Ten challenges the international IoT community needs to master (1/2)
We expect the Internet of Things (IoT) to be the biggest system that mankind has ever built: In a few years’ time, 75% of the world’s population will have access to the internet and by 2015 alone, we expect there to be more than 6.5 billion objects connected to the internet and cooperating partially without human intervention. The internet has always been a success machine. Looking back, we can see that the internet developed in three waves. The first brought easy exchange of documents, the second provided for commercialization of online businesses, and in the third wave people started engaging through social networks. The next wave is the Internet of Things, a gigantic wave that lets people, things and services interact autonomously around the clock. The Internet of Things will have an impact on how we live and move, how we produce and consume energy, and how we do business or manufacture things. The international IoT community is facing technological, societal and business challenges and must ensure that this next internet wave breaks smoothly on our beaches. There are ten IoT challenges to master. Some are new and some are old acquaintances that need a new interpretation based on specific IoT requirements.
Challenge #1: Robust connectivity
The basis for making the IoT happen is connecting things: billions of things, and diverse things from tiny acceleration sensors to video cameras and cars. Then, add all those objects that have a limited energy supply, are distributed or have an unstable internet connection. These factors all challenge connectivity and make engineering the IoT a tough task. One path we have to keep following is to work on energy harvesting devices, which can increase connectivity through their self-catering energy mechanisms. And second, we need an “IoT lingua franca”, a standardized way for these things to talk to each other, e.g. replacing TCP/IP with IPv6 or ensuring communications at the level of web services that are based on http and the REST principle.
Challenge #2: Useable security
Security is an issue when everything is connected. I believe that the technology is out there to implement proper security concepts; however, it does not fit to IoT requirements. Entering a password every time we access a device or sensor is far from reality. Security in the IoT needs to be as simple as a key that you give to your kid to open the front door, not as complex as millions of keys. Otherwise it becomes more complex to manage the keys than to provide the security itself. What I can see from a business development perspective and from my engagement in start-up initiatives is that the IoT offers niches for new security concepts that revolve around ease of use at known levels of privacy and protection. There is more to come!
Challenge #3: Big Data
Big data is inseparable from volume, variety and velocity – the three Vs. They are blended with the issues that distributed storage brings with it. When we look at where big data mainly emerges today, it is the content that users share through Web 2.0 and companies leverage to gain additional insights. So here comes my maybe unexpected thesis: the traditional V challenges will not have to be tackled by the IoT community, because they will be solved by Web 2.0 and the people producing all content. In the IoT, data comes in smaller pieces, so the bandwidth we need is much lower; what’s more, we don’t need all the data, we can aggregate it or even throw it away. When we’re talking about big data in the IoT, what we need is analytics and semantics that ensure we collect only relevant data, have mechanisms in place to filter irrelevant data, and use technologies that can handle massive amounts of unstructured data. What is even more important – and is something the IoT community is responsible for – is to create a functioning data market, as the current big data business model is based on advertising.
Challenge #4: Big Code
Maybe this is a new topic when talking about the IoT, and of even greater importance than big data. If we mess up big data, we will lose data, and in many cases this won’t be too harmful. However, if big code doesn’t work… Think of some of your applications: you are programming 1,000 or 2,000 lines of Java code, then you don’t just compile it, you interpret it on the machine; and how much code is there on the machine already, e.g. in all the libraries and operating systems? How do you ensure that it’s not just your code that really does what it’s supposed to do? You write only a tiny fraction of IoT code and rely on everybody else to perform as well. However, in reality, systems out there naturally come with hundreds of bugs and our current testing paradigm assumes the perfect world – which it isn’t.
Challenge #5: Information Models
For the IoT, it is essential to translate the physical world into a format that can be handled by IT. One way to do this is with information models. This is where domain knowledge is transferred into software. Let me give you an example: a connected home application needs to have access to an information model about rooms, floors, the location of devices and their functions. In product design, engineers make heavy use of information models in the development process, e.g. simulation models, prototypes, CAD drawings, or 3D modeling. However, we stop using these techniques later in the operational phase. For the IoT, we have to get used to constantly using these information models and blending them with lessons learned from operations. It’s at this point that the model becomes part of reality and reality becomes part of the model – which is basically the IoT: connecting the virtual with the physical world.
These are the first five challenges that I see for the IoT community to master – stay tuned for my second part, in which I briefly discuss IoT governance, accountability, open platforms, business models and ecosystems! Any comments so far?