Thursday, April 29, 2010

Socially Connected Devices

There's no shortage of people shouting about 50 billion devices being connected by 2020 but frankly I find all these futuristic discussions a bit shallow and without any real insight. People talk about everything being connected to everything and how your refrigerator will be able to order more milk when you run out (or something like that). But their arguments all seem to be built on an evolutionary model of connectivity expansion rather than a revolutionary one; there doesn't seem to be state change in how devices will communicate, just a whole lot more of them. So let me share with you where I think it is all going.

1st Generation Internet: All about getting big ugly machines to be able to transfer data between one another where the nature of the data was only meaningful to the computers at each end of the communication pipe.

2nd Generation Internet: People start getting involved and applications such as email allow people to start communicating with other people across the internet. The internet begins its amazing growth phase and websites designed for human interaction emerge.

3rd Generation Internet: Social Networking. Broadcast and community-based forms of communication emerge. Instead of using point to point emails to talk to the world, people broadcast their status using sites such as YouTube, Facebook, Twitter and Blogs.

OK, so hopefully you are still with me because here's where it gets interesting. We are currently in the 3rd Generation phase but the most interesting part is where its going from here. If we think that 50 billion devices are all going to start talking to one another, then what will they be saying? What will be the style of their communication?

4th Generation Internet: Socially Connected Devices.
By combining concepts related to the previous generations, 4th generation internet devices will communicate with one another in much the same way as humans interact on social networks; a device will broadcast its status / opinion / desire for information, etc to the 'cloud' and let other devices on that cloud respond in a manner similar to human social networks. Devices would comment on the status or opinion expressed by the original device; I agree, I disagree, I have a similar idea but my information is based on different inputs, etc. and they would adjust their own operation (opinion / status) based on the information they receive. They could supply information according to the original broadcaster's request, or they could attempt to open up a point to point communications channel with the original broadcaster in order to resolve, clarify or expand on the original topic of discussion.The point here is that (unlike the 1st & 2nd generation internet) communication is rarely point to point. It is far more communal in nature and provides all members of the device's social community an opportunity to engage further. Devices are no longer dumb robots that happen to be connected to the internet, they are intelligent and capable of forming 'opinions' as they communicate with other Socially Connected Devices.

To illustrate the concept, consider an example: On a hot day in summer, all the air-conditioners across the city are running flat out to keep homes and buildings cool. To help with their efficiency they monitor the open air humidity, temperature and sun load, and they broadcast this information to the cloud along with their location. As a cool change arrives and begins to travel across town, a sudden drop in temperature and sun load is measured by outlier buildings and they report this news to the cloud. Over time, more and more buildings report this change in temperature and a trend of information begins to emerge. Buildings that are yet to be hit by the cool change take note of the trend and check the local weather site to see where the prevailing winds are coming from. They quickly deduce that they are in the path of a cool change and so rather than continuing to listen to their own sensors that are telling them the sun load is still quite high, they listen to the outside opinions of other devices that are telling them that a cool change is on its way. With knowledge of the thermal mass of their building and an estimate of when the cool change will arrive at their specific location, they adjust the cooling output of their chillers at just the right time to ensure a more consistent temperature is maintained within the building and power is not wasted. And in the background of all of this, the power station throttles its output to manage the requirements of the network.

Maybe it all sounds a bit futuristic but the technology to do it is all here today. So very soon I expect to see my smart house Tweeting its status to a whole cloud of socially connected devices.