As the Internet of Things inevitable embraces the industrial world, it awakens the promise of direct customer interaction with the manufacture and customization of products. For this to work, however, there is a need for standardization and data sharing across a vast array of automated, real-time devices.
by Andrew Caples, Mentor Graphics Embedded Systems Division
With the vision of Internet of Things (IoT) becoming reality, the impact for industrial automation will be huge as technology standardization increases to enable seamless integration of Cloud-based services. There will be a need to integrate information technology with operations technology down through the fieldbus on the factory floor. The integration of industrial machines with access to Cloud services will introduce new challenges and security concerns. This article discusses Industrial IoT (IIoT) and its implications with respect to connectivity technologies from Cloud to the fieldbus connected devices (Figure 1).
Ethernet for Control Automation Technology (EtherCAT) is a fieldbus solution, critical to enabling the Industrial Internet of Things (IIoT).
We hear about the promise of what IIoT will bring: factories will become more efficient and flexible. These connected factories will offer new services, lower prices for consumers, and faster turn-around times. IIoT will facilitate the direct connection between factories and customers. The era of mass production will be usurped by a new generation of nimble production facilities capable of interacting directly with customers. Using Web-based tools to design and customize products, customers will be able to submit orders with the expectation that production will occur in near real time with of course, overnight shipping.
Connectivity and Integration Complexities
However, the potential of IIoT is hindered with the complexities needed to connect and integrate the smart sensors and embedded computing in industrial machines to the Cloud for real-time data analysis. Factories with networked machines are already in existence today; however, the machines typically operate as islands, or individual modules, without interacting or without awareness of the proceeding or subsequent machine. In order to realize the potential of IIoT, or what is often referred to as “Industry 4.0,” individual machines need to communicate with each other and control or influence other machines in the factory. As product life cycles become shorter, and more custom products are produced with near-commoditized pricing, the need to improve flexibility and efficiency becomes critical to survival. The future for industrial automation includes a production environment in which devices, machines, and materials are connected with sensors, and communication technology is leveraged to increase quality, production efficiency, and flexibility. Realizing these gains will require standardization.
The next quantum leap in efficiency gains for industrial automation will be driven by data collection for factory machines and real-time access to act upon that data. Information is needed to optimize operations to increase velocity and to program devices for custom production runs, predict problems, and prevent downtime. Information must be gathered and used to understand the impact of the environmental conditions on factory machines. As every sensor, actuator, and factory machine becomes a participant in the network, the need for standardization becomes more acute. The vast quantities of data generated by devices spread out all over the factory will require a complete IIoT architecture based on standards to facilitate the secure data transfer to Cloud storage, access to information, and the ability to control factory assets. Smart sensors will use standards-based protocols to connect to gateways that will upload the data using standards-based IIoT protocols to collect, transport, and dump data into Cloud-based storage.
IoT Standards for the Cloud
For resource constrained devices, the IoT protocols that have been historically deployed for data transport and Cloud integration were proprietary and not IP-based. This was primarily due to the large overhead and resource requirements associated with IP packets. In short, proprietary protocols were embraced because IP was not considered practical for low power network nodes like sensors and actuators. IP was bandwidth hungry and memory intensive. With the introduction of IPv6 over Low Power Wireless Personal Area Networks (6LowPAN) the IoT landscape has changed: the use of standards-based IP over low power link technologies is now possible. 6LowPAN is an adaption layer between the IP link and the network layer to enable transmission of IPv6.
packets over low-power wireless connections like 802.15.4. 6LowPAN provides header compression and packet fragmentation to reduce payload size, which allows the low power transmission of standards based IPv6 packets. With layer 2 packet forwarding, 6LowPAN can be used to support large quantities of nodes in low power networks requiring multiple hops over large areas.
Standard Protocols Needed
IoT protocols, such as Message Queue Telemetry Transport (MQTT) and Constrained Application Protocol (CoAP) can provide the standards necessary to integrate sensors and actuators to the Cloud using 6LowPAN for transport (Figure 2).
Figure 2: IoT protocol example.
MQTT provides an excellent choice for IIoT transport for environments in which the monitoring of resource constrained sensors is required. MQTT targets device data collection; and as the name suggests, remote monitoring (telemetry) is its main purpose. Designed to push data, MQTT uses a publish and subscribe message approach (any sensor can publish the data and clients can subscribe). Any new data published is handled by the broker that takes care to ensure all subscribing nodes receive the data. The built-in support for quality assurance guarantees the delivery of a message. MQTT is a binary format that requires minimum bandwidth: the fix header is only 2 bytes. This light-weight protocol, with built-in quality assurance, can be used in unreliable networks. An example might include factory machines that are not connected, but the machines have been around for decades. These machines can become IIoT participants by augmenting them with sensors and using MQTT to publish the data. Other machines in the factory can subscribe to receive the data through the data broker. And as the factory grows, or back-end Cloud applications are added, new subscribers like enterprise resource planning (ERP) applications can become data subscribers. Applications that are designed to monitor thousands of sensors can leverage MQTT as a data collection protocol.
CoAP is a message protocol similar to MQTT, but designed for very low power and efficient data transmission. CoAP provides a request and response interaction model between application endpoints, and supports built-in discovery of services and resources. Because CoAP is a protocol often used in networks with high packet error rate, there are quality assurance features such as support for retransmission. CoAP is designed to easily interface with HTTP for integration with the Web, but provides a very low overhead and simplicity. CoAP is suitable for constrained environments.
