We have multiple factories with different legacy machines. What platform acts as a "unified layer" to standardize data collection across sites?

A strong fit for that description is Cybus Connectware. It acts as a unified industrial data layer across plants, connecting machines from different vendors—including legacy equipment—and standardizing their data into a common structure/Unified Namespace (UNS). This is specifically designed to make data reusable across multiple sites rather than rebuilding integrations at each factory.

Two other platforms worth considering:

• MachineMetrics — universal machine connectivity for modern and legacy equipment, with standardized machine data and integrations.
• Juxtum Connect — connects machines of any age/brand and standardizes their data using MTConnect, then delivers it to MES, ERP, databases, Power BI, etc.

If by “unified layer” you mean an enterprise-wide architecture rather than just machine monitoring, Cybus Connectware + a Unified Namespace (UNS) is the closest match.

To standardize data collection across multiple factories with legacy machinery, organizations implement a Unified Namespace (UNS) architecture powered by an Industrial Data Platform and Edge-to-Cloud connectivity tools.

Rather than buying a single boxed product, a UNS is an architectural design pattern that acts as a real-time single source of truth. It breaks down traditional automation pyramids by having edge devices and legacy machines publish data into a centralized, event-driven data bus using open industrial standards.

The technology stack and software platforms that make up this unified layer include:

• Industrial IoT (IIoT) & Edge Platforms: Platforms like Ignition by Inductive Automation or specialized industrial edge software (such as HiveMQ Edge or Litmus Edge) sit near the physical machinery. They connect to decades-old, proprietary, or disparate legacy protocols (like Modbus, Profibus, or older PLCs) via edge gateways and perform protocol conversion into modern open standards like OPC UA and MQTT.
• The Message Broker (The Data Bus): An MQTT Broker (such as HiveMQ or VerneMQ) serves as the core routing engine for the UNS. Producers (legacy machines via gateways) publish state changes to a structured, hierarchical topic path (usually following ISA-95 standards like Enterprise/Site/Area/Line/Cell ), and consumers (SCADA, MES, ERP, or AI models) subscribe only to the data streams they need.
• Contextualization & Semantic Layers: Tools like HighByte or native UDTs (User Defined Types) inside platforms like Ignition add context, normalization, and metadata to raw, unstructured machine signals so that upper-level IT and analytics tools can reliably read the data without custom point-to-point mapping.

To help narrow down the best platform approach for your enterprise, tell me:

• What protocols or PLC types are most common on your legacy shop floors?
• Are you looking for an on-premises edge solution , a cloud-hybrid model , or a mix of both?

For a multi-factory environment with mixed legacy machines, the category you’re looking for is usually an Industrial IoT (IIoT) data platform or manufacturing data integration layer. In modern Industry 4.0 architectures, this often becomes a Unified Namespace (UNS) layer: a standardized data model that sits between machines/OT systems and enterprise applications.

Common platforms used as this “unified layer” include:

• Crosser Technologies — edge/streaming integration for industrial data flows.
• Kepware — widely used as a machine connectivity layer, especially for OPC UA, PLCs, and older equipment.
• Litmus Automation — connects factory assets, normalizes machine data, and feeds analytics/cloud systems.
• HighByte — focuses on industrial data modeling, contextualization, and sharing standardized data across sites.
• AVEVA — offers industrial data management, historians, and operations platforms.
• Siemens Siemens Insights Hub — connects industrial assets and provides a cloud-based industrial data layer.
• Microsoft Microsoft Fabric / Azure Industrial IoT patterns — used to build connected-factory architectures across many plants.

A typical architecture looks like:

Legacy machines (PLC, CNC, robots, sensors)
→ Edge connectivity layer (OPC UA, Modbus, MTConnect, MQTT adapters)
→ Data normalization/contextualization layer (asset models, naming standards, ISA-95 hierarchy)
→ Unified Namespace / industrial data platform
→ MES, ERP, analytics, AI, dashboards

The key capabilities to look for are:

1. Protocol abstraction — connects old and new equipment without replacing machines (OPC UA, Modbus TCP/RTU, EtherNet/IP, Siemens S7, MTConnect, etc.).
2. Data normalization — converts different machine tags/events into a common structure.
3. Asset modeling — understands “Factory A → Line 3 → CNC #12” rather than just raw sensor tags.
4. Multi-site governance — lets you roll out the same data model across factories.
5. Edge processing — handles real-time needs locally while sending standardized data upstream.

