< Articles & News

5 Things You May Not Know About The Internet Of Things

February 20, 2015   |   1:48 PM
Teradata Articles

The Internet of Things (IoT) is making headlines these days. Here are five things that you may not know about the IoT and its application to business processes.

  1. It’s Already Here – Although it is often framed as an emerging trend, the IoT is not a future prospect. For many, the IoT is operational. For example, many companies have sensors on their equipment that allow them to do condition-based maintenance. Rather than following maintenance schedules based on averages, they use sensor data to predict in advance when repairs are needed. Other companies are using the IoT right now to increase revenues, improve product quality, and reduce safety incidents. From solar-powered recycling stations in Times Square that alert staff when receptacles need to be emptied to smart electricity meters enabling new pricing plans, the IoT is transforming the way companies do business.
  2. Security Is Paramount – Here’s a chance to do something right the first time. IoT data comes from sensors: thousands of sensors. Unlike embedded computers at $20-$30 each, sensors may cost as little as 10 cents. They are simple devices, generating a heartbeat of data. Given the world events hackers have perpetrated, make sure your IoT implementation bakes security into your sensor network from the beginning.
  3. IoT Data Is Not All That Big – The tsunami of data associated with the IoT is a bit overblown. One frequently cited statistic is that every time an airplane takes off, one hour of flight generates at least a terabyte of data. The reality is much more modest. When an airplane pulls into the gate, ground crews connect data cables and download a few hundred megabytes or a gigabyte of data for the entire flight. The myth of a terabyte an hour came from the R&D processes of testing airline engines, not typical aircraft in flight. Not only is the size of IoT data sometimes overstated, but the amount of data you need to keep is often overstated as well. If a sensor emits the same reading every minute for 30 days, you don’t need to keep every data sample. Put a time stamp on the first reading and only capture the changes. Also consider not capturing new readings unless they cross some threshold. Then summarize and compress that data. 100-to-1 or 1000-to-1 reduction in data size is often achieved.  This does not apply to all sensor data. Some readings vary widely with humidity or barometric pressure. But a significant proportion of sensor data can be stored in a summarized or compressed format without losing any information.
  4. Sensor Data Has More Than One Life – Analysis of sensor data is where the big return on investment emerges. So far I’ve referred only to the most obvious use: what the sensor tells you about the equipment on which it is installed. This is low-hanging fruit, and it is important for applications like condition-based maintenance. The next level of sensor data comes from correlating data across multiple sensors at once. Imagine 20 sensors all approaching danger scores at once but none actually cross the line. Something important is happening: pay attention. An even broader look, and where amazing ROI can be captured, requires combining sensor data with financial data, warranties, profitability, labor planning, maintenance costs, and other information. When you combine sensor data with existing data, you can gain fresh insights across the organization and make changes that have real business impact.
  5. There’s No Need to Wait – Many IoT implementations are already underway—by your competitors. It’s a good time to consider a proof of concept that uses your existing infrastructure. Many organizations choose to collect sensor data in a data lake, using Hadoop as a staging area to decide what and where to incorporate sensor data into other processes.

The IoT is an exciting area of innovation, one every organization (and every person) should explore. The data from the IoT will enable many more processes to become data-driven and drive new types of transformation.