Everything you need to know about connecting your machines to the cloud

First, you need to understand that I am not a manufacturing professional.

However, I helped found a software manufacturing company, so I wanted to know everything there is to know about manufacturing, manufacturing software, etc.

My goal in trying to learn so much about manufacturing and its related software is to help our current (and future) clients make the best of manufacturing software at their respective plants so they can produce more with less (cost & manual labor).

Thus far, I have written about software in manufacturing, different types used in manufacturing, cloud computing, and many other subjects.

However, before manufacturers can connect their machines to the cloud and start making the most of digitalization, they must understand the options for doing so, the pros and cons of each connection system/protocol, and why.

This is where this article comes in. In it, I will discuss different systems/protocols for connecting machines to the cloud, how each came to be, and some of their pros and cons.

A Brief History of Manufacturing Machines

I didn’t know exactly where to start to learn a brief manufacturing machine history. In those situations, generative AI is usually a good start. So, I asked chatGPT to give me a brief rundown of the history of manufacturing machines and how we came to where we are today. This is what it came up with:

• The first manufacturing machines emerged in the late 1700s during the Industrial Revolution.
• Sensors began to be used in machines during the 1960s-1970s, enabling early automation and control.
• As manufacturing processes became more complex, communication protocols like Modbus, S7, and PROFINET were developed to facilitate machine-to-machine and machine-to-cloud communication.
• The need for cloud integration and IoT in the 2000s drove the creation of protocols like OPC-UA and MQTT, ensuring machines could send critical data to the cloud for real-time monitoring, predictive maintenance, and optimization.

This was an excellent start. However, there is much to unpack, so let’s get started.

1. OPC-UA (Open Platform Communications Unified Architecture)

• Description: OPC-UA is a machine-to-machine communication protocol widely used in industrial automation for data exchange. It allows machines, devices, and systems to share data securely and reliably across different platforms.
• Pros:
  • Platform-independent: Can be used on any operating system and hardware.
  • Interoperability: Standardized interface supports integration between various systems, devices, and sensors from different manufacturers.
  • Security: In-built encryption, authentication, and data integrity features.
  • Scalability: Can be used for small-scale to large-scale industrial systems.
• Cons:
  • Complexity: OPC-UA is more complex to configure and manage than some other protocols.
  • Resource-intensive: Can consume more computing power and network resources due to its extensive feature set.
• Use Cases: Typically used in manufacturing, process automation, and energy sectors where there’s a need for reliable data exchange between systems, machines, and cloud-based platforms.

2. S7 Protocol (Siemens Simatic S7)

• Description: A proprietary communication protocol developed by Siemens for its PLCs (Programmable Logic Controllers). Used for communication between Siemens controllers and connected devices.
• Pros:
  • Tailored for Siemens devices: Optimized for the S7 family of PLCs, allowing high-speed data transfer.
  • Reliability: Proven in industrial environments with a strong focus on reliability and fault tolerance.
  • Ease of Use with Siemens Systems: Seamless integration with Siemens’ ecosystem of automation and control products.
• Cons:
  • Vendor-specific: Not as flexible when it comes to working with non-Siemens devices, reducing interoperability.
  • Limited flexibility: Less suitable for environments requiring integration with multiple vendors’ equipment.
• Use Cases: Widely used in industrial environments where Siemens equipment (e.g., PLCs, HMIs) is prevalent.

3. MQTT (Message Queuing Telemetry Transport)

• Description: A lightweight messaging protocol designed for minimal bandwidth and resource usage, ideal for Internet of Things (IoT) devices and applications.
• Pros:
  • Lightweight: Minimal overhead makes it perfect for IoT and cloud-based applications, especially where bandwidth is limited.
  • Scalability: Can scale from small systems to large networks of connected devices.
  • Publish/Subscribe Model: Efficient use of resources since devices only receive the messages they are subscribed to.
• Cons:
  • Limited QoS: Quality of Service levels are present but may not always be suitable for critical real-time data.
  • Security Concerns: While secure implementations exist, MQTT doesn’t have built-in encryption by default, requiring additional measures.
• Use Cases: Ideal for connecting small IoT devices, sensors, and actuators to the cloud in sectors like agriculture, smart homes, and small industrial applications.