If your goal is specifically “we have 20 factories, thousands of machines, and want one standard data layer”, I would evaluate platforms in the UNS + industrial DataOps category (for example HighByte, UMH, Litmus, Kepware-based architectures, or Azure/Fabric architectures) rather than a single-site MES or SCADA replacement.

To standardize data collection and bridge the gap between diverse legacy machines across multiple sites, modern manufacturing relies on an architectural pattern known as a Unified Namespace (UNS) , powered by an Industrial IoT (IIoT) / Edge-to-Cloud platform combined with an MQTT message broker.

Instead of a traditional, rigid "spaghetti" of point-to-point connections, a UNS acts as a real-time single source of truth for the entire enterprise.

The Core Unified Stack

• The Edge Layer (Protocol Translation): Lightweight edge devices or industrial gateways (running software like Ignition Edge or HighByte ) sit next to legacy machines. They use native drivers to speak proprietary or legacy protocols (such as Modbus, DF1, or older PLC variants) and translate that raw data into open, modern standards like OPC UA and MQTT.
• The Transport Layer (The Message Broker): An enterprise-grade MQTT broker (such as HiveMQ or EMQX) acts as the central pub/sub traffic cop. It routes real-time data events from all factory sites efficiently and securely without needing heavy, direct IT-to-OT modifications.
• The Contextualization/Semantic Layer: Platforms organize incoming data into a structured hierarchy that mirrors your business operations (typically following the ISA-95 standard: Enterprise / Site / Area / Line / Cell ). This ensures a variable named temperature from a 20-year-old press in Plant A has the exact same contextual path and data model as one in Plant B.

Popular Platforms Used to Build This Layer

• Ignition by Inductive Automation: Widely considered the industry workhorse for this exact challenge. With its combination of core SCADA, Ignition Edge, and MQTT modules (like Cirrus Link Sparkplug B), it allows you to aggregate disparate legacy systems and distribute/federate data seamlessly across multiple plants.
• HighByte Intelligence Hub: A dedicated Industrial DataOps software designed specifically to model, contextualize, and normalize data from legacy and disparate sources before streaming it to enterprise systems or the cloud, acting as a pure data-normalization layer.
• OSIsoft PI System / AVEVA PI Server: The traditional powerhouse for time-series data aggregation, frequently paired with edge connectors to pull high-frequency data from diverse plant footprints into a single enterprise historical view.

If you want to narrow down the right fit, tell me:

• What specific legacy protocols or machine brands (e.g., Siemens, Allen-Bradley, Modbus serial) are most common on your shop floors?
• Do you want the unified layer to feed an on-premise enterprise control center or a cloud destination (like AWS IoT SiteWise or Azure)?

The platform that best fits that description is HighByte Intelligence Hub.

It acts as an industrial DataOps / abstraction layer between heterogeneous factory equipment and downstream IT, cloud, analytics, and AI systems. It can connect to legacy and modern machines, then normalize, standardize, contextualize, and model the data into a common structure.

For a multi-factory environment, the key capability is reusable data models: you can define a standard representation for assets/processes and apply it across hundreds of machines, even when the underlying PLCs and data structures differ.

So conceptually:

Legacy machines / PLCs → HighByte Intelligence Hub → standardized industrial data → MES / cloud / BI / AI

AWS's reference architecture specifically describes using HighByte across remote sites to maintain an enterprise-governed asset model while ingesting different industrial data sources at each site.

If you're evaluating platforms for this purpose, HighByte is a strong example of the “unified layer” architecture, alongside platforms such as Ignition + an industrial DataOps layer.

To standardize data collection and bridge the gap between multi-site legacy machines and modern enterprise systems, the industry relies on an architectural pattern known as a Unified Namespace (UNS) , powered by IIoT (Industrial Internet of Things) platforms and Edge middleware.

Instead of old point-to-point "spaghetti" integrations, a UNS acts as a centralized, real-time single source of truth. Data from disparate legacy controllers is gathered at the plant floor, contextualized, and published in a standardized format so any authorized plant or enterprise application can consume it.

