Much has been written about the vast number and variety of things that will soon be connected to the Internet—from milk cartons and alarm clocks to sensors and trains. Already in 2008, that number exceeded the number of people on earth. By 2020, when the next incarnation of the Internet—aka the Internet of Things—is in full swing, the number is expected to reach 50 billion. And it’s not just things that will add value and relevance to networked connections, but also people, data and processes.
Think about it. Through their interactions with the Web, social networks and devices—especially mobile devices—people have a massive multiplier effect on the amount of IP traffic traversing the network. In 2012 alone, new, more powerful smartphone technologies combined with growth in both mobile bandwidth and apps produced annual mobile data traffic nearly 12 times greater than the total Internet traffic in 2000 (Cisco Mobile VNI 2013).
Add to that a coming tsunami of constantly streaming data as sensors in just about everything become the norm—not just wearable sensors attached to our bodies, clothes and shoes, but also sensors, meters and actuators in our cars, machinery and infrastructure. And let’s not forget the critical role that processes will play in managing and automating this explosive growth in connections as well as in the collection, analysis and communication of data. People, data, processes and things. Together, they will make up the next phase of the Internet of Things—the Internet of Everything.
Data in Motion vs. Data at Rest
Zooming in on data in the age of the Internet of Everything, there’s another critical distinction that needs to be made. You see, not all data is created equal. Most of the new data being generated today is real-time data that fits into a broad category called Data in Motion. This refers to the constant stream of sensor-generated data that defies traditional processes for capture, storage and analysis, and requires a fundamentally different approach.
Let’s back up a minute. Historically, in order to find gems of actionable insight, enterprises have tended to focus their analytics or business intelligence applications on data captured and stored using traditional relational data warehouses or “enterprise historian” technologies.
However, the limits of this approach have been tested by the increase in volume of this so-called Data at Rest. The challenges inherent in collecting, searching, sharing, analyzing and visualizing insights from these ever-expanding data sets have led to the development of massively parallel computing software running on tens, hundreds, or even thousands of servers. As innovative and adaptive as these Big Data technologies are, they still rely on historical data to find the proverbial needle in the haystack.
This rising tide of Data in Motion is not going to slow down. In fact, as the Internet of Everything gathers momentum, the vast number of connections will trigger a zettaflood of data, at an even more accelerated pace. While this new Data in Motion has huge potential, it also has a very limited shelf life. As such, its primary value lies in its being captured soon after it is created—in many cases, immediately after it is created.
For instance, real-time traffic information from cameras, sensors and connected cars allows drivers to avoid traffic jams and use suggested alternate routes, potentially reducing hours of unproductive time spent behind the wheel. Similarly, manufacturers can connect their stock inventory with their suppliers’ production systems so that potential delays can be identified as early as possible and corrective actions taken on their respective shop floors to better prioritize people’s activities. In each of these cases, it’s easy to see the added value of connecting not just things, but also people, data and processes.
The real challenge for data-driven organizations is how to manage and extract value from this constant stream of information, and turn it to competitive advantage. Data in Motion represents a new type of data whose value can not always be extracted through traditional analytics. In a next post, we will look at examples of Data in Motion and how to extract value from it.