Header Ads Widget

What is Internet of Things?

Internet of Things or IoT has been a buzz word since last almost one decade. Lot of manufacturers have released their products with IoT integrated inside. What is Internet of Things? Is IoT a new technology? Is IoT a new concept? Let us see in detail. 

Internet of Things means connecting anything and everything around you to the internet, for collecting its information, using the information for some useful purpose and sending some command back to connected thing to perform some action. The word "Things" means any thing or object around us. It can be your car, lighting bulb, mechanical tools, any other vehicle, fan, motor, TV, fridge, your microwave, air conditioners, machines in factory, products in shopping malls etc. Any thing you see around you can be connected to internet using small electronic circuitry. Even we wear smart watches connected to our mobile phones and the phone is connected to internet collecting our vitals from smart watch and sending over internet to main server, where others can see the recorded information. Till these things around us are not connected over internet their status is unknown and such things are termed as "Dark Assets". As we don't know their current state.

If we look back in history, the first such connected machine was discussed in 1982 where a Coca Cola vending machine at Carnegie Mellon University was connected to internet to collect the inventory and to check if newly loaded drinks were cold or not. The term "Internet of Things" was coined by Kevin Ashton from Procter & Gamble in 1999 when he prefer to call it "Internet for Things". Since long time we have been using RFID cards for recording our presence in office and marking attendance, that is nothing but an implementation of IoT. We place our RFID card near a reader while entering any facility, it reads the identity of card which is linked with our employee ID, hence complete data of our movement with timing details are available with management/HR for review and corresponding action. If we don't have this facility, we may not know the current position of employee and hence will be called a dark asset as his/her location is unknown. 

Many companies already have automated process in their factory floor, like PLC (Programmable Logic Controller). They have multiple sensors connected over RS232 bus protocol to collect inputs from field and display at a central dash board. At the same time there are many actuator which can be given command from central dash board and they perform the action on ground somewhere at remote location in factory work shop. Actuators are devices which can take command and perform some actions like electronic locks, electric pumps etc. Again this is an implementation of IoT, which is now termed as "IIoT" i.e. "Industrial IoT".

So its not a very new technology, rather implemented and marketed more now a days now it is being implemented everywhere, not limited to industries. The thought process is that even every lighting bulb will be fitted with IoT chip which can keep sending information of its usage to the manufacturer and hence it will help for marketing purpose because it will know when the bulb has gone defective. Today we can see Philips and many other companies have come with LED bulbs that can connect over internet and can be controlled remotely from anywhere over internet.

Below factors are now helping the implementation of IoT more easily and at lower cost.

1- Technology and so the manufacturing process has enhanced and we have low powered smaller chips performing powerful actions hence can be embedded in any device easily. This helps small battery powered devices to operate for years without any maintenance. Even battery technology has also improved a lot from bulky batteries to small and powerful batteries which can keep supplying power for longer period. Even the solar panels have enhanced and are readily available to provide power on remote locations.

2- High speed and cheap internet and its availability in remote locations. Now the 5G technology will further enhance the implementation of IoT. Few decades ago, the reach of internet was not so easy in remote locations.

3- Earlier with IPv4 there was limitation of number of unique IP address to be provided to all devices or objects on the globe. Now with IPv6 we have over come even that problem. IPv4 uses 32 bit address hence it can be allocated to 2^32 different objects which comes to be around 4.29 billion. IPv6 uses 128 bits which means 1028 times more than IPv4. It is said that even if we connect all devices on earth on internet and give unique IP address to each device using IPv6 format, still we will be left with many unused IPs to connect more devices.

Internet of Things

Here we can see a simple diagram to understand how an IoT setup looks like. On the left we have sensor and an actuator connected to a microcontroller. Presence of actuator is not must if we only want to collect data with the help of sensor and analyze for some experimental purpose or marketing purpose. Small sensors are used to sense environmental conditions or object position, color, flow, pressure etc. and are sent to microcontroller which is connected to internet to send the collected data over the internet to servers lying somewhere in remote data centers. These data are collected periodically, may be every second, or at an interval of few minutes or hours or days. This totally depends on criticality of the data and how it is being used. Like if we are collecting weather report (temperature and humidity) of any location, every second data is not critical, it can be hourly data also, but if any doctor is remotely monitoring the health status of a patient, every second's data is important.

