Is the IoT world ready for sharing?
Today I noticed a nice overview of challenges for data processing of IoT data. Mostly based upon IEEE work it highlights the need for fog to support the cloud. And as we usually are concerned about the processing side of things, as this has long been the bottleneck to get the kind of AI and Deep Learning results we gradually are getting familiar with, we tend to overlook the complexities of data or more precisely information.
There are ample examples telling us how much data we have produced already and how fast the data creation will grow going forward. Good news for the data storage companies, because every bit of data needs to be stored on a physical device, whether it is in the cloud or not. And what is worse, data is not stored once, it is duplicated many times. Whether it for safety reasons (backup), or performance reasons (bring the data to the more performing CPU’s to process) or sharing reasons (how many copies in average of a single email exist?), data is duplicated many times. This only adds to the data management problems. Fog computing will help to solve the processing of vast amounts of data, but how will it impact the data management problems? And more specifically how will it impact the ability for different IoT systems to exchange data?
So far most of the IoT solutions are independent from each other. All data collected in the system is processed, analyzed and controlled within the system. A self driving car, as an example, could be considered as a single IoT system, data being collected from cameras, LIDAR and other sources, processed in the car (and cloud) and used to maneuver the car safely through the traffic. Whenever it will be there, not a small accomplishment.
Yet, so far there is no way in which cars can talk to each other. There is a development around V2X, but definitely not ready for prime time (a great video on the state of self driving cars can be found in the video: “16 questions about self driving cars”, highly recommended). Why? Not because it is not useful or the wireless communication technology is not ready.
The reason why it is so difficult to share data (even though some legal opinion might be to rather not share anything at all: Do You Really Need to Store That IoT Data? ), is because data needs to be in a context. It has to be transformed to information.
Just think of it. We as humans have developed language to exchange information, so the data we exchange can be put in a context to have the same meaning for everyone. It seems only logical that our IoT inventions should do the same. It is the information that is meaningful to be acted upon. Not the underlying data that led to creating the information.
Maybe an opportunity for linguists to add their knowledge to the IoT frenzy…