The Internet of Things promises insights, progress, and opportunity beyond what previous generations were able to imagine. Connecting physical objects to the Internet can reveal the path to medical breakthroughs, societal improvements, and business opportunities that could fundamentally change people’s lives.
But in our haste to harness that power, we are rushing headlong into this new world without pausing to consider the safest, most direct way forward, and without guaranteeing cooperation among travelers to ensure everyone’s successful arrival at their destination.
Dr. Sanjay Sarma, professor and Director of Digital Learning at MIT and co-founder of MIT’s Auto-ID Center, introduced the Cloud of Things concept in 2012 as an organizing principle for cloud computing in which physical objects are connected to the cloud.
“Just imagine if everything in the real world had an exact replica in the cloud,” Sarma said at the Auto-ID and Sensing Solutions Expo in Cambridge, MA, last year. “Think of it like an avatar. Your car – every time the check engine light comes on, all the sensors in your car are replicated in the cloud. In the cloud now, you have your car, and all its vitals, and its location. So now you can go on your phone and see exactly how your car is doing. If your kids drove your car, you can see where they drove, and if they drove too fast. If your check engine light comes on, you can see when it came on and what the issue is. You can pay your tolls with it. You can pay your insurance with it.
“Once you have big data,” he added, “the sorts of things you can start doing with it are staggering and tremendous.”
Dr. Sarma spoke with Data Informed about the Internet of Things and what he sees as the best way to develop it to its full potential.
Data Informed: What do you see as the current state of the Internet of Things?
Sanjay Sarma: The IoT is undergoing remarkable growth, but the current state of affairs is very “wild west,” with each product offering connecting in its own way. In a sense, this fragmented development is a land grab – an attempt to build an ecosystem of intraoperability, without consideration as to the bigger picture of interoperability. The opportunities afforded by the Internet of Things are incredible, but the current state of development doesn’t work toward the best possible big picture.
Is this state sustainable? What disadvantages does the current approach present?
Sarma: This state is sustainable, but chaotic. Sustainable models and growth models are not one and the same. IoT lacks a dominant architecture. Individual platforms can grow and thrive, but some of the core value of connectivity exists at the boundary layers between services. Until development begins to converge and the various silos unify, the IoT vision will not grow to its fullest potential.
Have these problems dogged other newly emerging technologies in the past? How did a dominant architecture solve those problems?
Sarma: Any sufficiently disruptive technology faces threats. Initially, these threats are competitors opposing the technology, but it doesn’t take long for the threat to transition to competitors with similar, but non-interchangeable, technology. Dominant architectures emerge from technical superiority, value-add, strong partnerships, and, occasionally, just plain luck. With RFID, Walmart was the driving force. With TCP/IP, it was the Department of Defense. The trick with a dominant architecture is to design intelligently for interoperability and growth. What good is an OpenOffice document if Word can’t read the file, or if an image can’t be included?
The problem with Laissez-faire IoT development is that the developers of a particular product or service might not understand every possible use case or interaction. Initially, Nest didn’t have an API, so the applications that developers could write that involved the thermostat were almost non-existent. Now that an API has been made available, developers can detect when a person leaves the office and tell the house to get ready proactively. Services like IFTTT (If This Then That) help to unify various architectures, but are still a long way off from forming a truly cohesive IoT.
What do you see as the best option for a dominant architecture for the IoT?
Sarma: The best architecture choice for the IoT is CloudThings, which describes the creation of digital duplicates of physical objects designed in such a way as to be interactive.
What advantages does a cloud architecture offer the IoT?
Sarma: An extensible mirroring platform, based on a hub-and-spoke model (where each device talks directly to the cloud) allows the most versatile device interactions, though it requires the most thoughtful design.
Architecture design can be unintuitive and contrary to what best business practices would suggest. Taking a step back and actively thinking about the design of the IoT will pay dividends long-term.
Cloud-based design choices provide the best interoperability and data-sharing opportunities. With a common-cloud design, it becomes easier to agree on an architecture and focus our efforts where they are most needed – in protecting security and privacy on the IoT.
What are the risks of proceeding into an IoT world without a dominant architecture?
Sarma: Aside from fragmented product and service ecosystems, security tends to get left behind. With the preponderance of IoT devices that people interact with every day, it’s only a matter of time before someone takes over your light switch – or worse.
Subscribe to Data Informed for the latest information and news on big data and analytics for the enterprise.