Fundamentals of IoT Systems | Gaurav Awasthi | Skillshare
Play Speed
  • 0.5x
  • 1x (Normal)
  • 1.25x
  • 1.5x
  • 2x
14 Lessons (1h 3m)
    • 1. Introduction

    • 2. Significance of IoT

    • 3. What is IoT

    • 4. Why IoT

    • 5. Data Transmission in IoT

    • 6. Primary Building Blocks

    • 7. Technology Makeup

    • 8. Case Study 1 Connected Health Care

    • 9. Case Study 2 Asset Monitoring

    • 10. IoT Data Lifecycle

    • 11. IoT Architecture Ingestion and Processing

    • 12. IoT Architecture Analytics, Visualization and Integration

    • 13. IoT Architecture Device Management

    • 14. Conclusion

  • --
  • Beginner level
  • Intermediate level
  • Advanced level
  • All levels
  • Beg/Int level
  • Int/Adv level

Community Generated

The level is determined by a majority opinion of students who have reviewed this class. The teacher's recommendation is shown until at least 5 student responses are collected.





About This Class


This course on "Fundamentals of IoT Systems' is intended to introduce and solidify your understanding on all aspects of Enterprise IoT systems

This course is NOT about programming or coding of embedded devices where you connect sensors to RaspberryPi or Arduino boards and push data to cloud to display it in graphs. This is course is about Enterprise IoT Systems, its core elements and their interplay

You would love this course if you are interested in understanding how Smart City, Smart Retail and Connected Healthcare use cases are designed and implemented. At the end of this course, you should be able to get an understanding of the end to end architecture of all IoT Systems.

It includes -

  • Core IoT Elements

  • IoT Data Flow

  • Technology makeup

  • Building blocks

  • IoT Phases 

  • IoT Reference Architecture.

This course also covers and Two Case Studies, one on Connected Healthcare and the other on Asset Monitoring

There is a also a Project Assignment as part of this course, for which you have to create an Architecture for a given IoT problem. The idea is to apply the knowledge acquired in this course to a real situation. The project submission would be evaluated and review comments would be provided to you.

Finally, please feel free to message me any doubts or question that you may have and want clarifications for better understanding.

Good luck, as you onboard yourself into this inspiring world of IoT!

Meet Your Teacher

Teacher Profile Image

Gaurav Awasthi

Enterprise IoT Architect and Consultant


Hello, I'm Gaurav.

See full profile

Class Ratings

Expectations Met?
  • Exceeded!
  • Yes
  • Somewhat
  • Not really
Reviews Archive

In October 2018, we updated our review system to improve the way we collect feedback. Below are the reviews written before that update.

Your creative journey starts here.

  • Unlimited access to every class
  • Supportive online creative community
  • Learn offline with Skillshare’s app

Why Join Skillshare?

Take award-winning Skillshare Original Classes

Each class has short lessons, hands-on projects

Your membership supports Skillshare teachers

Learn From Anywhere

Take classes on the go with the Skillshare app. Stream or download to watch on the plane, the subway, or wherever you learn best.



