What’s a Real IoT Project? A career-shaping question for technologists

— by Alok Batra

In 2010, I was a CTO at Cisco when it started to put increasing strategic bet on the Internet of Things (IoT). That year, we designed Smart City in Songdo. Soon after, I found myself at GE, architecting the “Industrial Internet”. When I started engineering on GE’s Predix platform in early 2012, there was still skepticism in the market towards the value of IoT. By the time Predix was revealed to the market in mid 2013, every other startup has become an IoT company and every other technology enterprise has hired an IoT head. In that year, there was customer who came to us and said, “I heard that you created the ‘Industrial Internet’. That is indeed what we need! Can we order 2?”

In a time when IoT projects seem to have mushroomed in large quantities overnight, and hundreds of thousands of people’s careers are suddenly tied to it, how to make sure that you launch (or join) the real IoT projects that will make careers, not the “bridge-to-no-where” IoT projects that will break careers?

Five things set real IoT projects different from just another conventional project under the guise of an “IoT” name.

Not Data, but Things

The Internet of Things is not the Internet of Data (IoD). Just because the Big Data hype and the IoT hype coincide in time, it doesn’t mean that collecting, monitoring or analyzing data from everything connected is IoT. A same problem can be defined as a “data problem” or a “thing problem”. The problem of trains running inefficiently is defined as a data problem if the solution is to collect data from sensors on the trains and analyze what affects fuel efficiency (with the hope that the learning will change the train operator’s behavior in the future). But GE defined it as a “thing problem”, and it came up with Trip Optimizer. Trip Optimizer uses a number of sensor inputs and processors to automatically control the locomotive’s throttle, helping keep trains on schedule while minimizing fuel use.

To distinguish an IoT project from an "IoD" project, ask the question: What is the expected output of the technology used? To collect and analyze data or to impact (system of) Things? To give people good information and ask them to take action, or to make things around people take actions for them?

Not rebranding, but new creation

Having had a career as CTO and chief architect at tech companies from large to small, I know what is on the minds of some technologists when something as hot as the “Internet of Things” comes into the market. “This will be another opportunity to push my platform (s) (or products)” or “this will be a way to revive my orphan idea that those non-innovators rejected”. The market is currently confused with myriads of IoT projects propped up by rebranded solutions and repackaged old technologies. Creating IoT in healthcare is not easy, and yet overnight, home monitoring solutions, which have been on the market for decades, suddenly become healthcare IoT success stories.

There is no evil in rebranding, the trap is, re-inventing from yesterday’s technology will delay you from getting to the future. This is the first time in history when physical things start to be merged with the virtual world of software in large scale. Last time when just one type of such physical things – routers and switches – came into the virtual world, the networking industry was created. Think about creation first, not just reuse.

Not making assets productive, but turning them into services

The Internet of Things eventually is not about Things. It is about people. Productive assets is about Things, new services is about people. Making assets 1% more efficient is linear progress. Creating new services can be exponential value.When an asset is turned into services, its utility is provided on-demand; it changes behavior with contexts; and it automates tasks instead of creating tasks for people. Nest did not create the most efficient thermostat in the world, it created a thermostat that becomes the temperature control service for people.

If you have a project to connect assets (refrigerators, cars, industrial machines…) and monitor their status so they can be running consistently and be serviced more promptly, then your project is not a hardcore IoT project. If you are allowing a parking meter to remind its user that metering is about to expire, then you are starting to think IoT. The next step is to provide a service to set up auto refill on parking meters.

Not just informative, but responsive

IoT is not a big dashboard of all things connectable. When the Internet itself first became popular, people primarily used it to publish and share information (browsers and emails). But today, the greatest value of the Internet is in the interactions it created – the number of clicks that created purchases, the personalized ads delivered, the games played on it, the money wired through it...

So being “informative” is a lower bar for IoT. Projects in Smart Cities have all thought about “City Dashboards”, but few are like Cisco, which is making city infrastructure, such as lights, responsive to people and its surroundings, in order to save energy and ensure safety. A “responsive” system of IoT is a platform for exponential value and monetization; an “informative” one at best offers linear monetization. More importantly, the software methodology to build a responsive IoT system is very different from that for an informative system.

Not just connected Things, but interconnected Things

For a long time we thought IoT was about connectivity. Just like for a long time we thought the Internet movement was all about laying down the cables and getting people online. It still is, in many parts of the world, but we discovered explosive growth comes in social networking – interconnected people. If you are working on getting devices connected to the cloud, or pushing device data into the cloud, you are working on connected things, not the Internet of Things. The wind farms where each turbine adjusts to both local condition and global outputs in order to balance output against wear & tear is a system of interconnected things. This type of project is a real IoT project. The challenge of connecting things is in the communications, the challenge of interconnecting things is in the architecture and workflow. In my future posts, I will elaborate on what is missing in the current “Connectivity + Cloud” IoT architecture and how to create interconnected IoT systems. Follow us on Twitter to stay tuned, and welcome to check out other posts on the Atomiton blog.