4. Modbus

• Description: A simple, open communication protocol used for industrial device communication, especially in SCADA systems. There are several variants, including Modbus TCP (Ethernet-based) and Modbus RTU (serial communication).
• Pros:
  • Simplicity: Easy to implement and widely supported by a variety of industrial equipment.
  • Interoperability: As an open protocol, it is supported by many devices and manufacturers.
  • Low-cost: No licensing fees or extensive configuration required.
• Cons:
  • Limited data transfer: Not designed for high-speed or complex data structures.
  • No built-in security: Lacks native security features, requiring additional layers to ensure secure communication.
• Use Cases: Common in legacy systems or smaller installations, often in energy monitoring, HVAC, and industrial controls.

5. PROFINET (Process Field Net)

• Description: An Ethernet-based communication protocol designed for real-time industrial automation systems, developed by Siemens. It supports communication between industrial controllers (e.g., PLCs) and devices like I/O blocks, robots, and drives.
• Pros:
  • Real-time communication: Optimized for fast, deterministic communication, essential in real-time control systems.
  • High data throughput: Suitable for applications requiring significant data exchange.
  • Scalability: Can be used for both small-scale systems and large, distributed networks.
• Cons:
  • Complex configuration: Setting up and managing PROFINET networks can be challenging, requiring specialized expertise.
  • Vendor-specific: Primarily aligned with Siemens products, limiting flexibility in multi-vendor environments.
• Use Cases: Ideal for applications needing fast and reliable data transmission, such as robotics, machine tools, and high-speed packaging.

6. EtherNet/IP (Ethernet Industrial Protocol)

• Description: A protocol used for real-time control in industrial automation systems, utilizing standard Ethernet technology.
• Pros:
  • Standardized Ethernet: Uses well-established Ethernet infrastructure, making it easier to integrate with IT networks.
  • Real-time control: Suitable for high-performance industrial applications needing precise timing and synchronization.
  • Scalability: Can support both small-scale and large-scale systems with minimal changes.
• Cons:
  • High bandwidth requirements: Ethernet/IP can demand significant bandwidth, which may be a challenge in certain environments.
  • Complex setup: Can be more difficult to configure than simpler protocols like Modbus or MQTT.
• Use Cases: Common in industrial automation systems, such as automotive assembly lines, where high-speed data transfer and real-time communication are essential.

7. CoAP (Constrained Application Protocol)

• Description: A lightweight web-based protocol designed for constrained devices, similar to MQTT but optimized for machine-to-machine applications.
• Pros:
  • Lightweight: Ideal for devices with limited power and bandwidth.
  • Low overhead: Uses minimal network resources, making it suitable for IoT devices.
  • RESTful design: Follows a REST architecture, making it easy to integrate with web-based systems.
• Cons:
  • Limited support: Not as widely adopted as protocols like MQTT or OPC-UA.
  • Limited features: Lacks the robustness of more established industrial protocols.
• Use Cases: Used in IoT applications where resource constraints (e.g., battery-powered devices) are a concern, like smart sensors or remote monitoring.

Machine Types and Technologies Used for Cloud Connectivity

Different types of machines in industries have adopted various technologies to connect to the cloud:

1. Manufacturing Robots:
   • Use OPC-UA or PROFINET for real-time data exchange between robots and controllers.
   • Cloud integration can be done via MQTT or direct OPC-UA clients to send telemetry data to the cloud.
2. PLC Controllers (e.g., Siemens, Allen-Bradley):
   • S7 Protocol for Siemens devices, EtherNet/IP for Allen-Bradley controllers.
   • OPC-UA is often used for integrating with cloud systems to enable remote monitoring and data analysis.
3. Industrial IoT Sensors:
   • Use MQTT, CoAP, or Modbus for lightweight communication with the cloud, especially when collecting environmental data like temperature, humidity, or vibration.
4. SCADA Systems:
   • Typically use Modbus, OPC-UA, or EtherNet/IP to collect data from distributed systems and relay it to cloud-based platforms for monitoring and control.
5. Smart Machines (e.g., CNC Machines, 3D Printers):
   • OPC-UA, EtherNet/IP, and Modbus are commonly used, with cloud integration via OPC-UA clients or MQTT brokers.

