Usually when people talk about the Internet of Things (IoT), the dominant thinking seems to revolve around how to leverage the APIs that are be being baked into new hardware. This includes household appliances, smart watches, smartphones, and HVAC equipment. But developers and businesses with the foresight to think outside the box stand to gain rich rewards by bringing new applications and services to market before these kinds of capabilities have been baked into these devices.
At the Parks Connections Conference in San Francisco, Jim Hunter, chief scientist at Greenwave Systems sat down with TheServerSide.com to discuss the future of the API for the Internet of things. At one level, new applications and services will require the use of programming languages like Java to make sense of the flood of information now available from devices, people, and processes now available. But Hunter observed that making sense of this stream of information will require new thinking about the data models for describing these things that can be easily leveraged by IoT related applications.
Interactions with IoT devices will be a much more natural part of our lives, said Hunter. "IoT object and interaction models will be represented much more like other things in your life such as your friends, your content and your schedules. People don't want to be CTO of their home, they do want to be CEO of their life."
Skipping to the finish line
Jim Hunterchief scientist at Greenwave Systems
One example of such a forward thinking company is Swann, an established enterprise in the closed circuit camera market. Geoffrey Schorz, Training Manager at Swann said that they are getting ready to release a new home gateway and security service that does not require dedicated IoT devices. This integrated home gateway listens to the sounds of fire alarms, carbon monoxide detectors, and breaking glass, to create an API for accessing information about these devices with no protocols in the traditional sense.
This approach also reduces the time, effort and expense that a consumer might otherwise suffer in getting access to the kind of services that satisfy their needs for security. On the backend developers can write applications and algorithms to generate alerts and provide homeowners with information about the status of their security. Schorz said that this approach is a game changer in the home security industry, since all of the other home gateways rely on expensive and complex installation of specialized IoT related hardware. In essence, Swann is taking advantage of new paradigm to bring value to consumers and developers with less cost and hassle.
Listen to the language of objects
Back in the early 1990s, when I was working at the Biosphere II project in Arizona, we were building a 2.5 acre closed system model of the earth. We had implemented a nerve system consisting of over 2,500 sensors to monitor the status of every aspect of this closed environment. One of the lead technicians on the project had observed that he could determine when one of the expensive air handling units was about to fail two weeks before this elaborate nerve system could detect a problem. He had noticed that the subtle sounds of these machines had changed and as a result would prioritize maintenance based on this observation.
A group of researchers at the University of Washington led by Shwetak Patel coined the term infrastructure mediated sensing in 2005 to denote the possibility that we could programmatically understand the workings of devices without dedicated silicon stationed at each monitoring point. This particular technology was a little bit ahead of its time. Patel's startup company, Zensi was ultimately bought up by Belkin. It might presumably be baked into the future home gateway products.
This research leveraged a single sensor on each of the electrical, water, and air handling aspects of a home. The electrical monitoring system listened for the distinct signal generated on the household AC wiring to determine the differences between a light bulb, a stove, a TV, and other devices. This allowed them to determine when any device was being turned on or off and how much power was being used without the installation of dedicated monitoring equipment on each outlet.
The water monitoring system leveraged the physical properties that water created as it was flowing through pipes. Household pipes are constructed such that subtle pressure waves called "water hammer" are generated that differ in volume and vibration between toilets, sinks and showers. This approach allowed Patel's team to see which fixtures were running and how much water they were using without a dedicated sensor on each plumbing fixture.
A single air handling sensor was used to gather information about the subtle pressure variations that occur when people walk between doorways in a house.
Listening to People
On the health monitoring front, most attention has been focused on Smart watches. But applications that leverage these are at the mercy of Apple's and Samsung's pace in making HealthKit and S Health information available. Furthermore, consumers that leverage these types of applications will have to buy these devices, wear them, and perhaps suffer the range anxiety associated with keeping them charged.
New thinking along the same lines as infrastructure mediated sensing could enable developers to create apps outside of these limitations. For example, researchers at MIT have released software for translating camera images of people into data about heart rate and even continuous blood pressure. None of the existing smart watches on the market currently support continuous blood pressure monitoring. Although these kinds of capabilities might be in their early phases. It speaks to the possibility of what can be monitored without dedicated hardware.
Intel is also working on the Context Sensing SDK and the RealSense API that are able to leverage information gathered from smartphones and a new breed of cameras. Today, the sensors make it possible to create 3-D models of a scene. But going forward, Intel is also looking at how to bake information about a person's health, fitness, intent, and emotional state into the APIs as well.
Why Wait for Silicon?
A key component for leveraging information from the IoT lies in creating the right data model for describing the properties of these physical objects said Greenwave's Hunter. This data model makes it possible to create higher-level programming constructs for orchestrating data flows, aggregating information into the cloud, and writing applications that bring value to consumers and businesses.
Large scale hardware manufacturers like Apple and Samsung are certain to make money by selling more intelligent devices. Brett Worthington, VP and General Manager at Wink, a smart home provider, said that consumers who buy an IoT device for their home typically return for another two to three devices within two-weeks after their first purchase. This is great for device manufacturers. At the same time, this business model might be limited to early adopters.
Widespread adoption of IoT applications by the mass-market is likely to be driven more by applications that can take advantage of existing things in the home rather than new purchases and the hassles of installing them. Why wait for these capabilities, and consumers to buy these devices to start creating value today? Innovative companies and developers that create APIs based on the physical properties of things might just jump to the finish line. "The power that transforms data into actionable information is context," said Hunter. "Context aligns a wide variety of relevance including time, density, location, ownership and activities. The greatest promise of IoT is automatically discovering and communicating sensed context and transforming data into the information and actions to move us forward."
What APIs do you find are most useful for working with the IoT? Let us know.