactoryTalk Historian | version 4.1.0; writes time series directly, without raw tables |
| Asset Framework (PI-AF) | PI Asset Framework | version 1.5.1 |
| FactoryTalk View SE | View SE DataLog, FTLD, Alarms and Events | numeric historical data only; DataLogPro not supported; the start date of the first history range cannot be changed later |
| FactoryTalk Alarms and Events | FactoryTalk Alarms and Events | listed as a connector |
| FactoryTalk Batch | FactoryTalk Batch | listed as an extractor |
| FactoryTalk Optix | a FactoryTalk Optix project | version 1.03.00; string, numeric and discrete time series |
| ODBC | any ODBC-compatible database | version 3.0.0 or later |
| Fiix | Fiix | added in 1.13.00 |
| Custom | other systems | for extractors not in the list |
The product page groups the sources more broadly: data gateways, historians, MES/MOM, PLM and digital engineering, and enterprise systems. At the control layer it points to FactoryTalk Optix, which can add OT context and assemble the data into a flexible common information model. The data flow page names historians and PLCs as edge sources. The list has no extractor for a PLC protocol as such, so PLC values arrive through a listed system such as an OPC UA server, an Optix project or a historian.
Registering an extractor needs the Data Configurator role, and each extractor has a secret that can be regenerated. A new extractor or M2M application gets a broad default role, so Rockwell strongly recommends a custom role with only the capabilities needed (least privilege).
4. From raw data to a contextualised model
The getting-started guide shows two paths. On the first, for example data from an SQL database, the extractor writes into raw tables in the cloud, and the engineer writes one or more transformations that turn the raw rows into assets, events, time series or properties in an object model. On the second, used by the historian extractor, time series go straight into CDF time series without the raw step. Data sets group data for one purpose, and they can also limit access to authorised people.
In the online help, assets represent real-world objects and are built in a hierarchical structure. Release 1.13.00 announces that the Asset Model feature and its APIs will be deprecated by 2027, and the online help recommends the Flexible Data Model Framework for all new implementations. Models built through FactoryTalk Optix also feed the flexible data model; release 1.13.00 notes that large model changes are sometimes not pushed correctly. FactoryTalk VantagePoint has reached end of life, and DataMosaix can import a VantagePoint data model from an .IIP file.
5. Who uses the data
- BI tools. Power BI and Grafana can be configured to use a DataMosaix project as a data source. The Virtual PostgreSQL Data Interface opens the data to Tableau (Desktop and Server), Power BI, Microsoft Excel and SQL Server Reporting Services. Release 1.13.00 lists its limits, among them SELECT queries only, a time series ID or external ID required when reading data points (last 24 hours by default), and an error above one million rows.
- Applications. Registering an application, which needs the Data Scientist role, gives a client ID and secret for reading project data in CDF. The REST APIs can be explored through Swagger UI.
- Rockwell solutions built on the platform. The getting-started guide names FactoryTalk Energy Manager, Batch Performance Analytics and Asset Intelligence for Mining.
- People and AI. The product page lists root cause analysis, dashboards and KPIs compared by plant, line and region, a built-in Charts application and data models for AI agents.
6. Security and access as documented
Access is set at two levels. FactoryTalk Hub decides who may open the product. In the Management Console, members get roles; the seven default roles are Application Developer, Applications, Application User, Data Configurator, Data Scientist, Extractors and Project Admin, and the help calls them a starting point for an implementation. Each role carries capability groups, and a capability combines a resource type, a scope and actions, for example timeseries:read. Only Project Admins can switch features on or off for a project.
Applications and extractors are machine (M2M) identities, authenticated by Auth0 with their role IDs added to the token; release 1.13.00 warns that more than 350 groups makes the token exceed 8 KB and fail. The audit log records who did what to which object, including sign-in attempts, data access and configuration changes; entries can take up to an hour to appear, users cannot change or delete them, and a filtered view can be exported to CSV.