So based on type of sensor connected to IoT device, respective data is collected and is sent to central server. These data which are sent at regular intervals are very smaller in size which is purposely designed to be small so that power consumption is less on the IoT device. Hence the data coming out of each IoT device is said to be small data. But imagine if all devices across globe are connected and and they start sending their small data over internet, the size of data reaching to servers will be very huge, may be some Terabytes in few minutes. So the data on the server side is called "Big Data". Big data is not only defined by its size, but there are other factors also which we will see in this article further. 

Now on the other side of world some where we can have a connected mobile device or a computer where we can see the data collected from the IoT devices, analyze it and take some action near IoT device. When we say action required near IoT device, then the actuator comes in picture. The action can be in form of action or even recommendations. Like if a light bulb has to be turned on over internet when it is dark around it, we send command from mobile or PC to turn on the light bulb. So here we have sensed the light condition using light senor and activated the bulb accordingly. In case if a doctor is monitoring health condition of a patient remotely over IoT, based on measured parameters doctor will send recommendations to the patient in form of medication or precaution etc.

There are many small implementation to realize IoT, one such small design by me was published in Electronics for You in 2019, click here to checkout the same.

IoT Protocols: 

Being low powered device, aim is to reduce the data load on IoT devices so that battery operated devices shall run for longer time. To achieve this, below data communication protocols are used.

MQTT: Message Queuing Telemetry Transport uses "Publish" and "Subscribe" method of communication. This is also used by Facebook messenger. In simple language when we subscribe to some channel in YouTube, we don't keep checking the channel for new updates, it will waste data unnecessarily. Rather if any new update is released, we get notification and then we connect YouTube and watch it. Same technique is used in MQTT protocol hence unnecessary data flow is minimized. This is most popular among other protocols.

CoAP: Constrained Application Protocol is named like this as the controller in IoT device works in constrained environment means it shall use less power, less data transfer and smaller program code on it so that execution of codes doesn't take much power.

6LoWPAN: IPv6 over Low power Wireless Personal Area Network is another protocol designed to to communicate with these IoT devices over personal area network. This ensures even very low power devices with limited processing capability should also be part of IoT ecosystem.

Big Data
We came across a the term "Big Data" that gets accumulated and processes at the Data Center at the back end which is received from all IoT devices. It is not characterized only by the size of data, but there are other features also which characterize a data in Big Data category. Big Data is defined by four V's that are "Volume", "Velocity", "Verity" and "Veracity". 

Volume: As the name suggest, the size or volume of data must be huge which is the first most important property of Big Data. If we say huge size means in the terms of Petabyte (PB) or Exabyte(EB) or even Zettabyte(ZB) also. For reference - 
1 PB = 1000 TB (Terabyte)
1 EB = 1000 PB 
1 ZB = 1000 EB

Velocity: The other characteristics of Big Data is the Velocity at which the data are transmitted, processed and re-transmitted is huge. Imagine if every light bulb on the earth is sending its data over internet at same time what will be the data traffic at the data center level. At the same time other equipment connected over internet will also be communicating over internet adding more data at same time.

Verity: Data at back end from different type of device will be reaching at back end and same has to be processed in time. Every devices in globe will be sending its own data, light bulb will send the hours of usage and current and voltage, AC will send data about room temperature, vehicle may be sending data of amount of fuel left, speed at which it is being driven, location, wear and tear of engine parts, kilo meters left for next servicing and so on. So there are N numbers of different verity of data expected from different connected devices. These are called "Unstructured Data" as they are not in any fixed format to fit in any table of content like "Name", "Age", "Sex", "DOB", "Address" etc. MS Access or SQL are the software which can work on structured data as mentioned here like "Name", "Age" etc. But SQL cannot handle the unstructured data received from such verity of devices.

Veracity: Next comes the uncertainty of which type or data and what volume of data will come from which region of globe and at what time. None of these attributes of data are predictable and such unpredictable volume, verity and velocity has to be handled for efficient implementation of Internet of Things. For handling such unstructured database one such tools is "Apache Hadoop" which is an open source framework used for distributed storage and processing of datasets of Big Data.

Post a Comment