The coming Internet of Things (IoT) is quite possibly the biggest change that will impact embedded designers over the next decade.
I'm a principal hardware and firmware engineer at Alfa Instrumentos, Brasil, a company that makes industrial weighing systems. In my 25 years designing hardware and firmware and doing system integration for embedded and mission critical industrial systems, I've experienced enormous changes in technology, tools, processes, and even the very culture of work.
The coming Internet of Things (IoT) is quite possibly the biggest change that will impact engineers like me over the next decade as we evolve our standalone embedded systems into massive and powerful networks of devices that deliver unprecedented amounts of data over the Internet.
In reality, the IoT is a phenomenon, not really a single application. And just like the Internet evolved, it is happening gradually. In short, the IoT effect will "emerge" from what is now a relatively disconnected grid.
If I look to the grid of embedded devices that are now interconnected into discrete networks, like industrial control subnets, I see something similar to early networks of the sixties and seventies. These clusters of embedded devices are starting to be equipped with TCP/IP gateways, being driven mainly by connection to the factory IT back-end infrastructure. In this industrial application space, an already intensive M2M connectivity exists, centered in specific control-oriented applications.
I see the current increase in industrial TCP/IP based fieldbus protocols, like Ethernet/IP and EtherCAT, driving universal adoption of TCP/IP connectivity all the way to the sensors and actuator control networks. In a few years, the promise of "pervasive computing" will have reached those sensor networks, allowing sensor information to feed more general applications.
From the embedded devices themselves, little more than reliable transmission of their sensor information will be necessary. But from the grid perspective, a much larger "cloud of data points" will be available, making new applications possible that use that data and run in other nodes on the network.
Now, what is necessary to make this pervasiveness happen? One main requirement is the availability of gateways to these proprietary sub-networks from the larger Internet. This is happening at an accelerated pace, as wireless and wired TCP/IP points are added to these grids, communicating PLCs (Programmable Logic Controllers) with Ethernet and WiFi, and direct connections of end points to the TCP/IP grid, as is the case with the weighing instruments my company manufactures.
When this interconnect reaches a critical stage, a multi-modal array of sensors will be available to any IT application, web-enabled human interfacing device, or cloud-enabled application that will exploit this dynamic data extraction to offer unprecedented levels of sensory information.
New applications will emerge: For example, monitoring of micro-weather and seismic activity from data extracted from temperature, humidity, wind speed, and weighing sensors distributed in a number of industrial grids. And real-time monitoring of damages for emergency response team management during large natural disasters, or even war scenarios.
Looking at what is happening today and where, the industrial embedded space may well be the first space where the IoT will emerge. Are you ready for it?