We have recently been seeing a lot of interest in the “Internet of Data”. We hear a lot about the Internet of Things, but what is the Internet of Data? I really see these two things as the same, but simply referring to different parts of the network. When people are talking about IoT or Industrial IoT a lot of times they are focusing on automation. That could be from intelligent devices, smart home sensors, weather sensors, autonomous vehicles, blockchain etc. This is really an emphasis on edge computing.
When people talk about the “Internet of Data”, what they are referring to is the collection of data from edge devices and performing deep analysis of that data to gain insights. The easiest way in my mind to look at the Internet of Things vs the Internet of Data, is big decisions vs small decisions.
Currently, a lot of IoT is focused on making small decisions. In the consumer world that manifests itself as turning my fireplace on when my SmartThings hub detects that I am home or collecting my steps from my Fitbit and displaying them on my phone or similar scenarios. In the Industrial world that presents itself as adjusting crop watering based on weather data or sending alerts when food in transit reaches dangerous temperatures or many many other use cases which are all incredibly valuable and interesting.
That said, things get even more interesting when we start to make big decisions using the data generated on the edge. This really comes to play when you begin to marry technologies like AI and predictive analytics with the Internet of Things. This intersection is where the “Internet of Data” really comes to life. Imagine if your Fitbit not only collected your steps and heart rate, but could also alert you to potentially life threatening conditions in real-time. This could be achieved by using machine learning to look at your historical vital signs like heart rate, perspiration, oxygen level, and comparing them with your current metrics and looking for patterns that could signal health conditions. This level of real-time analysis and data science could be potentially life saving.
We talk a lot about the data value chain. We see this as ingestion, analysis, and reaction. In order to see true ROI from the massive amounts of data that companies are collecting these three steps need to be executed. In typical Internet of Things use cases, ingestion and reaction are being performed but no real deep analysis. Part of the reason behind this lack of analysis is that intelligent devices are limited by their hardware, internet connection, and underlying software. Companies are focusing heavily on SaS based IoT platforms that require edge devices to migrate their data to the cloud. Cloud based services are awesome for a lot of use cases. That said, they can be expensive, and getting data to them can cause a real bottleneck due to internet connectivity and throughput. These solutions will work great for historical analysis; however, they will never be able to deliver on the true promise of the Internet of Things which is to solve problems in real-time using edge devices, machine learning and AI. That’s where the react portion of the data value chain comes into play.
We believe that to truly enable the “Internet of Data” machine learning and AI processing needs to be moved directly to the edge. In order to do this companies need to look for solutions that can handle the entire data value chain directly on the edge and do not create throughput bottlenecks, but use a more democratized architecture. Infrastructure that can handle similar workloads to existing infrastructure, but with a much smaller footprint will be required.
It's exciting to watch pioneering companies begin to adopt these patterns and think about solving big problems with IoT. Some of the more forward thinking companies today are solving problems by pushing the data value chain directly to the edge in logistics, healthcare, manufacturing and many other industries. Our belief is that as more and more of this compute is pushed towards the edge, the Internet of Data will become more of a reality. Companies in the industrial and consumer space will be able to provide real-time solutions to problems today while experiencing dramatic cost savings by leveraging their existing capital investments for compute.