OT Google Cloud IoT Gateway Strategies for Manufacturers: Connecting PLCs and Sensors to Google Cloud

Connecting operational technology to Google Cloud is now a strategic priority for manufacturers who need real-time analytics, AI-driven insights, and centralized visibility across distributed plants. An OT Google Cloud IoT gateway bridges the gap between industrial field devices — PLCs, RTUs, DCSs, and sensors — and cloud services like BigQuery, Pub/Sub, and Vertex AI, without requiring custom software development or compromising plant security. This guide provides a practical architecture roadmap for manufacturing teams and system integrators ready to make that connection work reliably at scale.

Why Manufacturers Are Moving OT Data to Google Cloud

The pressure to extract value from operational data has never been greater. Plant managers at companies like Nestlé, PepsiCo, and Pfizer are being asked to deliver OEE dashboards, predictive maintenance signals, and energy consumption analytics — all fed by real-time data from the factory floor. Google Cloud offers a compelling destination for this data: BigQuery for scalable time-series analytics, Pub/Sub for high-throughput messaging, Looker for visualization, and Vertex AI for machine learning models trained on industrial data.

However, the path from a Siemens S7-1500 PLC or a Rockwell ControlLogix system to a Google Cloud dataset is not straightforward. OT systems speak proprietary and industrial protocols — Modbus TCP, OPC UA, EtherNet/IP, DNP3, IEC 60870-5-104 — that cloud platforms cannot natively consume. The OT Google Cloud IoT gateway layer is what makes this translation possible, handling protocol conversion, data normalization, secure transport, and resilience in a single deployable component.

The Core Architecture: From Field Device to BigQuery

A well-designed OT-to-cloud architecture follows the Purdue Model reference framework, ensuring that data flows are controlled and that OT network integrity is not compromised. In practical deployments, the OT Google Cloud IoT gateway sits at Level 3 or 3.5 of the Purdue Model — the plant operations layer or the Industrial DMZ — acting as the bridge between the OT network and the internet-facing cloud layer.

Purdue Model Deployment Zones

1. Level 1–2 (OT Field Layer): PLCs, RTUs, sensors, and DCS controllers generate raw process data. Protocols at this layer include Modbus RTU/TCP, OPC DA, EtherNet/IP, and Profinet.
2. Level 3 (Site Operations): The gateway node aggregates data from multiple field devices, applies engineering unit conversions, and structures the data for upward delivery.
3. Level 3.5 (Industrial DMZ): The gateway operates in the demilitarized zone, initiating outbound connections toward the cloud while blocking inbound access to the OT network — a reverse connection model aligned with ISA/IEC 62443 zone and conduit principles.
4. Level 4–5 (Enterprise and Cloud): Google Cloud receives structured, normalized data via MQTT or REST API, routing it to Pub/Sub, BigQuery, or Cloud Storage for analytics and AI workloads.

Protocol Strategies for OT Google Cloud IoT Gateway Deployments

Choosing the right protocol stack for your OT Google Cloud IoT gateway determines how reliably and efficiently data reaches Google Cloud. Two protocols dominate modern IIoT cloud integration: MQTT and OPC UA.

MQTT: The Standard for Cloud-Bound IIoT Data

MQTT is a lightweight publish-subscribe protocol designed for constrained networks and high-volume telemetry — exactly the characteristics of industrial sensor streams. Google Cloud’s IoT ecosystem accepts MQTT natively through Pub/Sub bridge configurations, making MQTT the most direct path for sending plant data to Google Cloud. A capable gateway must function as an MQTT client, publishing tag values from PLCs and field instruments at configurable intervals or on change-of-value events.

A critical capability for manufacturing environments is Store and Forward. Network disruptions between a plant in Mexico City and a Google Cloud region are not hypothetical — they happen. Without Store and Forward, any connectivity gap means permanent data loss. An industrial-grade OT Google Cloud IoT gateway must buffer data locally during outages and automatically replay it to the broker once connectivity is restored, guaranteeing zero data loss.

OPC UA: The Industrial Standard for Structured Data

On the OT side, OPC UA (OPC Unified Architecture) has become the preferred interoperability standard for PLCs and DCS systems from manufacturers like Siemens, Schneider Electric, and ABB. A gateway that operates simultaneously as an OPC UA client (reading from controllers) and an OPC UA server (exposing data to SCADA or upstream systems) provides maximum flexibility in hybrid architectures where some consumers remain on-premises while others are cloud-based.

