Orginal article Infoworld
Big data has proven to have many uses, but one will surface in 2015: the support of the connected objects that constitute the Internet of things
With the Consumer Electronics Show in Las Vegas capturing a lion’s share of this week’s coverage in both IT and consumer media, it seems obvious that 2015 will be The Year of the Connected Object. Smart home appliances, wearables, digital health devices, robots, 3D printers, entertainment systems, car electronics, and other sensors and trackers of all sorts are setting the pace of the Internet of Things this year. Will Time Magazine go so far as to feature on their cover this year a Fitbit fitness tracker, a Roomba cleaning robot, or a Google self-driving car?
For example, a smart thermostat periodically measures room temperature, and reports this measure to its backend service (typically, cloud-based). The backend service combines the reported temperature, weather service data, information about the routine of family members (which is itself often collected by the thermostat’s presence sensor but also derived from online agendas, trip reservations, social media check-ins, etc.). It then determines the best temperature setting and instructs the thermostat to adjust the furnace or AC settings as needed.
The true value of the connected object does not reside in the object itself, but in the service it renders. This service is not rendered by the object (although actor objects can be the delivery mechanism) but by the backend service, and is based on the intelligence that is been built into the processing and analytics logic by the vendor. If we were to be literal about things, the so-called “smart thermostat” is nothing more than a “dumb thermostat” which connects to a smart cloud service…. Connected objects are just part of a large network of objects and services — the Internet of things.
Big data is therefore at the core of the services rendered by the Internet of things. The most recent advances in big data technologies have materialized in Hadoop 2.0 with YARN’s ability to support mixed workloads, Storm’s streaming data ingestion and Spark’s real-time processing capabilities. This ability of the Hadoop platform to become more real-time was critical to its use in the support of connected objects, and is set to enable any organization (with the proper programming skills) to fully leverage all the data it collects. For a provider of connected objects, that means rendering advanced and therefore high-value services based on the data collected by their objects, and (sometimes) executed by the same objects.