Toward the end of last year, GE began promoting an old concept in a new package called the “Internet of Things.” This is also known as IoT, or, machine-to-machine (M2M), and various other names as we attempt to connect our devices into our networks, allowing them to share data. I covered this concept in my Data Integration Predictions for 2013, and still stand behind the emerging use of this technology.
The examples of the application of this technology are many. We can monitor entire manufacturing floors in terms of the state of all of the production equipment, and when they are likely to need maintenance, even when they are likely to fail. We can then leverage this data to further determine productivity numbers that consider the health of the equipment and devices along with the business outcomes.
Of course it does not stop with industrial applications. As we’ve seen in our own homes, many devices we’ve purchased in the last several years, such as TVs, refrigerators, and even slow cookers, are now connected devices. They can report problems to their manufacturer, automatically upgrade firmware, even communicate with you using e-mail or smart phone applications. They can also create data, which is stored remotely, typically in a cloud-based database, ready for analysis by the device owners or the manufacturer.
At a Minds and Machines 2012 conference held in November 2012, GE released the results of an economic study stating the potential economic benefit of moving toward the “Internet of Things.” GE now promotes this technology as having the potential to add trillions of dollars in value to the GDP (gross domestic product).
GE is establishing partnerships with large consulting organizations that can assist them in making sense of all of the data that these devices are spinning off. The battle is to be fought around data integration to the devices, as well as the analysis of the data to make sense of what the devices are actually saying to us.
Count on huge databases leveraged to store and analyze this data. Business intelligence concepts once focused on such use cases as determining the profitability of a product line. Now we leverage the same concepts and tools to determine the likelihood that the engine in our car will fail within the next 1,000 hours of use, and perhaps see the recommended corrective action that can be taken to prevent the failure. These are many of the same concepts around analysis of the data, just different data.
In a Control Engineering article entitled “Data integration is key for overall equipment effectiveness ratings,” we see that those charged with gathering data from connected devices already see a need to solve the data integration problem. “Often various industries need to pull together disparate information for reporting purposes. Integration with heterogeneous data sources is often complicated, confusing, or a multi-step process. Many cases involve combining business data with process data or gathering more details into a process summary report for certain conditions.”
The core objective here is to make the most of the data coming from these connected devices. A sound and well-planned data integration strategy is key. Moreover, it is critical to have the right data integration technology in place to move the data from place to place, as well as to deal with the differences in formats, protocols, and semantics.
There are some differences with connecting devices that produce data, versus connecting traditional business systems or databases. They constantly produce data at a steady rate, not in small bursts, as is typically the case with traditional business systems. The data is usually in a primitive state, and needs to be manipulated before the information is sent to a target system, or, more likely, a target database for storage.
Once stored in a database, the device data can be leveraged in any number of ways. This includes comparing it with business data, such as productivity information, to better determine the state of the business. For instance, looking at productivity and profitability numbers in relation to problems reported from manufacturing devices to determine if there is a positive or negative correlation.
Another use might be to automatically adjust the manufacturing processes based upon information coming from the devices on the manufacturing floor. For instance, you could schedule more or fewer workers if the devices are at certain levels of efficiency. The use cases are almost limitless.
The need to connect to devices is not just a passing fad; it’s the way things will be now, and into the future. The ability to effectively leverage this data requires a sound data integration strategy and data integration technology. There is a huge potential here, let’s make sure we keep our eyes this opportunity to make our business work better.