Supporting Protocols for Heterogeneous Plant Floors

Real manufacturing plants are never homogeneous. A pharmaceutical facility might have Endress+Hauser instruments on Modbus TCP, a Rockwell Automation PLC on EtherNet/IP, a legacy DCS on OPC DA, and a building management system on BACnet — all in the same site. An effective OT Google Cloud IoT gateway must support multi-protocol acquisition simultaneously, normalizing all sources into a unified data model before cloud delivery.

1. Modbus TCP/RTU — legacy sensors, drives, energy meters
2. EtherNet/IP — Rockwell Automation ControlLogix and CompactLogix PLCs
3. Siemens S7 (300/400/1200/1500) — native Siemens controller communication
4. DNP3 and IEC 60870-5-104 — substations, SCADA telecontrol, utilities
5. IEC 61850 — power protection relays and smart grid equipment
6. REST API and SQL/ODBC — integration with MES, ERP, and historian systems

Key Requirements for Industrial-Grade OT Google Cloud IoT Gateway Solutions

Not all gateways marketed as IIoT-ready are suitable for demanding manufacturing environments. System integrators and IT/OT engineers evaluating an OT Google Cloud IoT gateway should validate the following capabilities before committing to a platform.

Reliability and Data Integrity

1. Store and Forward: Local buffering with automatic replay — essential for plants with unreliable WAN connections.
2. Redundancy: Hot-standby failover between a primary node and a backup node, ensuring continuous data delivery even during hardware or software failure.
3. Timestamp preservation: Data arriving at BigQuery must carry accurate OT-side timestamps, not cloud ingestion timestamps, for meaningful time-series analysis.

Security Architecture

1. Reverse connection model: The gateway initiates outbound MQTT connections toward Google Cloud, meaning no inbound ports need to be opened into the OT network.
2. TLS encryption: All MQTT sessions to cloud brokers must be encrypted in transit, meeting the baseline requirements of frameworks like NIST CSF and NIS2.
3. Role-Based Access Control (RBAC): Granular user permissions prevent unauthorized configuration changes at the gateway level.
4. Data diode compatibility: For critical infrastructure or highly regulated environments (energy, pharma, oil and gas), the gateway should support one-way data flow architectures.

Scalability Without Licensing Penalties

One of the most painful limitations in commercial gateway software is tag-based licensing. When a mining operation at Minera México or a refinery at CEPSA wants to add 10,000 new sensor tags to their cloud pipeline, per-tag licensing costs can make the project economically unfeasible. An OT Google Cloud IoT gateway platform with unlimited tag licensing removes this barrier entirely, allowing deployments to scale with operational needs rather than budget ceilings.

No-Code Configuration

Automation engineers and system integrators should be able to configure data sources, MQTT topics, tag mappings, and cloud destinations through a web-based interface — without writing Python scripts or custom middleware. No-code configuration drastically reduces project delivery timelines and eliminates a major source of integration errors.

Real-World Deployment Example: Renewable Energy to Google Cloud

Consider a wind farm operator managing a fleet of turbines across multiple sites — a scenario similar to the work done connecting the Taiba N’Diaye Wind Power Station in Senegal to a remote control center. In that type of deployment, each turbine generates data through IEC 60870-5-104 RTUs. The OT Google Cloud IoT gateway at each site acquires turbine telemetry, normalizes it into a structured tag model, and publishes it via MQTT with TLS to a cloud broker. BigQuery receives the data for performance analytics, while Vertex AI models identify anomalies in turbine vibration and power output curves. Store and Forward ensures that data from sites with intermittent satellite connectivity is never lost — a non-negotiable requirement when regulatory reporting depends on complete historical records.

Connecting to BigQuery: Architecture Patterns

Google Cloud offers several patterns for ingesting MQTT data from an OT Google Cloud IoT gateway into BigQuery for analytics.

1. MQTT → Pub/Sub → Dataflow → BigQuery: The most scalable pattern for high-frequency telemetry. The gateway publishes to a Pub/Sub topic, a Dataflow pipeline transforms and routes messages, and BigQuery stores the structured records.
2. MQTT → Cloud Functions → BigQuery Streaming API: A simpler pattern for moderate-volume deployments. Cloud Functions trigger on incoming Pub/Sub messages and insert rows directly into BigQuery via the streaming API.
3. REST API → BigQuery Storage Write API: When the gateway exposes a REST API server, Google Cloud workflows can poll for data and write batches to BigQuery — useful for legacy environments where MQTT is not preferred.

Regardless of the pattern chosen, the gateway’s ability to structure and label data correctly before it leaves the OT network is what determines the quality of analytics in BigQuery. Raw Modbus register values are meaningless without engineering units, tag names, asset context, and quality codes — all of which a capable OT Google Cloud IoT gateway applies at the source.

