The possibility of a true physical graph

As I prepare Gigaom Research’s upcoming technology roadmap for the smart home, I am increasingly considering what it will take to get us to a place where a true physical graph exists, a place where the physical world is connected, programmable and capable of communicating amongst itself.
We are, to be sure, a long ways away from this possibility. As broadband becomes ubiquitous and chipset costs continue their decline, we lay at a juncture where it’s very easy to connect devices, be it a dishwasher or a thermostat.
That connection has taken on a simple point product, singular application type model where one device is connected to the cloud and can be controlled via a desktop or mobile device. Typically this involves device specific apps, which is an intrinsic limiter of the ability of devices to communicate with one another. Imagine having an app for lighting, one for a thermostat, one for a stereo. You get the picture.
Now this model will work for a while because there still is value in being able to program and control certain home devices. The Nest learning thermostat is the most obvious example due to the energy savings, but I imagine there will be other applications like security (I could see parents wanting to be able to make sure their doors are locked even when they’re not home. Or imagine a backyard fence surrounding a pool and wanting to know whenever it’s open). Needless to say, many use case arguments will be made for application specific connected devices in the home.
But that’s an entirely different ball game to being able to:
1) Enable rules to easily control multiple devices.
2) Have a learning ecosystem where rules are suggested based on my own behavior that impact multiple devices.
3) Context awareness: Where devices themselves understand what other devices exist in a smart home and can communicate with them.
4) Have device to device communication in which devices can leverage hardware resources from other devices. For example, a connected smoke alarm that was aware of a connected media system that could leverage the speakers in the event of an alarm. Or if someone were deaf, the connected smoke alarm could use the connected lights as the alarm. The possibilities begin to become endless if you combine a true physical graph that is open, secure and where developers are able to unleash their creativity by being able to access the hardware resources of the physical graph.
Now we’re a long ways away from step 4. For starters, we’re grappling with multiple chipsets and radio protocols from Zigbee to Bluetooth to Wifi that must be corralled into being easily accessible from one controlling device like a smart phone. The hubs hitting the market, including Revolv, SmartThings and Staples Connect are all trying to solve this problem.
Some groups, like the AllSeen Alliance which employs the AllJoyn open source project and runs atop hardware protocols, wants to solve this problem. AllSeen has accumulated about 50 partners since its founding this year. The goal is to expand the open source so that functionality can be shared easily among devices.
Creating a framework that would allow the vast majority of devices to easily characterize themselves and pair with one another is a major ask, and antithetical to the way in which many companies like Apple and Samsung have historically thought about how to develop and monetize hardware.
But the past is always limited in what it can tell us about future technology value. And in a world where the true value and potential of the physical graph is in interoperability, there will be incentives for device makers, startups, semiconductor makers and platform providers to understand that realizing full value for the physical graph over the long term will require cooperation.
Because the alternative is mass fragmentation which will become unworkable.  And in that scenario the value of every device maker’s hardware will be lower.