1. Introduction: Hi, I am got above St. and I welcome you to this course called fundamentals of IoT systems. Internet of Things is the latest and greatest technology trend in the world today. Millions of people have invested their diets, energy, space, and money to get to learn IoT and to be part of this technological revolution. Personally, I am very excited about diarchy. I'm the prospect of injecting life into everyday objects. Sense putting sensors into the object, taking data out of them, and analyzing the data to know that state and behavior of these objects and even to send commands back to the devices to change their state. Incredibly interesting. My journey into IoT started around six years back when I started designing, architecting different IoT projects. And I also started consulting various organizations on IoT. Before that, I'd been working in software systems for around 14 years. But I think we have quite a smell. Technological Nirvana started when I started working in Africa. I've never looked back. And I think these experiences, skills, and knowledge that I have gathered so far in the field of IRD is what I want to share it with you all through this course. I would like to make a point that this course is not an implementation course. It doesn't involve coding, it doesn't involve implementing projects like connecting temperature sensors, do raspberry pie board, Arduino boards, writing some code into it and sending that data over the network to a cloud based service and then show it on some kind of a graph. Down BYOD is not the idea of this course. This course is fundamentally different from this idea and dual is one. We're talking about enterprise grade IoT systems here, which are very different. They are disordered, bigger problems. They, they added a different scale. So they're, they're inherently different from DIY projects. Secondly, we're also talking about the fundamental IoT knowledge here. I believe before we start implementing the project and there's nothing wrong with it. But before we delve into implementing projects, we need to understand fundamentally what our IoT systems. We need to understand all the technology, technological details of these projects, of the systems. And we also need to know, then we start implementing peer. Does it actually lie in the whole landscape? So a very thorough knowledge is what is expected. Implementing the technology. And discourse is about imparting that. Further, what all scores involves r. V would be discussing the essential elements of IoT. We will discuss what are the different kind of data flows that go through IoT. Black farmers. Components and services will also see what are the different technologies that make up an IoT system. Viewed also go through a couple of case studies to see real life demonstration of how IoT architectures solve real problems. Finally, this would culminate into an odd reference, IoT architect, which is an end-to-end architecture with all the bells and whistles. It covers all the different layers of IOT from ingestion to do visualizations. And we will understand different elements about what are the different variations of these architectures happened when we started from problems. So all in all, I would like to reiterate that this course is a fundamental course in do a complete IOT landscape. So welcome to this course, and good luck with learning. Onboarding yourself on this learning pot fire D. Let's get started. 2. Significance of IoT: Internet of things has the power to change the world. In fact, it is one of the most significant developments of 21st century. Well, these doll claims, but it is IoT is proving itself to be that value, but I wouldn't say it is. It is the most significant inventions of our time because IoT has been there serving its purpose across some of the industries. But the invention is in terms of the democratization of this technology. In the sense that a lot of technologies that make up IoT, like cloud computing, mobile applications, big data and analytics. Iot in that town and in that sense, has been a convergence of all these technologies which has brought itself so valuable that it is in the forefront of technology landscape today. And its adoption is across industries. It is not just an artist manufacturing, or health care or travel or government, even energy and utility companies. Almost all of the companies have started investing in IoT solutions because it is bringing back that kind of a value. It is fulfilling their business objectives. It is increasing operational efficiency. It is making energy mode aligned. It is, it is providing business value to organizations. But what we need to understand this behind all these doll claims, the technology behind it is very simple. We need to understand what is IoT? How does it work? What are its core essential elements? How do they inbuilt play to provide these many solutions, from very simple variable, mobile lab driven devices, to industrial automations, to massive devices do device twins, how does IoT platform or IoT systems provide that much value to a grass across these many industries. Let's see that. 3. What is IoT: Internet of Things is defined as the network of interconnected objects and devices that are able to communicate, and therefore, they are able to bind to transfer data over the network. The definition includes many elements that keep on discussing vertex trying to simplify this definition into simpler structure. So all in all, intended octane is basically about making Dante smart. Then I'd say smart. Smart means conversationally. When I say DOM, it means it is not able to speak, it is not able to communicate. So devices that are not able to go into a game make conversation by using Internet of Things. Internet of things in that sense, because not a single technology. It's a group of technologies that are many technologies that make the whole IoT ecosystem possible. Further, if, if I'm talking about conversation and it also means we are able to speak of the devices. Devices on April l is what is the problem? Where does their stage where does their condition? We are able to send commands to the devices through D2, then order instruct them to behave in a sodium. And therefore, we are able to make sense of what their devices are. Being. Able to understand the condition in which they're operating. Are the devices that behave in a certain way. You are able to collect that data and analyze their data and then make inferences are. So basically it is about talking with the devices in a sense that we are able to understand them. They are able to understand on intentions and being able to strike that kind of conversation. Betty simply is nothing more wired than making devices conversation. 