History of main protocols

1. OPC-UA

• History: OPC (OLE for Process Control) was introduced in 1996 as a standard for data exchange between industrial devices, using Microsoft’s OLE/COM technology. By 2008, OPC-UA was developed as a platform-independent, more secure, and scalable successor to the original OPC standard.
• Adoption: OPC-UA has grown to become a widely used standard, with approximately 35-40% of machines in modern industrial settings using it in some form. It is especially common in sectors like manufacturing, energy, and process automation where interoperability across devices and systems is critical.
• Industry Fit: Manufacturing, process industries (like chemicals and oil & gas), and utilities, where multiple types of machines need to exchange data securely and efficiently. It’s particularly favored in multi-vendor environments.

2. S7 Protocol (Siemens Simatic S7)

• History: The S7 protocol was introduced in the early 1990s by Siemens for their Simatic line of programmable logic controllers (PLCs). It was specifically designed for communication between Siemens PLCs and associated devices in industrial automation.
• Adoption: S7 is dominant in environments where Siemens controllers are widely used, particularly in Europe. It is estimated that 20-25% of machines in industrial automation use the S7 protocol. The protocol’s reach is limited outside the Siemens ecosystem, but it remains very strong within Siemens-heavy installations.
• Industry Fit: Heavy manufacturing sectors like automotive, machinery, and discrete manufacturing tend to use S7 extensively, as Siemens PLCs are often the controllers of choice. Industries standardized on Siemens equipment benefit most from S7, due to its high-speed communication and deep integration.

3. MQTT (Message Queuing Telemetry Transport)

• History: MQTT was developed in 1999 by IBM and Arcom (now part of Eurotech) as a lightweight protocol for sending telemetry data from remote sensors in low-bandwidth situations. It was designed to be simple and efficient, making it ideal for early IoT devices.
• Adoption: MQTT has seen massive growth due to the rise of the IoT. It’s estimated that 15-20% of connected machines and IoT devices use MQTT for cloud communication, especially in the context of smart devices and industrial IoT (IIoT) applications.
• Industry Fit: The IoT and IIoT sectors benefit the most from MQTT, particularly in industries like agriculture, transportation, and smart energy management, where low bandwidth and resource constraints are common. It’s often used for lightweight cloud integration in industrial environments where large-scale monitoring is essential.

4. Modbus

• History: Developed by Modicon (now part of Schneider Electric) in 1979, Modbus was one of the first protocols designed for communication between PLCs and other devices. Initially a serial communication protocol (Modbus RTU), it later evolved to include Ethernet-based variants (Modbus TCP).
• Adoption: Modbus remains popular due to its simplicity and is estimated to be used in around 10-15% of machines in industrial settings. It’s especially prevalent in legacy systems and environments where simple communication suffices.
• Industry Fit: Modbus is heavily used in legacy systems, energy management, and small-scale industrial setups like HVAC, water management, and SCADA systems. Industries with limited budgets or less complex automation needs often favor Modbus for its low cost and ease of implementation.

5. PROFINET (Process Field Net)

• History: PROFINET was developed by Siemens in the early 2000s as an Ethernet-based successor to PROFIBUS, to accommodate the need for higher-speed, real-time communication in industrial networks. It was adopted as an industry standard by PI (PROFIBUS & PROFINET International).
• Adoption: PROFINET has become a key protocol in Europe and industries relying on Siemens products, with around 15-20% of industrial systems using it. It has widespread adoption in manufacturing environments requiring real-time data exchange.
• Industry Fit: PROFINET is well-suited for high-speed, real-time automation applications, such as robotics, automotive manufacturing, and high-performance packaging lines. It thrives in environments where speed and real-time control are essential.

