Everyone is well aware of the hype around the Internet of Things in the consumer applications arena. But some of the most significant value that connected, intelligent machines will make possible comes from the industrial space. No longer is the Industrial Internet of Things (IIoT) just hype, either. With today’s highly developed sensors and big data analytics capabilities, manufacturers, oil and gas companies, and many others are optimizing their operations and transforming the traditional industrial process.
However, before these enhancements can be realized, operators need to feel confident that their systems can handle the always-on requirements of mission-critical applications. The implications of “on” switching to “off” for even a brief interval can lead to catastrophic business disasters. I believe there are three key issues that business and technology leaders should consider as they navigate the transition to the IIoT-enabled manufacturing enterprise.
1. Understand the potential vulnerability of virtualization.
Virtualization has transformed the way in which enterprises deploy technology, delivering significant efficiency boosts and cost savings. By enabling multiple applications to run on the same platform, virtualization delivers dramatic reductions in hardware costs, energy consumption, rack space and maintenance.
For better or worse, virtualization concentrates the potential points of failure by consolidating multiple processing workloads onto a single machine. In effect, it puts all your processing eggs into a single basket. If a virtualized server goes down, it could take several applications down with it.
Enterprises taking advantage of virtualization must think very carefully about how to maintain the availability of their virtualized applications. Deploying virtual platforms designed specifically for the rigors of industrial workloads at the edge is a crucial—and often ignored—success factor for highly automated, IIoT environments. This is because IIoT platforms are primarily deployed in environments that require:
• Always-on capability: Downtime can have severe business, financial and regulatory compliance implications. And because many of these platforms are deployed in remote locations, where an outage can be exceptionally costly, virtual platforms must be kept running until a repair is made.
• Little or no IT expertise: Most operational technologists are not trained to deal with multiple servers, switches, storage arrays and associated cabling.
2. Realize that operational technology (OT) and information technology (IT) function differently.
Say “enterprise technology” and the following image probably leaps to mind: a gleaming, climate-controlled data center filled with racks of blinking servers, switches and storage platforms, with a team of skilled technologists constantly at the ready to attend to the systems in their charge. That’s the world of IT.
Things are different in the world of OT. Consider a company operating thousands of miles of gas pipelines. It relies on a network of compression stations along the pipeline, which are used to maintain volumetric pressure and keep gas flowing. These remote stations require multiple applications to operate properly. Dedicating a server to each application is costly and complex, and takes up valuable space. OT platforms don’t have teams of technologists on hand to address problems or outages. If a server goes down, a new one must be configured back at headquarters, transported out to the remote compression station and installed. The application on that server could be down for two or three days, potentially disrupting operations. Even in environments that use virtualization techniques to provide a higher degree of availability, there are risks to the multiple pieces of equipment and cabling required when such a solution is deployed outside the controlled, pristine environment of a data center.
A better approach would be to equip the compression stations with a single, simple, fault-tolerant, virtualized server—designed specifically for industrial applications—that delivers continuous service while consuming less space. Using virtualization also simplifies the process of pushing new applications out to these remote stations, avoiding the slow, costly process of sending technicians to the physical location.
Deploying simple, compact, flexible platforms that can be easily maintained by OT personnel—or remotely by IT staff—is a critical success factor for achieving availability in an IIoT deployment.
3. Identify your downtime threshold and act accordingly.
For some business applications, such as e-mail, an outage of an hour or two is annoying but likely not catastrophic for the business. In the manufacturing plant, unplanned downtime of an application controlling a packaging line in a pharmaceutical plant or silicon chip production line can create measurable effects on that quarter’s revenue. Interruptions in systems that impact worker or public safety could have even more serious consequences.
By definition, the IIoT increases the scale and remoteness of deployments exponentially. Delegating the technical implementation of an industrial automation or IIoT project to those focused only on costs or budgets, without understanding the full business implications of potential weaknesses in the solution, is asking for trouble. It is critical to evaluate the potential business impact of a system failure and invest in an IIoT solution with the robustness required to meet and eliminate that threat. This factor simply cannot be overlooked.
John Fryer is the senior director of product marketing at Stratus Technologies, where he is responsible for go-to-market strategies and industry initiatives across all of the company’s product lines. He has more than 25 years of experience with systems and software products in a variety of engineering, marketing and executive roles at successful startups and major companies, including Motorola, Emerson Network Power and Oracle.