Sometimes one wonders why buzzwords are buzzwords. What’s so exciting about Internet-plugged cameras ? Plenty, actually. The Internet of Things promises systems of interconnected and constantly communicating devices, which range from household cameras to industrial sensors. This short video from IBM Think Academy neatly explained how it all works:
The headline of this Forbes article summarized the significance and problem of IoT: “The Internet Of Things is about data, not things”. The growing number of devices will generate abundant, rich, and hard-to-access data for analysis and data mining, an attractive premise for all data-driven decision-makers.
Through my coursework in my Data Analytics class, I understand that multi-sourced data must go through the Extract, Load, and Transform process to form an analysis-ready whole. How would one implement the ETL process on data created in the IoT environment, where hundreds of devices from different companies have their own data and meta data formats ? Will this become a Tower of Babel situation, in which data travel only between devices of branded systems ? If a standard is to emerge and unite the IoT landscape, what would it be ? One way to keep up is perhaps to keep watch on standardization initiatives and forming alliances in the industry, possibly through those on this comprehensive list by Postscapes. I believe that eyeing the future winners one of the most practical things to do as a student, because it reduces one’s risk of becoming obsolete.
Unlocking the value of IoT also requires powerful infrastructure to handle gigantic amounts of data. Some companies find itself having to stop relying on SQL databases and traditional data warehouses and to adopt NoSQL and big data solution, such as Cassandra and Hadoop to serve their expanding needs. Rockwell Automation Inc, whose data previously stored in its cloud architecture are “in the single-digit petabytes”, added a Microsoft’s Hadoop distribution, named HDInsight, to aid in pulling data from isolated databases holding data generated from different devices. Considering these changes, I realize that my current knowledge in databases will not be so useful in the world where IoT reigns, unless I use it to explore and understand more advanced and powerful analytical tools and infrastructures.