4. Why IoT: We have seen what is IoT. We learned a bit about the definitions and the fact that it consists of devices and an IoT platform. They communicate with each other to build an IoT solution. At least at that level, we know where desired. We also know that IoT has seen a massive adoption across business organizations, research labs, even universities. Given all of that, we still need to ask a very basic question and that is vitally needed. What are the problems that I will begin? Solid? I believe what any technology would be worthwhile to have darker and kind of adoption. It must be able to solve some problems that we cannot solve. Otherwise. It has a, it has to solve some real life problems next year. Some of them here in this lecture. Let's go through the first domain. The domain is called connected health care. And clinical health care has many use cases. The one that we're talking about here is remote monitoring of patients. So what happens is our doc or a health care worker can remotely monitored vitals. Are the nutrient levels like vitamin B 12. Or it can even operate remote medical equipment through an IoT system remotely. It is one of the essential functions. I would say that the patients do not need to travel to a hospital or clinic if they are not able to do so. And a health worker, a doctor can remotely monitor them and make adjustments to their to their doors, are made configurations to the equipment that they're using a code, our variable, they're using it remotely. So it's an essential functions. So to say. Let's see another use case. Now. This is one of the very interesting use cases that I know because I worked on a project within a village in Africa. The villages were using charcoal, stove to cook food and day in and day out every day. And we know it is not a healthy way of cooking food. So we devise greed for them to use ethanol based stores. Now the interesting part or the IOP part of the story is that this economic and also be consumed as an alcohol. So we have to ration the supply of ethanol or cooking purposes. So this company built kiosks to dispense ethanol. And, and everyone, every household was given a can, again, which could be filled with ethanol every day. Now, when a person goes to the chaos to fill that, there could be a handshake between both of them. And the chaos could identify that canister using NFC. And I'm dispense only a limited amount of ethanol every two days so that it can be used just for cooking the food. And an extra floor would not be allowed because again, the canister has to be authenticated and approved by the kiosk. So the whole solution was built using an IoT system. And this is, I think a very interesting use cases providing clean energy through the residents of a village. The toward US kids ever talk about is a waste management one. So the garbage collection trucks come to the localities every week or every day at some places. Now, does if there are bins that are that can talk themselves are in the garbage truck, that they're either not filled at all or half or depending on the domain, dependent on the US kids are the requirements. Iot system in the middle can decide can dynamically decide the route of the garbage truck and date is on the bins. So that these pins can have d level sensors and they can transmit the data back to an IoT platform that can make your decision based on a particular locality or how many beans need to be emptied. And at the root of the truck can be optimized to save fuel and time and resources of of people who are managing the garbage trucks. The fourth one is a smart supply chain. And so you know, supply chain has many angles if it starts from the Joanie, after all my data is being gathered or transported to a manufacturing site. And then the manufacturing processes itself. And then it goes to a finished products, go to a distribution center or a warehouse. And then again from there they are transported either to a retailer or to the end customers genetically. So smart transport is a very interesting part of the whole smart supply chain process. I'm sometimes depending on what kind of goods are being transported. Additional monitoring has to be put in place so that we can measure how the, how the quality of goods are when they reached the final. So these conditions could be temperature monitoring, shock, vibration. There could be many kinds of conditions or can be monitored. And the fifth is about smart workplaces. Smart look basis also comes on Dojo, smart buildings and smart work business has many use cases. The one we're talking about Period is the meeting room occupancy. So gone are the days when you look at the calendar and find out which meeting rooms are available who has brokered from wartime and whether you can use it or not. What is meeting rooms? Stem cells are able to tell you whether their book, they're not based on the proximity sensors or the heat sensors or the occupancy sensors, meeting rooms? No. Even if their vote at a particular given time, they're being used or not. So that is being used these days and smart buildings, smart workplaces. In fact, these smart meeting rooms can also set the edge vaccine stones in more energy-efficient tweets. So changing the heating and cooling and even lighting. That is provided to meeting rooms based on whether they had occupied at a given diamond. For D1, smart retail. A lot of retailers across North America and Europe are experimenting with having smart shopping colonies. So, so customers do not need to queue at the checkout counters. They can use mark trolleys, go two different ions because put it in their colleagues. And finally, at putting it in the trolleys, smartphones can scan the products on the goal. And once you are ready to checkout, the scanner on the shopping cart is able to take their digital payments, either too flashing the card or using mobile payments through anyway, and Checkout is completed. Within their colleagues said that there are many elements to it, like integration with other inventory systems or e-commerce engines. There many angles to it. But smart shopping carts is the wave of smart retail is going these days. So if you really look across all these use cases, we see that IoT has really provided the value to make these use-cases happened. Without which we could not have talked about a list, mass market based management or smart places and even connected health care. Here I have just presented six use cases. But door lighting there are more than 600 probably. Good IoT really provides value. So I think that gives us a reason to study IoT furthermore, and get deeper into the architecture and design burdens and see how it brings value. 