Internet of things: understanding and realizing the next big industry


A multi-billion dollar industry, bigger than the online advertising one (primary revenue model for giants like Facebook and Google), is the Internet of Things (IoT). If you don’t know what exactly Internet of Things is, here is how Wikipedia explains it, “The Internet of Things is the interconnection of uniquely identifiable embedded computing devices within the existing Internet infrastructure.” While most of the internet what we see today is just a network of computers, tablets or mobiles, Internet of Things aims to make everything (your air conditioner, toaster, washing machine and almost everything) a part of internet so that we can control these devices effectively, automate them and help them to connect among themselves to make our lives more easier.

Internet of Things is the most promising (may be hyped) industry at this time; around 20 billion + devices are expected to be a part of this network by 2018. The IPV6 addressing scheme has already made sure that there is room for every device to be added to this giant network. IDC has suggested that Global IOT market will evolve to $ 7.1 trillion by 2020, that’s greatest value added by any industry so quickly.

The main things to get connected to IOT are smart cars, consumer electronics, smart meters, and home-security systems etc. But I haven’t written this post to make you realize the potential or hype about the same. I’ve just devised very simple practical implementation architecture for Internet of Things to realize its full potential and help every part of economy to benefit from this boom.

At present, almost all the IOT startups are approaching this opportunity in basically the same way; they choose a hardware, make use of computer boards like Arduino or Raspberry PIE etc. and create very individualistic solutions for the problem. These types of solutions lack the scale, and don’t contribute in building any global ecosystem for the IOT. Internet is able to evolve itself so quickly due to the well accepted/implemented set of standard protocols, commercial viability, global architecture standards, availability of network devices with universal use cases (Ethernet card etc.), and an amazing combination of centralized and decentralized control (the way root DNS are maintained by multiple authorities). Thus the major contributors for internet revolution and its wide acceptance were Global Protocols, Universal Network Devices, Central DNS control, Commercial Motivation and Room for humans to innovate.

Now it is well understood that to achieve the full potential of the network of things we have to think beyond individual devices, beyond individual use cases and obviously beyond the walls of our startups. Here is my proposed architecture for the future of Internet of Things. This is in no-way a complete one but is definitely valuable as it can serve as a foundation for understanding IOT and building a universal layer of IOT.

This architecture would contain four components, each having some functions and one part of the economy (stakeholder) would lead its development.

Things: These are electronic devices which will form nodes in this network, for example a toaster can be a device that needs the ability to be controlled remotely by some other device on the network. In a scene where a universal homogeneous architecture is required, we will rely on device manufactures to communicate the state and controls of devices through a standard port (physical). For example, a toaster will have states like temperature and power, the port in this case should be able to send this information in the form of attributes and values, and should accept any change in those attributes and change the state of device accordingly. (Example, a television remote control sensor, which tells the present channel, volume etc. and lets us change the channel and volume as well.)

Thus ports will be universal points in the devices to accept, read and change information. At present, hardware guys use computer boards to build case specific solutions, which will become too cost ineffective in the future as this thing reaches scale. Major stakeholders will be device manufacturers or any B2B company which will help manufacturers in realizing those ports.

Network Interface Device: This will be the second component and would be analogous to present day Ethernet cards. Their function will be to act as an interface between the network and the device through the port. They will be connected to the ports on one side and the network on the other. They can be wifi enabled, or cabled devices with the only capability to pass PUT and GET requests from the server (our third component) to the device. The server would send instructions in a predefined format (analogous to JSON) to the NID and NID would pass them in a machine understandable format to the device through ports. These devices would be similar, specific in functionalities and would be required in large amount; any company manufacturing these can easily achieve economies of scale. Just the way CISCO is a master of Network Infra now. Device manufactures may also choose to ship the product along with this NID already embedded. Major stakeholders would be network infrastructure manufacturers.

The IOT Server: This is a network management layer above the IOT. The control and state requests will be made to the server and after authenticating users and devices this will transfer PUT (change state) and GET (get state) request to the NIDs and GET results from the NIDs. There will be a public API for this set up. The server will contain information about devices and all data will be stored here. Information will be provided by device manufactures like PUT requests possible, state requests possible, and device details etc. Consider it as a Facebook application server, where applications can read user data and modify it after being authenticated and can also communicate among themselves. Major stakeholders would be a joint organization which would enable this IOT Server like the ICANN.

The API: The IOT Server will have a public API used to talk to the devices. The developers will try to different recipes and create innovative applications to help us live an easier life. The API can make two or more devices talk with each other and with humans via simple PUT and GET request. These applications will be the new universal remote control interfaces. A public API like this will help create an ecosystem of developers just like Android and IOS have created over time. But nonetheless there is way more opportunities and room for innovative applications in IOT than in both of these stacks combined. Major stakeholder would be startups, individual developers and other businesses.

This is my first take on Internet of Things and is far from perfect. I may have missed some of the fundamentals here but I am very open to any discussions. Please don’t forget to leave your comments below and I would be happy to connect.


About the Author :

Ankit Shukla is the President at E-Cell ABV IIITM.



Updates from around the world