IoT: are we grasping it or groping at it?




Some time ago I was passively listening to a discussion at my office about making our office building "smart" as part of the IoT initiative. It was an interesting discussion. People were discussing everything about data - what kind of data will be available, how much, the database used to store it, data mining, finding out patterns and trends etc. etc. However, none of them was interested or even curious about knowing or discussing about the "things"!

Well, it was of little surprise to me as these days this can be seen almost everywhere. The use of acronym "IoT" has become very fashionable. Almost every company wants their product to be "smart" and "IoT ready" in someway or other. But the question is do we really understand or want to understand what IoT really means or are we just trying to rip benefits from the current hype about IoT?

IoT is internet of "things" like sensors, actuators etc. So the hardware is an inherent part of it. An IoT solution has both hardware and software parts. Only having a good sensor network is not sufficient, at the same time just thinking about data mining does not make a product an IoT product. It is an end-to-end system - not just sensors or not just data mining. So what is the current scenario most commonplace with so called IoT products? If we scratch just below the surface we will most likely find a data mining and/or visualization software. However it is important to understand that only having the big data part is no different but just like another credit card history data.

These days everybody seems to be interested in mining data, offer analysis, trend research etc. and there is nothing wrong in it. However, marketing a data analysis product as "IoT" product is the unethical. Most of the offerings are nothing but data analysis and presentation tools working at various speeds. There is nothing special with respect to IoT. Unfortunately many so called IoT product designers appear to be thinking about data only after it is delivered to them. There is no thought about gathering it leave aside controlling "things", rather there is general ignorance about it - "That is somebody else's job to deliver data to us {and our software is "smart" after that}!" Well, that's definitely hilarious! All these software products are just data analysis packages, nothing more and definitely not IoT products.

Then what could be a true IoT product? I think a true IoT product is the one which truly interacts with the "things" or at least offers well defined ways to interact with them. This implies bi-directional communication, control and security. This also implies scalability. It is also true that having these capabilities does not make it a true IoT product. Having capability of interacting with thousands of sensors does not make a system "smart".  Having a lot of data does not mean getting true trends.

The converse is also equally true - just having a sensor/controller network and support does not make it a complete IoT product. It's just a network, not benefit can be ripped unless the data it generates is consumed for further analysis. So, here comes the other half - the data analysis part.

If a product offers provides both these halves, it can be a true IoT solution. Even if a single half is being offered, as long as it is offering an easy integration and wide options of integration with the other half, the solution has a potential of being an IoT solution. But, I feel if somebody is trying to offer just a single half branding it as "IoT" offering, it really can not be called it that way!

Comments

Popular posts from this blog

Security in Linux Kernel - Part 2

Security in Linux Kernel - Part 3

Trusted Platform Module