We have made an announcement today about an exciting project we're leading that is designed to accelerate growth of the Internet of Things (IoT).
Many of you will already be familiar with what the Internet of Things is, but the basic idea is that over the next decade everything that can be usefully connected will be connected. The hundreds of millions of servers and billions of people who are connected to the internet today will be joined by tens of billions of new netizens, nearly all of them real-world objects. It's an astonishing idea that could make many aspects of everyday life better and more efficient.
Start with better. Devices that can communicate with you will be able to personalize themselves to your preferences, and offer new user interfaces. They might even anticipate what you want to do and respond accordingly - light bulbs with an app are just the start. Other devices will be real-world extensions of web services, providing you with ambient information, or instrumenting the environment.
Instrumentation allows us to measure things, measuring things allows us to understand them, and understanding them allows us to manage them more efficiently. Until recently we lived in a comparatively data-poor world where many things were difficult and expensive to measure. Even a simple question like "how many people live in this country?" required a stupendous effort every few years to collect the data. Now we can look at school registrations and mobile phone density and get a figure just as reliable the old one, at a fraction of the cost. As more activity moves online, where it's easy to measure things, we're getting incredibly valuable information almost for free. That's only increasing as open data movements make public data accessible, and as big data recombines and mines it all for insights.
Deployments of connected sensors are already bringing those online advantages to the real world. Bricks and mortar retailers are taking on their online competition by collecting equally detailed information about shopping habits. Insurers are already able to offer policies personalized to the driving habits of teenage drivers. Transportation for London is reducing disruptive breakdowns by detecting incipient faults and doing preventative maintenance. OEMs of all kinds are looking to connectivity to add new features and understand how their products are really used.
It only takes a simple cost-benefit calculation to decide whether it's worth connecting something: will you get more value from being able to talk to something than you would spend on connecting it up? ARM is working on both sides of that equation.
It's already possible to connect almost anything, and in many high-value applications they already are. To see many more things being connected we need the cost of doing so to drop a lot lower. Energy consumption is often the biggest factor in lifetime cost of ownership, so that's where we're focusing: not having to change batteries saves a lot of manual labor, as well as the cost of a mains connection. The biggest energy drain on a connected sensor is often its communications link, so we're supporting energy efficient communications standards such as Weightless and 6LoWPAN.
On the benefits side, we're working on a project based on a simple observation: data from connected systems often has value to a lot more people than originally anticipated. Things tend to get connected for a particular reason, so right now all the potential benefit from other uses for their information gets left on the table. Most of the time nobody even realizes there are other uses for it, and even if they do, it's not useful for them so they've no reason to do anything about it. Even when interests do align, a lot of those uses are unproven so it's simply too risky to incur the costs of setting up the technical and contractual connections to enable them. We're working with several of our partners in our Open IoT project on a way to resolve all these issues at once.
Our principles are simple:
- Discovery and use of IoT resources should be completely automatable, so that innovating new services is very low cost.
- New connected systems should strengthen and benefit from everything that was previously installed, creating a network effect.
- Data collectors - including individuals - should have confidence to share data, and should realize benefits in return.
- Dictate as little as possible. The IoT is a big place, and people have many problems to solve.
The Open IoT project brings together eight partners, enabled by a generous grant from the Technology Strategy Board under its IoT Ecosystem Demonstrator competition. Three of the partners - Enlight, AlertMe and IntelliSense.io already have fully fledged connected systems, connecting their own devices to their own web applications in order to deliver a service to their customers. This siloed approach to IoT has been the only way to do things until now, but for different reasons each of these three feel it is now an impediment to their growth.
In this project, they all have a common customer: ARM's headquarters site at Peterhouse Business Park on the edge of Cambridge. ARM has commissioned over 600 sensors from them, each one for a specific purpose. AlertMe home energy monitoring kits will let employees monitor their own energy consumption, part of an investigation of ways to meet our carbon reduction commitment. We're always short of meeting rooms, so AlertMe's occupancy sensors will monitor them, helping us use them more effectively. EnLight's street lighting system is being retrofitted on lamp posts throughout our car parks, saving 30% energy, improving lamp life, and cutting maintenance costs. IntelliSense.io's sensors are monitoring building heating, ventilation and water systems, letting us measure the effectiveness of efficiency measures and right-size investment in replacement plant.
The TSB funding lets us connect these IoT silos together and to new applications, and along with the other eight projects funded under the competition demonstrate that other applications for the data always exist. Later on in the project we'll be holding hackathons, integrating additional systems, and inviting another partner, Red Ninja Studios to create a new application exploiting the ARM site data-sphere. To test the idea that there are lots of small but useful applications almost everywhere, we'll also be deploying Weightless technology from another partner, Neul, and inviting people to use it to instrument and use additional data 'contexts' in the surrounding area.
Sharing data is not just a technical problem. Its most difficult aspects are actually the business models, contractual undertakings and privacy provisions that are every bit as necessary for our approach to be successful. We've spent the last three months developing a very lightweight technical approach though so we can share a first draft with you today.
Put simply, it is a catalog format that lists IoT resources and their descriptions. Anyone - or any machine - who understands the catalog format and understands the description of a resource can go and use that resource. We have deliberately not tried to define how to describe different kinds of resources - there are lots of projects out there already doing that, for example W3C's SSn, IPSO's Application Framework, OMA's Lightweight M2M, or OGC's SensorML. Although it would be helpful if everyone described a streetlamp in the same way, trying to force them to will only make people walk away. We're leaving it up to their own enlightened self-interest. It's worth noting that this approach is what has enabled all the 40-odd companies in all the funded projects agree to it, led by 1248.io, a startup that is also one of Open IoT's project partners. We can be confident this approach works for nearly any use case.
Each catalog record uses a linked data-style triple, so we can access the benefits that linked data brings, but simplified so that learning the catalog format doesn't require developers to know everything about linked data in general. The catalog specification and API are already public on the Open IoT website, and we're planning to now build up a community around it and make it progressively easier to engage with. In particular 1248.io are building an open source reference implementation, and we're going to create an identity for the catalog standard - Open IoT is unfortunately already taken!
You can find it here: http://www.openiot.org/
Check it out, follow us on twitter @OpenIoTProject, and if you'd like to get involved please get in touch!