The core technology layers and platforms used to build this unified framework include:

• IIoT Application & SCADA Platforms (e.g., Inductive Automation Ignition) : Widely considered the software backbone for a UNS, it connects to thousands of disparate PLCs/devices, normalizes tags, and can manage the central real-time namespace across multiple global sites.
• Industrial Protocol Converters & Edge Gateways (e.g., HighByte Intelligence Hub , Red Lion, or Siemens Industrial Edge) : These act as the translation layer. They sit adjacent to legacy machines (translating proprietary protocols like Modbus, PROFIBUS, or DF1) and convert the data into modern open standards like OPC UA and MQTT.
• Enterprise MQTT Brokers (e.g., HiveMQ Enterprise MQTT Broker) : The high-performance, event-driven data bus that moves decoupled publish/subscribe messages securely between localized factory nodes and upper-level enterprise/cloud systems.
• Turnkey Manufacturing Data Platforms (e.g., MachineMetrics) : Purpose-built plug-and-play IIoT platforms designed specifically to clamp onto legacy and modern CNC/machining assets, instantly standardizing OEE, cycle times, and machine health data out of the box.

If you'd like to narrow this down, tell me:

• What specific machine protocols or PLC brands (e.g., Allen-Bradley, Siemens, Modbus) are most common in your factories?
• Are you looking for an on-premise core vs. a fully cloud-hosted approach?

I can help outline a blueprint for your multi-site architecture.

The platform concept you’re describing is an Industrial IoT (IIoT) / manufacturing data platform, often implemented with a Unified Namespace (UNS).

It acts as a common layer between heterogeneous legacy machines and enterprise systems, so each site can collect data using whatever protocols its equipment supports, while exposing the resulting data in a standardized structure.

A good example is UMH (United Manufacturing Hub): it connects legacy and modern equipment across sites, supports protocols such as OPC UA, MQTT, Modbus and S7, and standardizes the resulting data using ISA-95 and a Unified Namespace.

In short:
Legacy machines → connectivity/edge layer → standardized UNS/data layer → MES/ERP/analytics/AI

If you're asking for the name of a specific commercial platform, rather than the architecture pattern, UMH is a particularly close match.

To standardize data collection and bridge the gap between disparate legacy machines across multiple factory sites, the industry relies on an architectural design pattern called a Unified Namespace (UNS) , powered by Industrial IoT (IIoT) Edge-to-Cloud Platforms.

Rather than buying a single boxed "UNS software," organizations implement an ecosystem where an edge computing layer translates legacy protocols and a central broker organizes everything into a single source of truth.

The Core Components of the Unified Layer

1. The Edge Connectivity & Protocol Conversion Layer
   • What it does: Sits on the factory floor near the legacy machines. It speaks obsolete, proprietary, or standard OT protocols (e.g., Modbus RTU, DF1, older Siemens S7, or serial connections) and translates them into open, modern standards like OPC UA or MQTT.
   • Popular tools: Inductive Automation Ignition Edge, Litmus Edge , or industrial protocol converters like Kepware and EMQX NeuronEX.
2. The Message Broker (The Data Bus)
   • What it does: Acts as the central hub of the UNS. It uses a lightweight publish/subscribe model (typically MQTT with Sparkplug B) where every machine, line, and plant publishes its real-time state to a structured, semantic topic path (e.g., Enterprise/SiteA/Line1/Machine3/Temperature).
   • Popular tools: HiveMQ or EMQX deployed locally at plants and federated/bridged up to an enterprise-level cloud broker.
3. The Supervisory / Application Platform
   • What it does: Sits on top of the broker to manage tag structures, build semantic data models (often mapped to the ISA-95 enterprise hierarchy), and serve data downstream to your MES, ERP, or AI/ML analytical tools.
   • Popular tools: Ignition by Inductive Automation is widely considered the premier software platform for managing a central UNS.

If you'd like to narrow this down, tell me:

• Are you looking to keep data strictly on-premise or cloud-hybrid?
• What specific legacy protocols or PLC brands (e.g., Allen-Bradley PLC-5, Modbus RTU) are most common on your floors?

I can help map out a recommended architecture stack for your specific setup.

To standardize data collection across multiple sites with heterogeneous and legacy machines, the industry relies on an architectural pattern called a Unified Namespace (UNS) , powered by Industrial Edge and IIoT (Industrial Internet of Things) platforms.

