I’m an IT guy. Have been in software for 25 years. I have two Co-founders at Atomiton. Jane is a physician by training. Oleg is a nuclear physicist by training. Both of them keep me grounded – in the physical reality.
Oleg tells me, taking something abstract and behaving as if it was true is called “belief”. I call it “software engineering”. The world of IT is all about information and abstraction, through which we software engineers think we control the world – like Marc Andreessen said, we are "eating the world".
But back when I was CTO at GE Software, architecting the Industrial Internet, I learned humbling lessons from my GE counterparts – the OT guys. When the OT guys from the Energy, Oil & Gas, Transportation businesses told me “over my dead body” about some of my ideas, I knew I was not going to fight with the laws of physics.
Joking aside, what exactly is OT, or Operational Technology? Gartner defines it as “hardware and software that detects or causes a change through the direct monitoring and/or control of physical devices, processes and events in the enterprise”. Today OT is widely used in industrial processes such as manufacturing, water treatment, energy distribution, oil pipelines, and in facilities management.
In contrast, IT is “the application of computers and telecommunications equipment to store, retrieve, transmit and manipulate data”, or information.
IoT = IT + OT
If you have not derived this already, think about Smart Home as an example. An IT-centric “Smart Home” solution would collect data on every aspect about your home environment, your living habits, your sleep cycles, via any possible means, including sensors and smart fridges. You would get a report everyday (or every minute if you like); you may receive suggestions and reminders, based on “Big Data” analytics, on how to live better, when you should buy milk, or whether your cat is home. But nothing much about your life would really change.
An OT-style “Smart Home” would feature a “dinner meal process”, where your fridge selects and presents the ingredients, your cutting board measures and chops them up, the kitchen mixer works out the recipe, and your stove and oven work in tandem to produce the edible meal. If anything gets burned or spilled during the process, an alarm would alert you, and an automated process would be sent down to make sure no fire or slippery floor happens. You, on the other hand, would relax, watch what your TV selects for you, and only intervene if there’s any unusual incidence. The only problem is, you would get the same dinner every day.
Obviously, a sensible IoT Smart Home solution would take the merits of both IT and OT and create an entirely new experience. But don’t dwell on the Smart Home thought – it is only meant to be an analogy. Think about cities, schools, factories, wind farms and retail stores. Think about a wind farm that can not only collect and analyze data to determine weather patterns, but also adjust the torque of selected wind turbines from second to second, in response to changing weather conditions. You would need the information from IT and the physical action from OT, and something in the middle to link them logically.
The idea that IT and OT need to be merging is not new:
- In 2011, Gartner said the convergence of IT with “physical-equipment-oriented technology” – what it refers to as “OT”, would optimize business processes, enhance better decisions, and reduce costs.
- GE’s “Industrial Internet” succinctly states the intention to bring together the “Industrial” (OT) and the Internet (IT), or "bringing software and machines together".
- If you want to learn more, this report from Schneider Electric elegantly explains how IT and OT convergence enables smart grid development.
In a sense, the IoT hype today reflects the accelerated development of this trend, which is greatly enhanced by
- The cheap and ubiquitous accessibility of mainstream IT technologies (networking, cloud computing, storage);
- The maturation of newer IT technologies (communications, containerization, Big Data);
- The dramatic drop in the cost of sensors.
What happened to the “O”?
Over the last two years, in pushing IoT to the peak of the hype, we have kept, and blown up, the sensor, data, and cloud – the IT mindset, but dropped the control, process, and coordination – the OT mindset. In the technology innovation landscape today, pockets of OT-minded activities happen in consumer electronics and homes, or among hobbyists, but far less in business and industries than it is deserved. Here are the signs & symptoms you have seen:
- Every business wants Hadoop, while most do not understand;
- Getting data out of machines becomes the top priority. Enterprise partners frequently fight about data ownership;
- Most IoT platform stacks resemble a data stack;
- Cloud becomes the Holy Grail;
- Applications are mostly dashboard-like.
It seems that the IT guys have won the upper hand in this IoT gold rush.
So what would happen if we let go of the “O”, and become “IT-on-steroids” hooked up with billions of sensors and smart phones? Three consequences:
- ROI will be delayed. In industries, failure to take actions on the physical world within seconds’ or minutes’ time window will cost us majority of the near term productivity gains. Big data insights can bring us the intelligence to change processes or designs, which are mid-term and long-term gains.
- People’s experience will get worse. Deficient on physical actions, IT will always put people in the loop, creating ever more work for them. The Schneider Electric paper describes how operators could struggle to process the higher volumes of information while trying to select operating options from an increased set of alternatives.
- Data will be become unmanageable. IoT is not about how fast we acquire data; it is about how fast we can “get rid of” data. Data on which relevant actions have not been taken have to be stored, like most enterprises are doing today (hence Hadoop). In an ideal IoT world, 80% of the data require local and in-line actions, after which only metadata is kept.
IoT needs new “OT”
There are reasons why IT and OT have been in separate worlds, and why the efforts to bring them together have mostly been integration, not technological convergence. Traditionally OT has been designed for mission critical tasks where:
- Time and sequence matter;
- Response time required is often milliseconds;
- Must be in sync with physical reality;
- Deterministic control is needed
Most of such concepts are foreign to IT software development.
But on the other hand, OT is rigid and inflexible. At GE, I got a shock when I first learned that bringing a new feature to a plant equipment could easily take 2 years, including design, development, validation, deployment, instrumentation, validation (again), sign-off, operation. (i.e. you would be eating the same dinner for 2 years)
There is no doubt IoT needs new technologies to bring the capabilities of IT and OT together, and take them to many more sectors than just the OT-dominated industries. What are the solutions is the challenge question I will give to the readers and to the ecosystem currently innovating in IoT. At Atomiton we are creating our own solution to address this need: TQL System. But I believe the answers could be many. I’m laying out three categories of cases:
- First, the traditional OT-run sectors, where PLCs and SCADAs are the norm, for example manufacturing, industrial plants, energy, etc. We need to bring the intelligence and flexibility of IT to the OT.
- Second, the machine-rich operational enterprises, such as hospitals, labs, buildings and offices, where smarter machines operate in silos. We need to create an OT-like layer to coordinate the machines, while relaxing some of the control system requirements from the industrial domains.
- Third, the newly connected worlds such as in homes, cities, logistics. We need to bring OT characteristics to the IT, but with much more contextual responsiveness than traditional OT.
As my physician partner Jane often states, “When your head is in the Cloud, your feet still need to be on the ground.” That is what IoT needs.