5. Data Transmission in IoT: Then we discussed the definition of internet of things. We talked about devices being Conversation that we talked about the data flow between devices and the IoT backend on IoT platform. And that is the crux of IoT is here and see what kind of data actually flows from devices to the platform. Iot back-end are in the reverse direction. Will IoT back-end DO devices that are different kinds of data that flows from, from each side. And we need to understand it because this is the crux of the whole of IoT ecosystem. So let's see. Iot backend here as shown, is able to receive sensor events, alarms, and device. Let's go through them one by one. Then you say sensor data. I mean data that is sensed by the devices. Here devices, not just the end devices, but it represents sensors. It drips engagements. And it also references the devices that are capable of sending the sense data are directly to the black fall. It may not need a good proof in that sense, it may have an IB, IB stocked with inbuilt in it so that it is able to go to, to communicate on IP-based protocol. So Devices here is a logical which means any, any of the sensors, devices, gateways, objects that are able to send data to do an IoT back-end. So in that sense, sensor data is actually at time series data that the sensor sends in the device ecosystem, something like temperature, pressure, vibration, shock, location on a few levels or garbage level, or our presence of people. It could be any of the things. So this, this is generally time-dependent. So this data can come every second, every human, every few milliseconds, every minute, or maybe sometime in few hours. So this data is already time. And this sensor data is then captured by the devices and sent back to the IoT platform. Second kind of data is the events. Events are something like a refrigerator hazard door opened every hour. So every opening offer DOD is an event or for that matter, if there's a connected car that is being monitored than a driver dating arrived at a sharp angle could be an event to depends on how we define events for each domain. Events are happening in the normal functioning of devices that have alarms. No, alarms are different. There's alarms are of two kinds. One is that devices themselves are able to identify whether it is an alarming condition for an error condition. So so they are able to send that data back to the IoT doctrine. Dark, there was an alarming condition that has happened because an alert notice happened needs to be based. So they send an edit gold on device itself is able to do that. The second kind of alarms, business allowance or business edit conditions, which are gone for good on the IoT platform itself based on the kind of visual data that is coming. So such a condition could be, for example, a refrigerator door is opened. Five dimes. In a few minutes and the temperature goes below a threshold for a longer period of time. Not this kind of multiple error conditions. Multiple conditions can lead to a business at a condition which is generally configured as part of an IoT service in the backend. So here we are not talking about those. Although such business rules can be applied on devices itself in what we call as edge computing or Edge Intelligence, in which gateway aggregates that data from our Sensors and Daron. And locally it is unable to apply certain rules and find out at a conditions. In such scenarios. These alarms, I have both the elements, something that device itself generates. Our gateway analyses some data and gender, age, those alarms and sent back to the IoT backend to process them. And finally, there is something called device. All the devices that are connected to a platform have their health monitored by the platform. Not going always knows or the IoT back and always knows what kind of devices are connected or not reporting at the moment. And if some of the devices are not reporting, it knows that those particular devices have died. So when appropriate error condition needs to be raised on, maybe there would be ignition that goes under multi-colored reconfigured or whatever. It has to be a process to deal with. And therefore, depending on what kind of protocol is used for data transfer between devices, on the backend Flatiron device has to be either investment zones are sometimes doctrine. It sets, the protocol itself, manages that if it's a persistent connection and the connection is broken, we know that the device is not responding. So in that sense, the protocol being used, for example, MQTT in this case, is able to is able to notify the backend dark. The device is not responding. So these are the kind of the streams are strong devices do IoT back-end. On the reverse side. I would be backend is also able to sing certain commands back to the device is something like reconfiguration of our ignited quiz devices to send back log files. So these are, these are the kind of commands. Not sending the commands is a different process. We can discuss in detail when we talk about the connectivity part. The MQTT broker, for example, offers persistent connections between the backend and the devices, which makes it easier for the platform to send back, come on to the devices. And software upgrade or devices is generally about upgrading the software founder off the devices and when devices are GO spread their located across the geography. I'm Dale funded or software needs to be updated and we do not have generally send technicians who do a manner let it needs to happen over the ear. So these grids require commands to be sent back to devices to notify them of a certain location like a CDN repository on a database from where these devices can download the software patches and install and then falling back to the IoT back-end. Software is upgraded. Similarly for control. So the troubleshooting or devices, there has to be some sort of mechanism, right? And SSH Donald learned SVD predominant there is formed between a backend API and our device so that ignitions can fix the issues within the devices. It could be restarting the devices are doing a software package or