7. Next to FactoryTalk Historian, FactoryTalk Optix and a plant SCADA
In the architecture Rockwell documents, the plant systems remain the sources. FactoryTalk Historian keeps the plant time series and PI Asset Framework, which extractors read. FactoryTalk View SE keeps running the operator screens while its connector reads data logs and alarms (FactoryTalk View SE, ME and Optix). FactoryTalk Optix sits at the control layer and sends modelled data through its connector. MES/MOM is one of the source groups (What is MES). The wider Rockwell family: What is Rockwell Automation; the SCADA layer: What is SCADA.
What we check first
This is TDL's practice, not a Rockwell Automation requirement. Before a plant connects to DataMosaix:
- We list where the data lives today (FactoryTalk Historian and PI AF, View SE data logs and alarms, FactoryTalk Batch, OPC UA servers, databases) and compare each extractor version with Rockwell's supported versions list.
- We check tag and point naming across lines and sites. A model is easier to build when the same equipment has the same names everywhere.
- We trace the network path from the extractor host to the cloud: which zone it sits in, which firewall it crosses and who owns that firewall. Zones and conduits are covered in IEC 62443, port rules in firewalls for FactoryTalk and Logix. The pages we read list no ports or addresses, so we confirm them with Rockwell before the firewall request.
- We agree the IT/OT boundary with both teams, so that the extractor host does not become a path from outside into the control network.
- We plan roles before the first extractor: custom roles with only the needed capabilities, who keeps the secrets, and who owns the FactoryTalk Hub tenant.
How TDL approaches it
TDL Industrial Automation is a System Integrator in the Rockwell Automation PartnerNetwork. Our Rockwell Automation page covers the plant-side systems DataMosaix reads from, among them FactoryTalk View SE, FactoryTalk Optix, FactoryTalk Batch and FactoryTalk Historian, and our SCADA and historian page lists the historians we work with.
Frequently asked questions
Does FactoryTalk DataMosaix replace FactoryTalk Historian?
Rockwell's documentation treats the historian as a source. The FactoryTalk Historian (OSI PI) extractor and the PI-AF extractor copy its time series and asset framework to the cloud; the getting-started guide counts historians among the edge data sources.
Can FactoryTalk DataMosaix run on-premises?
The SaaS online help says the product is available only as Software as a Service in the cloud, with the extractors deployed on the edge. Rockwell documented a separate private cloud product, which its 8.03.00 release notes say is now rebranded as UnifyTwin.
What is Cognite Data Fusion in DataMosaix?
Rockwell's getting-started guide describes Cognite Data Fusion (CDF) as the other part of FactoryTalk DataMosaix SaaS, next to the Management Console. Extractors push data to CDF, and applications and BI tools read the project data there.
Can Power BI or Excel read DataMosaix data?
Yes. Power BI and Grafana connect to the CDF project, and the Virtual PostgreSQL Data Interface supports Tableau, Power BI, Microsoft Excel and SQL Server Reporting Services, with SELECT queries only.
Does DataMosaix read FactoryTalk View SE data?
Yes, through the FactoryTalk View SE connector. Release 1.13.00 lists DataLog, FTLD and Alarms and Events as sources, numeric history only; DataLogPro is not supported.
Sources
- Rockwell Automation, FactoryTalk DataMosaix, product page (page metadata 26 August 2026)
- Rockwell Automation,
FTALK-SP020C-EN-P , Industrial DataOps: Accelerate your data journey (FactoryTalk DataMosaix flyer), March 2024, p. 1 - Rockwell Automation, FactoryTalk DataMosaix SaaS online help (current, last update October 2026): Introduction, Prerequisites, Extractors, Supported Extractor Versions, Register an FactoryTalk Optix Connector and Asset Creation
- Rockwell Automation, FactoryTalk DataMosaix SaaS online help: Applications, Analyze Data, Virtual PostgreSQL Data Interface, Default Roles, Capability Groups, Features, What is M2M? and Auditing
- Rockwell Automation, FactoryTalk DataMosaix SaaS Getting Started Guide: FactoryTalk DataMosaix SaaS Architecture and Typical Data Flows
- Rockwell Automation, FactoryTalk DataMosaix SaaS release notes: 1.13.01 (current) and 1.13.00
- Rockwell Automation, FactoryTalk DataMosaix Private Cloud release notes, 8.03.00, December 2025
Read 15 November 2026. This article paraphrases the sources and does not quote them; read the original documents before designing or working on a system.