As urban dwellers inhabiting 80% of our increasingly crowded planet, we are surrounded by countless helpers in our cities, our houses and our transport. These sensors and other “things” are designed to make our lives simpler and safer and remove the mundane tasks from our complex lives.
Thanks to the Internet, these inanimate objects can now connect and communicate with us and each other in that network known as “The Internet of Things” (IoT). From sensors in a garbage bin in Barcelona sending a message that the bin needs emptying, to those embedded in a Rolls-Royce jet engine checking for potential points of failure, these devices improve our lives. They also have the potential to cause immense harm if they are captured by cybercrooks.
There are already more IoT devices on this planet than human beings, and their numbers will multiply rapidly over the next few years. While the estimates for 2020 vary wildly from 31 billion in the Statista graph below to 200 billion according to Intel, the growth will be consistent.
If you were wondering how there can be such a difference in magnitude between the two estimates, there is a valid reason for this: a single definition of what constitutes IoT does not exist. The Institute of Electrical and Electronic Engineers Standards Association (IEEE-SA) published a definitive study on the IoT Ecosystem in 2015, where they pointed out that there were no standard definitions but two main schools of thought:
The IEEE-SA defined IoT in their study, opting for the evolutionary model:
IoT refers to any systems of interconnected people, physical objects, and IT platforms, as well as any technology to better build, operate and manage the physical world via pervasive data collection, smart networking, predictive analytics and deep optimization.
Whether you prefer the conservative or the explosive estimate, there is no doubt that the IoT is here to stay. IoT development is a particularly attractive choice for startups, as startup costs can be very low with a Raspberry Pi or an Arduino kit, or any of the other hobbyist alternatives out there.
The lack of barriers to entry in the IoT market also underpins the premise on which Intel's 2020 prediction is based. What is certain is that we will be surrounded by machines talking to other machines(“M2M” or machine-to-machine communications) and producing petabytes of data. In our cities, homes, factories, hospitals, and cars, there is already a stream of communications that will grow in size and complexity. Cisco, in their Cloud Index Report, predict that the volumes of data created will reach 600 zettabytes (ZB) by 2020, four times the 145 ZB generated in 2015, although they also estimate that only 10% of the data generated by IoT devices will actually be stored.
Most literature on IoT attempts to divide it into different categories and sectors, such as:
While this makes the IoT market easier to understand, it does not mean that the “things” cannot communicate across these imaginary boundaries and share data. Quite the contrary. A wearable device, which belongs to the consumer sector, can transmit your vital signs to your physician’s computer for analysis in the healthcare sector. A baby monitor, for instance, is both a consumer item and a health-tracking device critical to avoiding the risk of the sudden infant death syndrome.
The graphic below gives some idea of how pervasive IoT devices already are in our daily lives.
While some of these devices are extremely simple, the environment within which they operate successfully is complex. It takes far more than a technological solution; the IoT device needs to be integrated into an ecosystem, which has many components and different players.
There is a niche for virtually everyone to fill in the connected world. There are other sectors, such as agriculture, that are not specifically mentioned, where there is also great potential. Spanish IoT device manufacturer Libelium has exploited many aspects of agricultural IoT, and there is still plenty of opportunity there, from assessing soil and water quality to managing vehicles and outbuildings.
However, the IoT market is very immature, and engaging with it can be reminiscent of the days of the old Wild West. Some enterprises who have already made the move have already found this out.
At Cisco’s IoT World Forum, held in May 2017, they released the rather depressing news that 3 out of 4 IoT projects were reported to have failed or not met their goals post-implementation. This was the finding of a survey they conducted among approximately 1800 business and technology decision-makers in the US, UK, and India.
Over 60% of respondents admitted that they underestimated the complexity of the undertaking. When an audience poll was taken, based on the findings established during the survey, complexity and lack of expertise featured heavily as roadblocks. Security was also a major factor.
