Take a look around you. How many devices do you see that are connected to the internet? Your cell phone, your computer, your TV, perhaps even your car or your fridge. More and more devices are becoming connected to the Internet and to each other. I can now use my smartphone to monitor basically my whole house. The home temperature, potential intrusions, utility usage. From wherever I am, in near real time. Not only can you monitor all those things, they can also communicate with each other and adjust automatically. Also outside of the home environment, objects of all types are becoming connected. From the traffic light, to the power. This hyper connectivity is becoming more and more relevant for industrial use also, for transportation, traffic management or production. What is behind those connectivity? What is the Internet of Things? It is a network of connected objects, that can build an aware, autonomous and actionable system. Let's take the self-driving cars as an example. The self-driving car, is aware of its surroundings. For a data, a GPS system and probably a few cameras. It is autonomous and analyzing its position, and drawing a map of its surroundings. It can also communicate with other cars or via a central server, to check the traffic conditions along its route. It is actionable when it decides from all those inputs when and where it will go next. And it can translate that into commands for the engine, the brakes to steering its wheel. This system as a whole, can get pretty complex with multiple components interacting with each other. Let us try to make it simple. Imagine there are three different layers interacting with each other all the time. One layer is a connected objects. The rada, the cameras, the intrusion detectors, but also your utility meter, your watch, all those devices need to interact with a central platform that can consolidate the data and decide what to do with it. That's the second layer. On the other side, as a third layer, your few services that can understand the data aggregated by the platform and translate it into action. Between those layers, data is constantly exchanged using networks, which can be anything from Wi-Fi to satellite, or fixed to mobile operator lines. So, we have some processing, some transmission, and some storage. Does this sound familiar to you? Do you remember what we have discussed in the videos on big data and cloud services? The technology required for each component became more stable and cheaper to manufacture. It's not consuming less energy and space for high capacity. This made it economically efficient to embattled more and more objects, which has then driven the increase of IoT devices in recent years, and will further fuel this. The number of connected objects is expected to increase by 30 percent per annum until 2025. This means by 2020, we will have between 50 to 75 billion connected objects. Just think about that for a second. It would be seven to 10 devices per human being on the planet. This would translate into a total incremental spend of 250 billion dollars across industries. All major players in the high tech industry already offer Internet of Things solutions, but also traditional players do, the utility providers for example. So what drives the adoption of IoT? This is what my colleagues, Seer Yusuf, Michael Reeseman and Nicholas Hunke have looked at in the recent study. One of the major finding is that IoT adoption is driven by specific use case scenarios rather than by industries. Some of those use cases do not only incrementally improve existing operations, but also create innovative products or even business models that were not possible before. The team has done a fantastic job identifying a wide range of IoT use cases. They've selected 10 use cases from a long list, that have the highest spend potential and a time for maturity lower than 5 years. Let us have a look at those use cases. We will start with the closest to maturity, Demand response. If you live in Europe or North America, you probably know that our energy production is under a lot of stress in peak periods during winter. Smart appliances can sync with a grid and figure out when is the right time for them to run, depending on the overall network supply and demand. Another use case building on this is distributed generation and storage. This is when a company can remotely decide when and where to excess energy, and when it triggers additional capacity into the grid for its peak demand. Then we have smart meters to track and optimize your water electricity consumption. Other use cases are fleet management, track and trace, automated inventory, predictive maintenance, self optimizing production and connected cars of which we had just talked before. The last one of the list is patient monitoring, which is expected to still need up to five years to reach maturity. You will probably agree that this is already quite an extensive list of things to expect in the next five years, and there's more to come in the next five to 10 years. But as you've seen from this use cases, they are not equally distributed across industries. 50 percent of the expected IoT spending is generated by just three industries: Discrete manufacturing, transportation, logistics, utilities. Overall, IoT is still in its early stages of maturity. Still, there is no one single IoT. Multiple players are trying to consolidate the use cases we've just talked about and win the platform war. The key questions that companies need to ask themselves are on operability. Will the platform they are using today be inter-operable with others tomorrow? Will we see the days of full interoperability between devices, services and platforms? This reminds me of the journey the Internet as we know it today went through. You might not remember it, but there was a time when it was entirely possible that we will end up as multiple connected networks. They don't connect to each other globally. This was before people like Van Jacobson standardized the TCPIP protocol, a protocol which guarantees a single fully scalable internet as it is today. IoT is not there yet. We have a number of platform service providers. We counted more than 400 this year. The reality is only a few would survive. And they will do so if they are able to federate a large and AI community of application developers around them. Open source might be one path to full interoperability. In the meantime, what do I advise my clients when they are looking for a platform for data? Obviously, a technology offer that's compatible with the company's system. A modular architecture, and easy to use APIs, that can prepare the company for future integrations with other IoT ecosystems, other application or service providers. Let's recap the key messages of the last few minutes. A complete IoT system will have functions of sensing, processing and action. It's aware, autonomous and actionable. By 2020, we'll have seven to 10 IoT devices per human being on the planet. This is driven by further technological advancement in processing, transmission and storage. We are not yet in the era of the one Internet of Things. We need to think industry by industry, company by company. What is the specific use case scenario a company wants to pursue? A firm solid partner. Whatever a company chooses, it should keep in mind, that it needs to factor in the challenge of interoperability. The more modular the solution architecture is, the more likely it is to be durable.
