How will the Internet of Things (IoT) impact the enterprise? Asking this today is like asking in 1900 about the impact of the internal combustion engine. There will be technical and commercial developments over the next 50 years, and we cannot clearly see that far ahead. There are some trends that we can predict with confidence for the next five years, and we can see principles that will continue to apply beyond that, but we cannot forecast their effects.
By Chris Harding, Director for Interoperability, The Open Group.
The IoT is the network of connected devices that enable computer systems to monitor and control aspects of the physical environment. Sensors and actuators have been used in computer systems for many years; it is the ability to connect to such devices anywhere in the world through the Internet that is new. The IoT has applications in many areas, including personal and home environments, smart cities, factory automation, and transport.
The first trend analysts are predicting is growth. IoT software and solutions are at the top of the list of emerging technologies Forrester thinks you need to follow closely. Gartner projected that 8.4 billion connected “things” would be in use in 2017, up 31% on the previous year. Boston Consulting Group estimates that companies will spend an incremental €250 billion on IoT in 2020.
Another clear trend, contributing to the first, is increased functionality at lower cost. Moore’s law applies to intelligent devices, just as it does to IT in general. The cost of basic sensors and actuators is already low, so the trend will be to add to them communication and processing capability. A basic wired temperature sensor currently retails at under $5. Make it wireless, and you will at least double the price, but that will often not be important; for example, it costs much more than $5 to dig up a street, if that’s what you have to do to connect with wires. Adding storage and processing capabilities might double the price again, but the device will still not be expensive.
The internal combustion engine affected different enterprises in different ways. Rail companies lost traffic to road and then air transport. Farms produced more by replacing horses with tractors. Enterprises generally could operate more effectively in many ways because they had faster, more flexible transport; for example they could sell more effectively through travelling salesmen.
Enterprises will benefit from the IoT in different ways. Transport companies will be able to use vehicles that can sense their environment and drive themselves. Farms are experimenting with automatic cultivators that can distinguish weeds from crops. Enterprises generally will operate more effectively because they have intelligent buildings; for example their offices will be more comfortable to work in and more economical to run.
Some enterprises will design IoT systems, but most will use IoT products and services supplied by other companies. An intelligent building control system that regulates lighting and heating in a company’s offices might be a purchased product or provided as a service by a specialist supplier; few companies will develop their own building control systems. An automatic cultivator could be a product purchased by a farm. The same farm might subscribe to a crop monitoring service that combines information from sensors in the fields with weather data and analysis of images taken by drones to deliver data on crop growth, forecasts of harvest dates, and alerts of pests and diseases.
The challenge for enterprises using IoT products and services will be to integrate them with other systems. A building control system that includes control of access to secure areas, for example, will be integrated with HR systems, to ensure that the right people have the right access. An automated cultivator will use field plans and crop records, which will also be used by the crop monitoring service, and by other farm systems, and will share with them data gathered during its operation. The farm’s data will also be shared with government systems, for example to contribute to nation-wide analysis of crop health.
The ability to integrate IoT products and services with other systems within and outside the enterprise is a critical factor for successful use of the IoT. Other critical success factors include some, such as reliability and cost-effectiveness that apply to any system. They also include security, which applies to other systems, but has particular considerations in the context of the IoT.
Integration of the IoT products and services used by an enterprise requires them to have open interfaces and a shared interpretation of the data that they exchange.
Unless there is a need for sub-second response times, open software interfaces usually take the form of Web APIs. Many IoT systems do have real-time requirements, but sub-second responses are generally required within systems, rather than between them. IoT products can expose Web APIs over the Internet. Services that use the Cloud for data processing and storage typically do expose Web APIs over the Internet. Enterprises should select IoT products and services with stable, well-documented Web APIs, for integration with other systems.
Data exchanged using Web APIs takes the form of name-value pairs, in which the name identifies the item of data that has the value. The systems exchanging the data must interpret the names in the same way. It is no good, for example, if one system interprets “temperature” to be the temperature at a given time in degrees Centigrade while another interprets it to be the average monthly temperature in degrees Fahrenheit. Enterprises should select IoT products and services whose API documentation uses standard vocabularies to specify the interpretation of the exchanged data.
Security is a particular concern for the IoT. Some IoT systems can control their environment, so that a security breach can directly result in physical damage, as well as damage to the system’s software and data with consequential financial, emotional and reputational damage to users and their associates.
The infrastructure of most countries contains IoT systems. Their security is a matter of national importance. A breach that renders the power grid unavailable, for example, could have severe consequences for the comfort and even lives of its inhabitants, and for the health of its economy. A hostile country might use its national resources to engineer such a breach, as an act of war. A commercial enterprise that has IoT devices in its infrastructure might suffer such an attack by a rival enterprise. Criminals could use the threat of such an attack for extortion.
Some IoT devices are in physical situations where it is difficult to maintain them. Particularly with older devices, it may be difficult to apply software patches to address security vulnerabilities. These devices leave their systems open to attack. They can also be “captured” by attackers and used to attack other systems, as in the Dyn cyberattack that disrupted the Internet in 2016.
The game of chess, which is an intellectual model of warfare, is also a model for IT security. There is no strategy that guarantees success; you can only win by playing better than your opponent. Enterprises using the IoT must analyse the risks involved, and use appropriate countermeasures to eliminate or mitigate them.
The applications of the internal combustion engine are all very different, but they do have a common characteristic: more of what, in a person, would be muscle power. Enterprises using them gain physical strength. With the IoT, by this analogy, they gain sensory perception, and the ability to apply their strength precisely. The IoT gives them what, in a person, would be eyes, ears, and hands.
Processing IoT information will take a substantial proportion of enterprise computing power just as, in a person, a substantial amount of brain activity is devoted to sensory perception. It is estimated that about 30% of neurons in the brain’s cortex are devoted to vision, 8% to touch, and 2% to hearing. Enterprises are using big data analysis and cognitive computing to gain business insights from IoT data. This will account for an increasing proportion of IT spend over the next 10 years, and may well reach the 40% suggested by the human analogy.
The current trend is for this to be spent on vertical industry applications and the computing resources to support them. It is unlikely that this will change. The idea that a general enterprise sensory perception system will evolve is carrying the human analogy too far. So we will see major vertical IoT solutions that include both physical devices and specialist analysis and control software.
The IoT will have as big an impact on enterprises as the internal combustion engine did. It will give enterprises the ability to sense and control their environments in a coordinated way. There will be major challenges for enterprise architects to overcome, but like the internal combustion engine, the Internet of Things will ultimately benefit not only enterprises but also the people in them, and society as a whole.