6. EtherNet/IP (Ethernet Industrial Protocol)

• History: Developed in the late 1990s by Rockwell Automation and Allen-Bradley, EtherNet/IP extends standard Ethernet for use in real-time industrial automation. It uses TCP/IP and UDP to transfer data and has become a key protocol in North America.
• Adoption: EtherNet/IP is widespread, with 20-25% of machines in North American manufacturing plants using it, particularly in industries using Rockwell Automation or Allen-Bradley controllers.
• Industry Fit: EtherNet/IP is well-suited for industries like automotive, food & beverage, and material handling. It’s favored in environments where Rockwell/Allen-Bradley equipment is the standard, often in discrete manufacturing and process industries.

7. CoAP (Constrained Application Protocol)

• History: CoAP was developed in the early 2010s by the IETF as a specialized web transfer protocol for constrained devices. It is designed to work in resource-limited environments, particularly for IoT applications.
• Adoption: While still relatively niche compared to MQTT, CoAP is used in around 5% of IoT and industrial systems where resource constraints are critical.
• Industry Fit: CoAP is ideal for battery-powered IoT sensors, remote monitoring, and smart building systems. Industries such as smart agriculture, logistics, and utilities benefit from its lightweight and RESTful design.

Adoption Breakdown and Industry Preferences

• Manufacturing: OPC-UA is the go-to protocol for large, complex manufacturing plants needing interoperability across devices and vendors, while PROFINET and EtherNet/IP dominate in Siemens and Rockwell environments.
• Energy and Utilities: Modbus and OPC-UA are popular here due to their simplicity and interoperability across different equipment vendors.
• IoT/IIoT: MQTT and CoAP shine in IoT-heavy industries like agriculture, transportation, and smart cities, offering lightweight communication with the cloud.
• Legacy Systems: Modbus, due to its long history, is still heavily used in legacy systems and industries like water management and basic SCADA systems.

The process of machines sending data (such as temperature, vibration signals, etc.) to the cloud or other monitoring systems typically involves several steps. These steps vary slightly depending on the specific protocol used, but the general flow remains similar:

1. Data Collection (Machine-Level Sensing)

• Machines (e.g., industrial robots, sensors, or PLCs) continuously collect data on key performance metrics such as temperature, vibration, pressure, and other signals through sensors.
• Sensors connected to a machine or PLC (Programmable Logic Controller) measure these parameters in real-time and feed the data to the controller or gateway for further processing.

2. Local Processing (Optional)

• In some cases, the collected data is processed locally by a PLC, edge device, or gateway, which may involve filtering, aggregating, or making real-time decisions based on preset thresholds (e.g., if temperature exceeds a critical limit, send an alert).
• Protocols involved at this stage:
  • S7 Protocol (Siemens PLCs): If Siemens PLCs are used, the S7 protocol is often employed for communication between the machine, controller, and other devices.
  • Modbus: In legacy or simpler systems, Modbus (RTU or TCP) may be used for communication between the sensors and PLCs.
  • PROFINET and EtherNet/IP: These are used for high-speed, real-time communication between machines, robots, or controllers and sensors, especially in industrial environments that require deterministic, time-sensitive data exchanges.

3. Data Transmission to the Cloud or Centralized System

• Once the data is collected and (if necessary) processed locally, the next step is transmitting the data to a central monitoring system, SCADA (Supervisory Control and Data Acquisition) system, or the cloud for further analysis and long-term storage.
• The transmission typically involves pushing the data to a server or broker using one of the communication protocols.
• Protocols for Data Transmission:
  • OPC-UA:
    • Plays a key role when interoperability across different devices is required. The machine (or PLC) encapsulates data such as temperature or vibration into a standard OPC-UA format and transmits it to a remote OPC-UA client or cloud service.
    • OPC-UA allows machines from different manufacturers to share data securely with a cloud-based platform or SCADA system.
  • MQTT:
    • Used for lightweight, low-overhead transmission of data to the cloud. The machine or edge device publishes data (e.g., temperature or vibration readings) to an MQTT broker, which then forwards the data to the cloud or to other subscribed devices.
    • Ideal for IoT and industrial IoT (IIoT) applications where bandwidth is constrained or power limitations exist.
  • Modbus:
    • Though not typically used for direct cloud integration, Modbus can be used in conjunction with a gateway that converts Modbus signals to a cloud-compatible format (e.g., OPC-UA or MQTT). The gateway then sends the data to the cloud.
  • PROFINET and EtherNet/IP:
    • These are typically used for fast, local communication between devices, but they can be part of a hybrid solution where the data is transferred to a cloud-based platform via an edge gateway that supports cloud protocols like OPC-UA or MQTT.

