Industrial IoT Is Different

by Wes Mukai, CTO, GE Transportation Digital Solutions

Déjà vu all over again

We’ve seen it all before, right? Back in the early 2000s, LG introduced the Internet Digital DIOS - the world’s first internet refrigerator. Checking news and your daily calendar from your comfort of your kitchen was going to change the world. RFID tags, active and passive, were going to solve all our problems in the manufacturing and logistics space. The death of bar code readers was inevitable. But at that time, there were some things that gave us pause. Connectivity to the home was adequate, however wired Ethernet and powerline were the real in-home networking options. The rapid flip to pervasive home wireless networks was not obvious. At over $20,000, consumers found the DIOS to be unnecessary and expensive at the time.RFID, with all its promise, gave MBAs all over the world a collective headache as they struggled to solve cost challenges and the Auto ID business case.

Now, Internet of things (IoT) hype is back in full force. The consumer space leads the hype, with interesting internet-connected key locks, toaster ovens, and power strips. Gartner publishes research reports on wearables; and mobile phones in the enterprise stretch the definition of IoT and edge computing. Cloud computing and the distributed reach of cost effective computing, storage and network is driving real consumer and enterprise IoT value.

The challenge of Industrial IoT

Industrial IoT is different. In the transportation space, freight locomotives have over 200 sensors streaming data to control systems, operational applications, and analytic applications. Connectivity loss during a freight trip is still an issue. Proprietary networks and systems persist in trains and power plants alike, making the sharing and integration of data a challenge.

Therefore, one cannot ignore the unique challenges of Industrial IoT:

• Industrial data is growing twice as fast as any other sector. Yet today, less than 3 percent of that data is tagged and used in a meaningful fashion.

• Datasets can be fragmented and siloed such that they can’t be used elsewhere in the business.

• Operational technology (OT) and information technology (IT) systems often operate separately, leading to duplication as the roles these functions play in organizations converge.

• Edge devices are not always connected, may be air-gapped due to privacy/regulatory/security implications, or may need to continue operating when the connection is temporarily unavailable.

• Mission critical analytic or operational applications may need to adapt to local conditions at the edge in real-time, and cannot wait for the data to get to the cloud, be analyzed, and send back insights.

In the transportation industry, railroads are constantly looking for new ways to increase velocity, avoid unplanned downtime, save fuel, and improve safety. At the same time, they look to embrace industrial IoT, the cloud and digital technologies to drive tangible outcomes. The stakes are high. The Railroad industry gains $2.5 billion in value for every 1 mph rise in velocity. They save $2.2 billion for every 1 percent improvement in terminal dwell for railcars. However, to unlock this level of value, the challenges above need to be addressed.

Industrial IoT Considerations – Data, Edge, and Security over the long haull

When embarking on the industrial IoT journey, there are some specific unique considerations that require additional focus over enterprise or consumer IT.

The first consideration is not technical at all, but one of stamina and focus. For the industrial space, it will seem as if many of the innovations and existing technologies of the consumer and enterprise IT space are always 18 months away. We see this in the enterprise cloud space—where companies look to migrate traditional mission critical on-premise applications to the cloud. All of the software-ilitiies never seem good enough. Industrial companies are going to need to work through the IoT and cloud challenges with longer time horizons and with focused, long term investment in their internal IT organizations, partners, and vendors. Technology is advancing rapidly and the companies that stick-to-it, will see the big transformational outcomes.

Data and the edge is the next big consideration. Public IoT clouds are developed to support IT data or con­sumer facing devices. Industrial data is often time-series based existing as multi-terabyte structured streams sampled continuously, resulting in unprecedented volumes of data (In Aviation, one plane flight, for exam­ple, can generate 1 TB of data). To handle this volume and often variety of data, industrial companies need to broaden their definition of the edge. Edge device management and ana­lytics will be separated into multiple tiers, from embedded controls/sen­sors, to gateways, and to on premise appliances. These contrasts with Enter­prise IT edge computing which often begin and ends with compute gate­ways on always connected networks. Data volume and variety, real-time performance, or regulatory/ privacy needs will dic­tate the need for an expanded defini­tion of edge with integrated edge services and intelligence. Without this expansion and integration, data must be centralized which can be impractical in industrial settings.

Security for Industrial IoT is also of utmost concern. Se­curity should be embedded at every level of the industrial cloud. Remem­ber, we are bridging the worlds of IT and OT in a matter that can establish end-to-end security and trust. Pillars of trust around governance and cer­tifications (e.g. ISO 27001/2, NIST 800-53), platform hardening, secure software development environments, and continuous monitoring are criti­cal considerations.

Unlike consumer IT, industrial IoT often involves mission-critical controls or operations applications involving complex, near real-time data analytics. That is what makes the industrial IoT unique and drive the considerations above.

Take the leap

As mentioned earlier, Industrial IoT will always seem 18 months away. But unlike earlier IoT hype cycles, the technology and economic underpinnings of IoT, with the economies of the cloud and advances in computer, network and storage, make this cycle seem different. The time has come for Industrial companies to “take the leap” toward Industrial IoT to unlock the huge potential outcomes that are possible and transform their businesses.