Organizations must overcome design and implementation challenges of industrial IoT architecture to establish an effective IIoT deployment.
IIoT technology addresses the connection of sensors and devices with industrial control and energy systems to make improvements in productivity and processes possible. IIoT also integrates advanced computing, data storage infrastructure, analytics and machine learning using strong communications platforms.
IIoT architectural design can be complex, particularly in the communications network and control layer. However, the fast pace of innovation in IIoT and investments that align across several fundamental industry technologies will make IIoT more resilient to change and provide more improvements in the years to come. Operations analytics and performance optimization are particularly effective today when implemented with machine learning-based tools.
What is IIoT architecture and where should organizations start?
IIoT architecture is defined by layers of interaction, including device, communications and semantic. In the device layer, architects design how devices interact with communications systems to connect and interconnect in a structure. In the communications layer, systems use protocols to exchange actionable information. The semantic layer applies meaning and context to the system and identifies system features in the context of business goals.
Architects must start the design of IIoT deployments with an ontology -- the relationship between each of the layers -- and invest in four categories of fundamental industrial IoT technology to accelerate development.
Communications and security
Industrial IoT architecture requires investments in communication within the plant, facility and distribution supply chain to interconnect and optimally interoperate. Organizations have extensive potential uses for IoT sensing and the exchange of data throughout the industrial domain to address management, safety, flow and state of material and machines, environmental conditions and general operations concerns. Communications technology and the promise of 5G with 10 Gbps-capable broadband wireless connectivity increases the capability of IIoT applications. For example, virtual networks available today can securely host computing and analytical workloads in the cloud for broad applications that apply remote monitoring, analytics and decision support through modeling across disparate industrial locations. Local control connectivity between systems in the plant will occur on wired, wireless and optical networks, which are generally private and have a security posture equal to cloud systems.
Open data platforms and industrial device networks
Open systems guide the development of sharing information using a common information model. Industrial device networks share transport and communications protocols in general, which improves their management. In a plant or facility, data governance automation should strictly prohibit information sharing and access to data and devices based on need. Architects must consider the need for data privacy and the development of strong forms of data and communication path protection when developing IIoT platforms.
Intelligent algorithms can complete many classification, prediction and optimization tasks within operations, especially the maintenance and performance optimization of machinery and predictions of system failures based on sensing. Some forms of AI support decision-making and learning from data in real time, whereas other AI models address longer optimization-type problems. Organizations must know which form of AI is needed to contribute to process optimization and to improve expected results months in advance because it is a time-consuming process to develop and embed process optimization into industrial systems.
Systems that interact across a factory, facility or smart building require a broad perspective into the many independent systems that already exist. A successful architectural design depends on the system in terms of abstraction of services, information and data exchanges, and the timing of exchanges in the system as a whole.
Tools to build your organization's IIoT architecture
The Industrial Internet Reference Architecture (IIRA) serves as a guide to develop approaches to complex systems in the IIoT space. Frameworks generally recommend organizations develop an architecture using a system approach for IT and operational technology applications, including specific reference architectures found within an industry segment, such as transportation, energy, healthcare or government.
Early IIoT plans should refer to an implementation viewpoint on patterns of use between components in the IIoT system. The implementation viewpoint is the full lifecycle of technology and its utilization, the collection of infrastructure, communications, sensors, machine learning and the technologies to implement the functional components of the system. For example, organizations can consider a three-tier architecture pattern that consists of the edge, platform and enterprise service tiers aligned with systems currently in use. IIRA also defines other patterns of use. Technical IoT project leaders that commit to reading through IIRA first can shape how organizations define their approach and consider the goals and systems within a framework-driven perspective.
Architects must also take steps to design an IIoT architecture specific to their organization's needs:
- Define the vision and goal of the architecture: Who are the stakeholders? What are you trying to achieve? What will machine learning do better or inform over time? Where are the opportunities for improved decision support? How is the system used, and what determines success? An output-based planning approach is key.
- Have architects and IoT project leaders read and evaluate IIRA to guide their approach. During the design phase, technical stakeholders should align the business viewpoint with the functional viewpoint to define the communications layer, interfaces, data and interactions of systems with the environment.
- Define a scope that considers an implementation future and the viewpoint. Specify functionality, including business, information, operations, control and application domains, and determine a deployment pattern of use that fits best.
- Identify fundamental capabilities of the architecture in terms of tasks, roles and activities within each functional domain, which will coordinate functionality across the system.
- Define feedback, KPIs and success measurement criteria. Define program relationships to ensure all stakeholders and viewpoints are considered.
- Ensure the organizational structure is ready to plan, execute, measure and correct in an ongoing, iterative cycle of continuous improvement.