While there are many other contributing factors, here is why complexity, lack of expertise and security issues can derail an IoT project. When we talk about the “things” in Internet of Things, there are virtually no limits. From a simple sensor, like a thermostat, to a whole collection of sensory devices encapsulated in the modern motor car, each of them is busy collecting data to relay back to us. There is no common language or protocol that is used, and the storage and computing capacity of the simpler devices can be extremely limited. The simpler devices also have to use protocols that will use the least battery power available. This has given rise to communication protocols that support the smallest sensors, whereas the protocol alternatives, such as the low-power wide area network (LPWAN), Zigbee and Bluetooth Release 5, are growing, mainly to support the requirements of the IoT. The types of data vary too, and have different format, from streaming video from a drone to a person’s vital signs picked up by a wearable device.
An agreed set of standards would reduce much of the complexity, but the wide variances between all these components of the IoT make standardization something that is unlikely to happen in the near future. So for now, any IoT project will probably need to use most or all of the protocols.
In selecting a protocol, one has to consider many factors around what the operating requirements are for the device, the importance of the data, the frequency of transmission, quality of service and inherent security. Each protocol differs in some or all of these aspects.
It’s not all doom and gloom. There is a wide range of platforms available (at least 450 at last count) that assist the company in its endeavor to build their own IoT ecosystem. Basically, an IoT platform is middleware that polls and receives data from different devices using the suitable protocols, makes sense of what is received and can provide some or all of the analytics functions, which are the whole reason for entering the world of IoT.Selecting the right platform is another piece in the complexity jigsaw. It is highly advisable to get expert advice when making this decision. Development organizations have been exposed to the benefits and limitations of a variety of platforms at an operational level and are probably the best advisors. Lack of expertise can cause IoT project failure to the same extent as complexity, and wise companies have engaged a number of partners to assist along the way.
Bearing in mind all the different device types and protocols, expecting to find the expertise within the organization to successfully complete an IoT project is extremely optimistic. The consensus reached in both Cisco’s survey and at the IoTWF was that you should not attempt to “go it alone”. Partnerships and collaboration are an essential ingredient.Special pain points are device and protocol management, interoperability and integration, security and data management. Besides, if you are not operating a DevOps environment, a culture change will be required as well. The IoT does not allow for service unavailability, while demanding frequent changes to be introduced as swiftly as possible. Even if you follow the agile methodology within your IT division, your current rate of deployment may not meet the requirements of your IoT ecosystem.
Cyberattacks in the connected world are becoming more frequent, such as the Mirai botnet attacks in 2016, which infected over 300 000 IoT devices. While the perpetrators have been caught, there will be more attacks in the future. The challenge points back to complexity again: how much security can be embedded in a simple sensor? Not much, so there needs to be external firewall protection.
There is another aspect to security that is poorly understood by those implementing IoT, namely the privacy issue. With new statutes and regulations like GDPR coming into force, the protection of consumer data is paramount—the problem is that much of the data being transmitted to the company’s IoT platform is not recognised as personal data. For instance, a smartmeter transmits constant data about energy usage in a home. If that information can be siphoned off and analysed, it reveals when the house is empty and provides criminals with easy pickings, instantly becoming personal location data.
Expertise in securing the Internet of Things goes beyond conventional security as every device is a threat. The security has to be embedded in the design and not bolted on as an afterthought. If GDPR is one of your constraints you need expert advice on data privacy. This is also an opportunity to assure that you are completely compliant in your management of personal data.
As Rowan Trollope of Cisco commented, while many IoT initiatives failed in 2017, at least 26% succeeded. If you understand the challenges of the connected world, and that the road is not an easy one, it should be easy enough to find your company in that top quartile. It will be essential to:
The rewards are there. Even organizations that did not succeed in their initial IoT implementations remarked on how much they learnt from the experience and are willing to try again with the knowledge they have gained, aided by seasoned partners.