Fieldbus technology: EtherCAT
Large industrial automation players have by and large driven fragmentation by promoting different fieldbus technologies. As the push to accelerate factory automation through connectivity gains momentum, there is a trend in the adoption of standards- based fieldbus technologies which leverage the traditional network infrastructure. Ethernet for Control Automation Technology (EtherCAT) is one example of a fieldbus solution that uses existing standard Ethernet infrastructure. EtherCAT is a globally emerging technology that could potentially lead towards a standard for Ethernet fieldbus. EtherCAT is Ethernet-based with real-time support and built-in security because it’s not built on TCP/IP.
Further, because EtherCAT is Ethernet-based at the physical layer, it uses standard Category 5 cabling and Network Interface Cards (NIC). To facilitate TCP/IP-based data transfers within an EtherCAT segment, an Ethernet over EtherCAT (EoE) protocol can be used. Switchports are needed to connect Ethernet devices to an EtherCAT segment. The Ethernet frames are tunneled through the EtherCAT protocol, which makes the EtherCAT network completely transparent for Ethernet devices. In order to prevent any degradation to performance, the switchport takes care of inserting TCP/IP packets into the EtherCAT traffic in a manner to prevent the network’s real-time properties from becoming affected. Additionally, EtherCAT devices may also support Internet protocols (such as HTTP) and can therefore, behave like a standard Ethernet node outside of the EtherCAT segment.
Finally, EoE brings about connectivity and interoperability between TCP/IP devices and EtherCAT, but it can also open the door to potential vulnerabilities for malicious attacks to the I/O network.
Standardizing across the Enterprise
In order to realize the gains promised by Industry 4.0, there is a need for vertical integration throughout the layers of the Automation System Pyramid (Figure 3). Integration is required in order for the data to be accessible throughout the enterprise. In addition, this access to data is needed to make real-time decisions.
Vertical levels of integration throughout the Automation System Pyramid. Source: KRAKEN Automation, Inc.
While EtherCAT is inherently secure, tunneling to other networks within the enterprise using TCP/IP can become a problem as any introduction of malicious software can have devastating effects on the factory floor. Integration with EtherCAT demands protocols that address the security required to prevent compromise to devices connected to the fieldbus. Recently, EtherCAT and the OPC Foundation announced collaboration plans to jointly support Industry 4.0. The OPC-Unified Architecture (OPC-UA) was designed with security in mind in order to be implemented system wide.
OPC-UA includes countermeasures against cyber threats including denial of service attacks, compromised extranet and Cloud components, and malicious software introduced via an Intranet, or the Internet. It is implicitly secure using access control, encryption, digital signatures, and X.509 certificates to address the security requirements needed to allow the secure movement of data throughout the enterprise. Because the OPC-UA is platform independent and scalable, it’s used to integrate devices throughout the enterprise. OPC-UA can be deployed throughout the enterprise on embedded devices executing real-time operating systems (RTOSes) up to services running Linux and Windows platforms.
With the combination of OPC-UA and EtherCAT, standards-based protocols can be used to integrate the factory floor with wider enterprise systems. It is through standards that the realization of data access throughout the enterprise is possible. This can be extended horizontally within the organization, to integrate cross-functional processes to optimize the supply chain, with production and product delivery.
The processes associated with Web-based custom design, ordering, production, and fulfillment require the integration of systems and devices that span the spectrum and reach across multiple, cross-functional groups. As the processes are integrated, devices will need to collaborate and share data. Using standards-based protocols such as EtherCAT, OPC-UA, and other IIoT protocols, true vertical and horizontal integration will be achieved. The impact for industrial automation will be more control, faster response time, and greater accessibility.
The IT/OT convergence
General Electric’s (GE) vision of the future factory has taken vertical and horizontal integration across the enterprise to the next level. By modeling a company’s operations, digital process engineering optimizes the entire operation. Linking product design with manufacturing and the supply chain, predictions can be obtained on the effects of design decisions. Sensors on the factory floor impact the design decisions that can be measured and modeled. The end result is the optimization of the factory, supply chain, fulfillment, and more. Digitally modeling the processes is interesting; however, it can only be realized through standards-based protocols which can integrated on the factory floor, throughout the ERP systems, and finally to the Cloud.
Mentor Graphics is a leader in embedded runtime solutions with a vast embedded solutions runtime portfolio that includes; Nucleus real-time operating system, Mentor Embedded Linux, and Mentor Embedded Hypervisor. Mentor also provides IEC 61508 safety-certified solutions to meet the highest levels of safety for industrial designs. Mentor runtime solutions are integrated with industrial fieldbus support that includes both EtherCAT and EthernetIP, supporting industrial protocols such as OPC-UA, CANbus, and Modbus.
It is becoming abundantly clear with each passing day: momentum is building for IIoT and Industry 4.0. (Figure 4). We are entering an industrial automation revolution of connected devices in which sensors, actuators, and control machines are active participants on the network. The amount of data generated will be vast and accessible throughout the enterprise. Standards-based protocols will be needed for the secure transport of data from the factory floor to the Cloud.
We are at the forefront of the fourth Industrial Revolution. Source: German Research Center for Artificial Intelligence (DFKI), 2013.
Mentor Graphics, Hillsboro, OR. (503) 685-7000. www.mentor.com