How vNode Solves This

The vNode Industrial Data Platform is purpose-built to serve as a production-grade OT Google Cloud IoT gateway for manufacturers, system integrators, and industrial end customers across more than 40 countries. With 10,000+ installations across oil and gas, pharma, renewable energy, mining, and utilities, vNode addresses every challenge described in this article through a unified, no-code platform.

1. Multi-protocol OT acquisition: vNode simultaneously acquires data from Siemens S7 (300/400/1200/1500), Rockwell EtherNet/IP, Schneider Electric OPC UA servers, Endress+Hauser Modbus instruments, ABB DCS systems, DNP3 RTUs, IEC 60870-5-104 substations, and many more — all configured through a web-based interface with no programming required.
2. MQTT Client with Store and Forward: vNode’s MQTT module publishes structured tag data to any MQTT broker, including Google Cloud Pub/Sub bridge endpoints, with built-in Store and Forward that buffers data locally during network outages and replays it automatically — guaranteeing zero data loss.
3. OPC UA Client and Server simultaneously: vNode reads from OPC UA-enabled PLCs and DCS systems while simultaneously exposing a unified OPC UA server to on-premises SCADA and historian systems — enabling hybrid cloud and on-premises architectures without duplication.
4. Unlimited tags, no licensing penalties: vNode does not charge per tag, making it economically viable to connect thousands of sensors from large-scale industrial plants to Google Cloud without cost escalation.
5. Sparkplug B support: For deployments requiring standardized IIoT payload structures, vNode’s Sparkplug B module delivers self-describing, structured MQTT messages that integrate cleanly with Google Cloud analytics pipelines.
6. Industrial DMZ and cybersecurity-ready architecture: vNode supports reverse connection (gateway-initiated outbound MQTT), data diode-compatible deployments, RBAC, and network segmentation — aligned with ISA/IEC 62443 zones and conduits and NIST CSF Protect functions.
7. REST API Server and SQL/ODBC: Beyond MQTT, vNode can expose industrial data via REST API or write directly to MySQL, PostgreSQL, SQL Server, and Oracle databases — providing additional integration paths into Google Cloud via Cloud SQL or AlloyDB.
8. Built-in redundancy: vNode’s hot-standby redundancy module ensures automatic failover between primary and backup nodes, maintaining continuous cloud data delivery even during node-level failures.
9. MCP Server for AI integration: vNode’s MCP Server module makes structured industrial data available to LLM-based tools and industrial copilots, enabling AI workloads in Google Vertex AI to consume live OT context directly.

Whether you are a system integrator delivering a cloud connectivity project for a pharmaceutical plant, or an operations team at a mining company looking to feed Google BigQuery with real-time conveyor and crusher data, vNode provides the architecture, the protocols, and the reliability to make it work. Contact the vNode team to discuss your specific deployment scenario, or explore the full technical capabilities in the vNode User Manual.

Frequently Asked Questions

Can vNode send PLC data directly to Google Cloud Pub/Sub using MQTT?

Yes. vNode’s MQTT Client module publishes tag data from any connected PLC or field device to any MQTT broker, including configurations that route to Google Cloud Pub/Sub. The connection is outbound from the vNode gateway, requiring no inbound firewall openings into the OT network, and Store and Forward ensures no data is lost during connectivity interruptions.

What makes an OT Google Cloud IoT gateway different from a standard IoT gateway?

An OT Google Cloud IoT gateway must support industrial protocols like Modbus, OPC UA, EtherNet/IP, and DNP3, which standard IoT gateways typically do not handle natively. It must also meet OT reliability standards — including Store and Forward, redundancy, and cybersecurity-oriented architectures like reverse connections and DMZ deployment — that consumer or IT-grade IoT gateways are not designed for.

Does vNode support Sparkplug B for Google Cloud IIoT integration?

Yes. vNode includes a Sparkplug B module that wraps MQTT messages in the standardized Sparkplug B payload format, enabling self-describing, structured data delivery to MQTT brokers and downstream Google Cloud analytics pipelines. This is particularly valuable for large-scale deployments where data schema consistency across many devices is critical.

How does vNode handle data loss when the internet connection to Google Cloud is interrupted?

vNode’s Store and Forward capability buffers all tag data locally on the gateway node during any network disruption — whether lasting seconds or hours. Once connectivity to the MQTT broker or cloud endpoint is restored, vNode automatically replays the buffered data in sequence, ensuring BigQuery and other cloud analytics systems receive a complete, uninterrupted time-series record.