There are experts who believe adoption of the Internet of things (IoT) will progress substantially in the next decade and bring with it considerable benefits. Others think the problems will outweigh the benefits. South African-born Dirk Swart, co-founder of US-based Wicked Device (www.wickeddevice.com), an electrical engineering consultancy that develops Internet-enabled devices, shares his insights on what it all means.
Brainstorm: Gartner predicts that, by 2020, 'the installed base of the IoT will exceed 26 billion units worldwide and will consist of a very diverse range of smart objects and equipment'. These are pretty hefty figures. What do you make of them?
DS: The term 'Internet of Things' encompasses a huge range of things. I'm going to talk specifically about small, low-power, cloud-connected embedded sensors and actuators, because that's my particular field of interest. If anything, the figure of 26 billion is at least one order of magnitude too low if we count each device. There are already a massive number of embedded devices that are not connected to the Internet. What we're doing now is connecting them.
It's not as if you go out and buy an 'Internet thing'. People are purchasing specific solutions to specific problems. For example, they might want to measure how much pollution is present in the air they breathe, or turn their house lights off when they're out. This specificity is what allows devices to be cheap, reliable, low-power and easy to use. A major hurdle in the near future is going to be creating a management infrastructure for all these devices that allows a lot of control to be done automatically. Consider your car - it has numerous microcontrollers on it, none of which you actively manage, or are perhaps even aware exist. The manufacturer has created a sophisticated control system so you don't have to. A good Internet of Things management interface would allow you to orchestrate meta behaviours without having to specify each device's specific operation.
Scepticism
Brainstorm: Green Internet consultant Bill St. Arnaud is a vocal sceptic when it comes to the IoT, saying it "has been in the red zone of the hypometre for over a decade now. Yes, there will be many niche applications, but it will not be the next big thing, as many pundits predict. If the Internet of Things had any true validity, you would think you would start to see evidence of its presence on early-adopter Internet networks." Your view?
DS: Connected devices offer great efficiencies. We're seeing evidence of this every day. In a resource-constrained world, the advantages will be overwhelming.
On the subject of niche applications - the 'things' themselves are tremendously specific, as each is manufactured to do only one or a small number of tasks in their specific niche. It's all niche applications. That's what the Internet of Things is: lots and lots of niche applications tied together by a management infrastructure in the cloud. Overall, each niche, each possibility for creating efficiency, will be filled with a small embedded device.
Few people today imagine driving somewhere new without a GPS for the same reason - it's more efficient. The key will be to make sure the efficiencies benefit you, not your supplier, and this is going to be very difficult to achieve.
Consider this analogy: if I want help with my phone service, I have to wait on hold for several minutes (at best). Why? Because that is most efficient for my phone provider. It's annoyingly inefficient for me, but they don't care about that. In the same way, there's going to be tremendous pressure for a big chunk of the benefits of these devices to accrue to the providers of cloud services, not individuals.
Brainstorm: Professor Peter Jacoby believes that the effects will be "widespread but pernicious. We might as well inject ourselves into the Internet of Things. By 2025, we will have long ago given up our privacy. The Internet of Things will demand - and we will give willingly - our souls."
DS: Privacy is a big issue, but for the most part, the privacy ship sailed with the advent of smartphone apps. And for most people, the smartphone is going to be the interface of choice for interacting with Internet-enabled objects. One consequence of this is that we're entering the age of differentiated pricing. Micro data analysis allows large data owners to dissect your data, and offer you prices and service offerings that are customised to you. The goal is to maximise revenue, by knowing what you want and when you will pay. That's a somewhat unattractive choice for most of us. Unfortunately, the solution is to restrict the power of the data collectors via legislation and other means. These are difficult levers for most of us to move.
I'm reminded of some satire in a recent edition of the New Yorker, in which somebody's banking smartphone app makes a variety of humorous comments on purchases, becoming more and more HAL-like as the story progresses, to the anger and detriment of the user. Devices really are going to have a real impact on our purchasing behaviour.
Another impact will be the inability to buy features you don't want. For example, it's possible that you won't be able to get a fridge without a connection to the manufacturer.
Open source is one answer to privacy concerns, because it separates out this desire to profit by adding features not wanted by the customer. People need to build management infrastructure that is not obligated to a particular manufacturer or cloud services owner. But that is difficult and expensive to do. I guess we can all hold thumbs that another Linus Torvalds emerges for the Internet of things.
Interoperability
Brainstorm: What about interoperability? More devices, more sensors, more software. How can we ensure they all work together and don't sit wastefully on separate islands of data?
DS: Interoperability is often viewed as a technical issue, but most interoperability is based on business decisions. The reason the iPhone app store doesn't provide Android applications is because Apple made a business decision to create a walled garden. In the same way, it's reasonable to expect that several islands of interoperability will develop in this space.
The resulting data will be freely available to be exchanged, and there will be no technological reason not to do so. On a technical level, once the control infrastructure is in the cloud - which is, after all, a goal of the Internet of things - adding a management hypervisor system will solve the interoperability issue, and it's absolutely achievable today. Of course, whether this happens or not depends on the respective business strategies of the companies involved.
At a 50 000-foot view, things like net neutrality are going to be much more important than the interoperability of individual devices. These devices all need the Internet to gain these efficiencies, and that clearly makes it a common carrier.
First published in the July 2014 issue of ITWeb Brainstorm magazine.
Share