The Intelligent Internet of Things
Insightful intelligence on the performance of business and societal operations is obtained by processing and analysing IoT parameter and event data. The established ability to generate near real-time information enables informed decisions to be made at the local level. The relatively recent evolution from information to intelligence elevates IoT’s functionality by enabling various repetitive, intelligent tasks to be automated. In healthcare, for example, they include image-recognition tasks such as detecting breast cancer from mammograms with remarkable accuracy 30 times faster than a radiologist. And in industrial environments machine learning enables visual quality control inspection of identical items being transported on conveyor belts at high speed. So far so good.
IoT solutions have been deployed across the business landscape: from manufacturing to logistics, energy, health and smart cities and multiple other vertical industries. They are known as silo solutions because normally they cannot communicate or interoperate with each other. They are therefore managed separately. This does not equate with one of the key benefits of insightful intelligence, the ability to share intelligence. However, combining those sources and intelligent workflows in order to extract valuable information is problematic. Silo systems have different connectivity, security, data storage, system integration, device hardware, application development, and other requirements.
The intelligent infrastructure
The Intelligent infrastructure, which is a relatively new term, addresses those issues and constraints. There is no agreed definition, but it will be enabled by intelligent platforms that can monitor hundreds or thousands of feeds, on premise or in the cloud. They will make extensive use of edge computing and employ machine learning, artificial intelligence, and streaming analytics as well as runtime software tools that monitor, alert and support interactive decision making. In addition support for enterprise resource planning, customer relationship management and manufacturing execution system will be required. In a nutshell, most everything will be integrated, from the manufacturing shop floor to the executive offices.
Between then and now
Gartner predicts that it might be 5-10 years before mainstream adoption of Intelligent Platforms. Right now innovative vendors are addressing the issue by adding infrastructure machine learning, analytics, and artificial intelligence for IT operations as software overlays. A key objective is to create a platform that combines diverse data sources and intelligent workflows.
Meanwhile, fuelled by 5G and despite economic uncertainty, IoT deployments continue to grow at over 20% per annum. When fully operational, 5G networks will have the capacity to connect 500 times more devices than 4G. This is the foundation for the future of “Massive IoT” — a world with a million or more connected devices per square kilometre. A lot of the data will run at low rates, e.g. NB-IoT/LTE-M, and operate over LoRa, Wi-Fi and satellite networks. In addition different connectivity types will be employed for collecting data cost-effectively.