The many faces of IoT

There have been years of hype generated around innovative IoT technologies that we previously never imagined possible. From connected fitness equipment to industrial smart gloves to intelligent water management systems — the possibilities now seem endless.

Change is certainly on the rise and is transforming our industries, lives and world. Change driven by fundamental technological shifts: cheaper, more powerful hardware, nearly ubiquitous connectivity and cloud computing. As with any massive transformation, the nature of exponential growth takes time and often comes more slowly at first.

However, this giant market we call IoT — encompassing everything from wearables to autonomous vehicles to smart homes, factories and cities — simply does not exist. There is not a broad, homogeneous set of applications that we can call IoT. Instead, there are many, varied sets of applications, each enabled by the same tech trends, but manifesting themselves in different ways.

In order for us to ensure we develop a world where the benefits of connected devices outweigh their risks, we need to start looking more closely at how we are defining IoT and how we can ensure the success of all different types of IoT deployments.

IoT and the evolution of computing

We are witnessing an evolution of computing as it expands from mainframes, one computer for many people; to desktops, one computer per person; to mobile, multiple computers per person; to the phenomenon we see today: one-to-many computers per thing.

The devices in this latest computing wave are different from the ones before it in an important way: They’re incredibly diverse.

PCs were nearly identical, almost all running the same OS — if you remember, we even called them clones. Mobile phone hardware is somewhat more varied, but we’re down to two operating systems — and the need for app developers to have a consistent set of APIs has reduced hardware variability even more.

IoT devices, however, are defined by a variety of constraints: available power, connectivity/bandwidth, computation and cost. Often these constraints are entwined: less available power leads to lower power data transmission or longer duty cycles between engaging the radio which leads to lower available bandwidth. These constraints are set by the environment within which these devices are serving. For example, connected home products are often not energy limited — they’re often powered via electrical outlet — and enjoy high bandwidth via Wi-Fi or Ethernet, but may be cost-constrained by consumer budgets. On the other hand, sensors used in oil and gas may have a larger budget, but with limited power and network access because of the remote nature of the work.

Source: Timescale

These constraints can also translate into entirely different network topologies. For example, an Amazon Echo talks directly to the internet via Wi-Fi. But in a factory, low-powered sensors can communicate via a low-power protocol — for example, Zigbee — to a local gateway, which then could use Wi-Fi or Ethernet to communicate upstream. And in a remote mine, sensors may communicate via multihop mesh back to a gateway, which may then use a cellular network to transmit upstream.

Source: Timescale

Of course, these environments can also lead to entirely different businesses — for example, a direct-to-consumer or retail model for consumer products, an enterprise sales model for industrial sensors or an RFP-driven process for smart city devices.

There is a rich diversity at work here, far greater than what we have seen before in PCs or mobile phones. These IoT devices represent a broad spectrum of reds, greens, blues and violets, yet we continue to lump them all under a bland white umbrella, losing what makes each color so unique. This hampers our ability to understand and serve these markets.

IoT data is different

In order to better understand the behavior of various IoT applications, we must first recognize that the survival of any IoT project depends on the data it collects.

After taking a look at the examples above, you can imagine that the complexity of data varies from application to application. One thing that remains consistent is the need for a scalable, reliable and easy-to-use database that can work with you as your business grows.

IoT organizations often times find themselves collecting different types of data, such as relational, geospatial, time-series and metadata. In order to avoid creating data silos, it’s important to be able to join this data together in a single-pane-of-glass view. This enables various teams across the organization to access and gain a better understanding of the data that can impact future decision-making.

Another attribute of IoT data is that it is extremely high volume. For example, a single connected car collects 4,000 GB of data per day. Organizations must build their data infrastructure around a system that is able to scale with them. By taking a proactive approach and building a solid data infrastructure, IoT companies can mitigate risks that are associated with the abundance of complex data that they are undoubtedly collecting.

Expanding our world

As we enter a new wave of computing, one that will produce even more change driven by a world of connected devices, we need to be educated and prepared.

This phenomenon is enabling change in many different ways, each of which is its own precious little snowflake: smart home, connected vehicles, preventive maintenance, precision agriculture, asset tracking, fleet management and so on.

I don’t know how long it will take for this phenomenon to develop across these various applications, nor which of these trees will bear fruit sooner rather than later. But I do know that the sooner we expand our knowledge and change the way we treat different IoT devices, the sooner we can expand our world.

If you are building a new IoT application, be sure to carefully consider the technologies you work with before you find your infrastructure crumbling to the ground.

Are you working on an IoT project you want others to know about? Leave a comment in the section below!

All IoT Agenda network contributors are responsible for the content and accuracy of their posts. Opinions are of the writers and do not necessarily convey the thoughts of IoT Agenda.

Data Center
Data Management