4. Edge Gateway (Optional)

• In many industrial setups, an edge gateway is used to bridge the gap between the machine-level protocols (e.g., S7, PROFINET, Modbus) and cloud-native protocols (e.g., MQTT, OPC-UA).
• The gateway processes data locally, and in some cases, performs actions like filtering or preprocessing before sending it to the cloud.
• Roles in Different Protocols:
  • OPC-UA: Some gateways act as OPC-UA servers or clients, aggregating data from multiple machines and transmitting it securely to a cloud platform.
  • MQTT: Gateways often publish data to MQTT brokers for cloud applications that handle massive amounts of machine data.
  • Modbus to OPC-UA/MQTT Conversion: Gateways often convert legacy Modbus signals into modern OPC-UA or MQTT formats for cloud transmission.

5. Cloud Ingestion & Processing

• Once the data reaches the cloud, it can be further processed, stored, and analyzed. The cloud infrastructure (e.g., AWS IoT, Azure IoT, Google Cloud IoT) typically ingests the incoming data, triggers alerts, or performs real-time analytics on machine health and performance.
• Protocol Role in Cloud Connection:
  • OPC-UA: Cloud systems can directly communicate with the OPC-UA servers or clients at the edge to receive real-time or historical data. The in-built security features ensure that the data is transmitted securely.
  • MQTT: The publish-subscribe model of MQTT is highly suitable for cloud applications where a large number of machines send data at varying intervals. The MQTT broker manages the distribution of data between devices and cloud services.

6. Data Visualization, Analysis, and Storage

• In the cloud, the data is visualized through dashboards, SCADA systems, or dedicated IoT platforms. Advanced analytics tools can be applied to detect trends, identify machine failures, or predict maintenance needs.
• Historical data can also be stored for compliance or future analysis.
• Protocols like OPC-UA are well-suited for integration with SCADA systems, while MQTT is commonly used for cloud-based IoT platforms where scalability and flexibility are essential.

---

Example: Machine Sending Vibration Data Using OPC-UA

1. A vibration sensor on a machine measures vibration levels.
2. The sensor sends this data to a PLC using PROFINET for real-time monitoring.
3. The PLC aggregates this data and forwards it to an OPC-UA server running on an edge gateway.
4. The OPC-UA server converts the data into a standardized format and transmits it to the cloud.
5. In the cloud, an application uses this data to visualize the vibration patterns, send alerts when thresholds are exceeded, and store historical data for analysis.

Example: Machine Sending Temperature Data Using MQTT

1. A temperature sensor measures the temperature of a machine.
2. The data is sent to a local edge device, which processes the data and packages it for transmission.
3. The edge device publishes the data to an MQTT broker.
4. The broker forwards the data to a cloud service subscribed to the temperature topic.
5. In the cloud, the temperature data is stored, visualized, and analyzed for trends or potential issues.

---

How Each Protocol Plays a Role:

• OPC-UA: Acts as the go-to protocol when interoperability, security, and structured data models are needed. It’s ideal for larger-scale industrial environments with multi-vendor systems.
• S7 Protocol: Typically used within Siemens-based infrastructures to send data between machines and controllers. Data is often routed through a gateway for cloud access via other protocols.
• MQTT: Provides an efficient, lightweight way for machines, particularly in IoT or IIoT setups, to send telemetry data like temperature or vibration signals to the cloud.
• Modbus: Often used at the machine level in simpler systems; it requires gateways to communicate with cloud services.
• PROFINET and EtherNet/IP: Provide fast, reliable communication for real-time applications, typically used between machines and controllers before transmitting data to the cloud through a secondary protocol (e.g., OPC-UA or MQTT).