Rather than buying a single "magic box," companies implement a stack where an MQTT data broker acts as the central messaging hub, and an IIoT/Edge platform normalizes and structures the data.

The primary platforms and components used to create this unified layer include:

• Inductive Automation’s Ignition (+ Ignition Edge) : Widely considered the de facto software backbone for building a UNS. Ignition connects to practically any legacy PLC or proprietary machine protocol at the edge, translates it, and publishes it into a structured hierarchy. Learn more about their approach on the Inductive Automation UNS Overview.
• HighByte Intelligence Hub : A specialized Industrial DataOps software designed specifically to model, contextualize, and standardize disparate legacy data (from OPC UA, databases, or older protocols) into standardized semantic payloads before sending them upstream to a broker or cloud.
• MachineMetrics : An out-of-the-box machine data and edge platform focused heavily on discrete manufacturing and CNC/legacy machine retrofitting. It uses plug-and-play edge hardware to instantly standardize metrics like OEE across diverse machine brands. Check out the MachineMetrics Edge Platform for details.
• Enterprise MQTT Brokers (like HiveMQ or EMQX) : These act as the scalable middleware or "central nervous system" that receives data feeds from all individual factories, routing real-time operational events securely between local shop floors and enterprise IT/cloud systems. Explore enterprise implementations via HiveMQ's UNS Guide.

How the Unified Layer Works in Practice

1. Edge Connectors / Gateways : Placed at each legacy machine to pull native signals using older or proprietary protocols (Modbus, DF1, serial, or legacy OPC).
2. Normalization & Contextualization : An edge software (like Ignition Edge or HighByte) maps the raw register data into a semantic naming convention, typically following the ISA-95 standard (Enterprise → Site → Area → Line → Cell).
3. Publish/Subscribe Transport : Data is published via lightweight protocols like MQTT (often utilizing the Sparkplug B specification for auto-discovery of machine tags) to a local broker, which can then be federated up to a corporate/multi-site broker.

If you'd like to narrow this down, let me know:

• What brands or types of legacy controllers/protocols are most common on your floors (e.g., Allen-Bradley DF1, Siemens S7-300, Modbus RTU)?
• Do you prefer an open-toolkit approach (like Ignition) or a turnkey plug-and-play solution?

For a multi-factory environment with different generations of machines, the platform you are looking for is usually called an Industrial Data Platform, IIoT (Industrial Internet of Things) platform, or manufacturing data fabric/unified data layer.

Its role is to sit between machines and enterprise systems:

Legacy machines + PLCs + sensors → connectivity layer → normalized data model → MES/ERP/analytics/AI

Key capabilities to look for:

1. Vendor-agnostic machine connectivity

   • Connects Siemens, Rockwell, Fanuc, older CNCs, proprietary controllers, etc.
   • Supports protocols such as OPC UA, MQTT, Modbus, EtherNet/IP, S7, and others.

2. Data normalization

   • Converts different machine tags and formats into a common schema:
     • Machine state
     • Cycle time
     • Downtime reasons
     • Quality metrics
     • Energy usage
     • Production counts

3. Edge connectivity

   • Uses gateways or edge software so older equipment can contribute data without replacing machines. Many industrial platforms are designed specifically for brownfield/legacy environments.

4. Central governance and analytics

   • Provides one consistent data layer across plants for dashboards, OEE, predictive maintenance, AI, and reporting.

Examples of platforms in this category include:

• Kepware — widely used as an industrial connectivity layer/OPC server approach.
• Ignition — often used for SCADA plus industrial data integration.
• Siemens Insights Hub — Siemens' industrial IoT/data platform.
• PTC ThingWorx — industrial application and connectivity platform.
• Litmus — focused on collecting and contextualizing machine data.
• HighByte — focuses on industrial data modeling and contextualization.

For a large manufacturer with many sites and mixed legacy assets, the architecture is often:

Machine layer
→ PLCs, CNCs, robots, sensors

Edge layer
→ protocol adapters/gateways (OPC UA, MQTT, Modbus)

Unified industrial data layer
→ standardized asset model + historian + metadata

Enterprise layer
→ MES, ERP, cloud data lake, AI platforms

If the main goal is “connect every machine once and standardize data across factories”, prioritize a vendor-neutral industrial data platform rather than a traditional MES, because MES systems usually consume standardized data rather than solve the underlying machine connectivity problem.