How can artificial intelligence help organizations solve digital transformation challenges? Consider these examples - from customer service to resource optimization.
Why IoT is not an island
As IoT increasingly becomes an integral part of many enterprise technology stacks, I see a future for IoT mixing and matching
A few years ago, there was a lot of talk in the Internet of Things space about platforms. The idea was that vendors would create standardized software platforms that could be customized by individual companies for their own use. Think of them as ERP systems but for physical devices rather than transactions.
(I’m talking here about industrial and other commercial IoT applications and installations; the issues in the consumer space are sufficiently different that the two can’t really be discussed together.)
But these platform ambitions haven’t always played out as expected. For example, a recent article in Channelnomics quotes Dima Tokar, co-founder and head of research at IoT analyst MachNation, about GE’s struggles with its Predix platform and digital business generally. He says "This shows that it's difficult to build a software platform that works across many verticals."
[ Read also: How blockchain and the auto industry will fit together. ]
To be sure, the struggles at any single company can’t be extrapolated too far. However, in this case, it does serve to highlight how IoT usually isn’t a discrete hunk of technology that can be dropped into a company’s IT infrastructure. Indeed, the analogy to ERP is telling. Expensive delays and failures associated with ERP projects are an all too common CIO headache. With IoT, multiply that by all the physical sensors and gateway layers that have to be integrated in addition to existing software systems.
Where IT and OT meet
We often talk about IoT as the intersection of Information Technology (IT) and Operational Technology (OT). It’s where power-optimizing or outage-avoiding software meets the grid. It’s where traffic management algorithms meet a network of cameras. It’s where analytics software that can predict upcoming failures meets acoustical sensors and maintenance records.
But we may not have collectively internalized what that level of intersection and integration means.
How standard do you think all those largely physical OT systems are? Oh, industries do develop certain standards to allow them to interoperate with partners and competitors (who are sometimes one and the same). But both the implementations and strategic priorities of individual companies are often quite different. Some may focus on reducing costs. Others focus on the best customer experience. One size doesn’t fit all for either IT or OT.
And that’s just within a single industry. Different industries tend not to even speak the same language. It’s hard to imagine a common platform that serves electric power utilities, telecommunications providers, hospitals, and railroads equally well.
A role for open source
What we can imagine and we do see, however, are common interoperable building blocks (increasingly based on open source) and architectural patterns.
For example, one common IoT pattern is that there’s urgent data that needs to be acted on immediately and other data, pre-filtered or otherwise, that gets sent back to a datacenter for later analysis and non-time critical action (such as predicting when to send out a preventative service tech).
In a positive train control system (PTC), you see both types of actions. Rail-side sensors collect and record data about the train's route, speed, and load characteristics. All data passes through the control tier, where messaging and business rules software determines what to do with the data. As the train nears an intersection, messages to alter speed are relayed to the conductor's dashboard with highest priority (with automated fallbacks as needed). Information about speed, fuel efficiency, weight, and more are stored in the datacenter to be analyzed with less urgency. [ Read a case study about PTC. ]
This example also highlights some of the software building blocks common to many IoT solutions. Business rules management, messaging, reliable intelligent gateways, enterprise service buses, real-time data caching, and scalable storage are among the elements often present in IoT. The manner that they’re used can differ considerably but open source products, in particular, are well-suited to this sort of mixing and matching.
The bottom line is that IoT is increasingly an integral part of many enterprise technology stacks. But, precisely because it is such an integral part, it often needs to be tailored for individual company needs and critical IT and OT systems.
[ Why is adaptability the new power skill? Read our new report from HBR Analytic Services: Transformation Masters: The New Rules of CIO Leadership ]