Why the Internet of Things Will Drive Data Integration
There are lots of numbers flying around about the Internet of Things (IoT). Gartner has IoT on 25 billion devices hitting $263 billion by 2020. Cisco sees it hitting $19 trillion by 2025, with an impact ten times the Internet itself, which is at least one of the ten most important inventions of our species. For my personal data points, I now own a connected thermostat, I wear a fit bit, and that fit bit is connected to a Bluetooth-enabled water bottle, to track my water consumption.
Device connectivity is really what IoT is all about. Thus, IoT is really a data integration problem, at the end of the day. Its secondary problem is data analytics.
Data-oriented technology has a vast array of clear benefits, such as the ability to have more energy efficient homes, better health, and even better and safer cars. But, the amount of data that will be transmitted from these devices is on track to exceed the amount of data that’s currently being transmitted from traditional enterprise applications. Will this emerging industry be prepared to manage this amount of data? Will the data integration approaches and technologies be able to keep up?
The problem is a lack of standards within the interfaces created by the IoT device manufactures. You don’t get data from two devices the same way, and thus those who are charged with bringing together the data from IoT devices need to speak to very different APIs (Application Programming Interfaces).
Of course, we always have abstraction in the form of data virtualization. Thus, we could look at each API using an interface translator that’s able to make various different interfaces appear as a single set of interfaces, using the least common denominator approach. While effective, these abstraction solutions quickly become complex, when considering that the number of interfaces will continue to grow, and thus the abstraction layer will become more diluted and overly complex.
The call to action for the device providers is to come together on a common set of interfaces that provide some sort of standard from device to device. This should be the objective in 2016, as we move from just standing up the devices, which has been the activity thus far, to allowing these devices to share data, which is the real end objective.
So, what’s a stake? One small example would be the ability for smart grids to turn down cooling and heating systems as needed to deal with spikes in power consumption, which could save millions of dollars, as well as huge amounts of carbon. Or, for clinical systems to be integrated with the millions of wearables out there, to determine patterns from gathered data that indicate a stroke or heart attack. Those are just two of the thousands of use cases that will be better known in 2016, and will drive a renewed interest in data integration by providers that may not have practiced it before.