The concept of the IoT as a strange and wonderful nexus dovetailing the digital and the physical world already is a modern-day truism. Just, we’re not there yet.
Given the hype over the past few years, one would expect we’d see mass deployment across virtually any vertical by now. But we don’t. As a matter of fact, the adoption of IoT is behind schedule. You may remember the predictions on the installment of smart connected devices shrink over time, as the forecasters did their reality checks, figuring that anticipated growth did not match the reality. Until recently.
As we speak, IoT adoption is indeed accelerating. For the first time, the industry is operationalizing the promise of smart tech. Eventually, we’ve reached the tipping point of the IoT. But why now?
Until recently, IoT lacked the hotbed to breed innovation and implementation at critical pace. One key condition is only coming together now: ecosystems. Coming from individual, isolated applications, IoT architecture has matured into complete end-to-end systems, with data centers and use cases, which actually involve real consumers. This has become possible due to four technology pillars that are driving the current IoT wave.
Sensors to capture every facet of the physical world
Sensing technology has come to a point where sensors are super small, tremendously energy-efficient, and super versatile, which makes them a critical asset at the heart of the Internet of Things. Whatever use case one may imagine, we have reached a state of technology where any physical signal in our human world can be translated into electric signals and then into machine-processable digital data. Be it the mood of a person monitored by a home entertainment system, the saturation of a soil with nitrogen in a smart agriculture application, or the vibrations of a turbine in a predictive maintenance scenario – as an integral part in the perception of the human environment, digital sensing has established a new connection between our lives and technology.
The second driving force at work here is connectivity. Established, field-proven standards are now covering the entire scope from ultra-short-range connectivity like NFC over mid-range (WiFi, Thread, Zigbee, V2X) to wide area networks and entire cities (Lte, LoRa). For wide area networks, the promise of 5G is looming at the not too distant horizon. As we speak, the first infrastructure trials for 5G layouts are rolled out in Europe, China and the US. In addition, more advanced and powerful technologies such as Ultra-Wideband (UWB) are being implemented, bridging the last gaps and enabling an entire new suite of use cases.
At the same time, adoption rates for the well-established technologies are picking up momentum. Just a month ago, the 100th Chinese city has rolled out NFC-based contactless payment for public transport, enabling consumers secure and easy access to mobile ticketing. The full spectrum of wireless connectivity, and the enablement of fast, cost-effective commercialization of easy-to-use mobile payment devices with the ability to scale rapidly, is obviously another key driver for the IoT.
Compute has shifted to the edge, AI will be for everyone
The third pillar is “edge”: A fundamental change in innovation is unfolding that is pushing more smart tech to the edge than ever before. The old model, where data is generated at the edge and processed in the data centers, has come to its limits. Consequently, computing is shifting to the edge – rapidly. In fact, IDC forecasts that in just a year from now, already 43 percent of all IoT computing will occur at the edge.
And there’s good reason for this. Dedicated edge processing reduces response time and network congestion. Autonomous vehicles for instance, will rely crucially on real-time processing to make correct decisions within a fraction of a second. Dedicated processing eliminates the need to build inefficient and unresponsive centralized cloud data centers to handle the significant increase in data collection. It is more reliable at the device level, and it will result in better user privacy, since raw data won’t be uploaded to the cloud.
Edge processing is based on distributed resources that may not be continuously connected to a network in such applications as autonomous vehicles, implanted medical devices, fields of highly distributed sensors, and a variety of mobile devices.
To make use of machine learning (ML) and to enable artificial intelligence (AI) in this challenging environment, an agile application is necessary that can retain learning and apply it quickly to new data. This capability is called inference: taking smaller chunks of real-world data and processing it according to training that the program has done. For inference to work in edge environments, processing architecture and hardware are required that are optimized and come with certain requirements on processing capacity, energy efficiency, security, and connectivity.
Advanced hardware fulfilling these requirements is readily available and has given IoT adoption a strong boost. What comes on top is the democratization of previously exclusive technology. Machine learning, until very much recently, was available only to those with sophisticated cloud architecture, advanced algorithms and access to massive real-world data sets. With a set of new toolkits, edge node developers are now able to efficiently integrate and run cloud-trained ML models in resource-constrained edge devices, and thus enable a broad range of industrial, IoT and automotive applications. Removing the heavy investment necessary to become ML experts has enabled tens of thousands of customers whose products need machine learning capability.
According to a recent Cisco study, only 9% of respondents in the US have high levels of trust in IoT devices. Or, in other words, what are all the blessings of life in the digital age, when one must fear one’s fridge betraying one’s privacy?
The same study says, 42% believe that IoT services deliver significant value to them. While IoT adoption is accelerating across all industries, businesses and end users alike are facing constant privacy and security threats. This paradox underscores that establishing trust in applications is paramount for the IoT to succeed.
In fact, we see vulnerabilities to IoT devices being announced at a rapid rate, such as the Meltdown and Spectre vulnerabilities of CPUs, the ROCA or Heartbleed attacks. Many of these incidents highlight that we are increasingly reliant on a few, dominant system building blocks, which have not been thoroughly security vetted.
Against the backdrop of these threats, stakeholders now seem to have understood that a joint effort in building trust is mission critical. And indeed, change is happening. For the first time, we see frameworks such as the General Data Protection Regulation (GDPR) in Europe or the California Consumer Privacy Act of 2018 (CCPA) come into effect that are defining common guidelines on data security and privacy. On the industry side, initiatives of leading players like the Charter of Trust demonstrate strong commitment in advancing cybersecurity and securing the digital supply chains.
To fulfill that promise and create an Internet of Trust, advanced providers design their products, systems and services with security testing in mind. They are upgradable so that patches can be applied after a security breach, which will inevitably happen at some point. Systems are designed in a resilient way to prevent the collapse of an entire network due to an attack. And, because security requirements vary considerably for different markets and applications, scalable security is created in the architecture of products and services.
Exactly this is what secure connectivity at NXP is about. With a successful track record of providing solutions to secure ecosystems such as secure microcontrollers, NFC, payment, access control and high-speed network switches, our engineers and businesses are playing a key role in the creation of the Internet of Trust. Success in this field is rooted in a system approach, where not only secure products are delivered, but it’s also ensured that design for security and design for privacy solutions are provided to customers. This security by design approach also encompasses design for secure manufacturing, secure trust provisioning and secure delivery, thus building the fourth pillar of the IoT ecosystem.
2020s will see the IoT getting real
A new world of sensing, world-spanning connectivity, the shift from center to edge, and advanced security to shield systems and devices are the key drivers for the IoT to take the step from exploration to actual implementation.
If we look into the future, we expect 75 billion connected devices in 2025. With literally trillions of end-point devices, new applications and business models evolving around them, the shift from center to edge is nothing less than a tremendous opportunity.
This is not about adding value and creating new business models alone, but to make our lives easier, simpler, more convenient and more secure. If technology isn’t helping to advance people’s well-being, then there is little reason for technology to exist. The IoT holds the power to transform our lives, and it’s upon us, to deliver on that promise.
This article was originally published on Kurt Sievers’ profile on LinkedIn.
We are curious to hear your view. What is driving today’s adoption of the IoT? Engage in the LinkedIn discussion under this link.