reconfiguration. So these are, these are the set of commands that we need to consider when we design an IoT system. That helps chart out the data transferred between a backend and devices from the reverse kind of the backend is able to send data to devices, are combined to the devices. All in all, the states are transferred forward and reverse is set out to achieve certain things. Those, those things up those objectives reach operational efficiency. How to make operations beta for any business process, don't wanna go certain systems because the sensor data, events and alarms and device healthy are able to monitor basically the whole inventory of devices connected to the IoT platform. Cost optimization. And V are able to do energy management and reduce the cost of operations. And finally, we are also able to do in font decision-making because once we have all the device data in the back-end, we are able to run analytics on it to infer in foot important information based on which we can decisions, operational decisions or business decisions so that the systems perform better. And that is the whole idea of IRA. Basically, we just don't want to collect data to store it. We want to analyze it, to visualize it and make sense out of it so that, so that we're able to make better decisions for the overall efficiency of the system. 6. Primary Building Blocks: Let's talk about the primary building blocks of an IoT system in this lecture. Based on whatever we have studied so far, you must have some idea or four down to components. But let's go through in detail what all is involved. When we, when I say building blocks I'm talking are at a very high level. So after dark level, we have devices, IoT, Black, Enterprise and external systems that interact with the IoT platform that comprise off an IoT ecosystem. When I say devices, it represents the end devices. The equipment also represents a sensors are put on the end devices and the gateways collect the data from the sensors and is able to transmit that data back into the IoT platform. So devices is in that sense, a structure which captures victim, which consist of all these kinds of n devices and sensors varying you are able to measure the conditions are the contextual information like embraced or humidity, level, events like door opener and chugging down or the speed or the fewer levels. So all these conditions are monitored and the data is collected in devices or gateways and then transmitter and I already dark phone. So this bidirectional communication is an expert written V, as we talked earlier, devices are able to communicate bidirectionally with IoT platform. On the other side, IoT platform is able to send commands back to the devices. The second loop component, and here is the IoT platform itself. And I'll IoT platform consist of multiple functions like data ingestion. It consists of platform services, business services, and our data is stored it as analyzer. It is processed some stream analytics algorithms and apply it, then sum it did on Nixon's audits are done on it. And then it is after the data is processed, it is then visualized. And further, diode. Black phone also integrates with enterprise and external systems. I would get blocked from, could be built in any way. It is just those set seven or eight functions on a few more functions that comprises of what we see as an IoT platform that can be built using cloud-based technologies, could be built on cloud infrastructure. It could be based on, deployed on on-premise data centers. Using open source components like Robert M Q, or Kubernetes or any of the other. And then I say cloud-based models. It could be based on infrastructure services, like using only the virtual machines are the Azure VMs, for example, load AWS, Elastic, Elastic, Compute, and easier to instances are it can also use lack from services using components. Like IoT Hub are Azure, Kubernetes Services are DynamoDB is these, these are platform services that are provided by cloud providers. But you need to configure them to use them as IoT services. Finally, you can also use SaaS services. Something like Software as a Service and read him that our components are available to be used as an IoT Service art as an IoT component. Something like Azure has recently come up with your IoT central, which is, which covers a lot of device management and data ingestion components. And it is available right out of the box. And I don't need to build it using past components. There are pros and cons of building systems using all three of the options. So based on the problem at hand, we decide how to build these IoT platforms. But IoT platform as such, the building block channels built using any anymore, it just needs to use of all those services. Further, there's again, a bi-directional communication enterprise at external systems. These systems could be any system and I say external, it could be. But that API would be any other system which is outside the scope of an enterprise. Enterprise systems, typically, typically our CRM systems, SAP system, ERP systems are onwards engines in case. It is a very significant, very significant component in the whole all sort of building blocks. I can give you a small example. It will make a point here. For example, in scenarios in a retailer, supposing a retailer employs a new channel to sell its goods and goods using Kanban goal kind of an option beta. Beta customer goes to a store, scans of product himself, makes payment and just goes out, checks it out himself and goes out. It's a completely new kind of a channel. It's actually a kind of scenario if you see to really see the buddy deal jargon. So in this scenario, the customer gone through a proper checkout process using the services of a customer service agent who was there and checks it out, checks all the goods, production bar and makes the payment. Those boss systems are point of sales systems are integrated with something like an inventory system. If someone buys a marginal bottle of ketchup, for example, vaunted has checked out on the transaction, completes. The inventory, has to be updated. Burden that bottle of ketchup of Doug Grant has to be the inventory has to be reduced by one. What if there's a new channel like scan band goal? The customer, although it's a new channel with an IOT implementation, unless that IoT landfall interacts with that Mendeleev system to upgrade, upgrade, then Ben-Gurion reducing it by one. The whole business process is not complete. So IOT brings in a new way of selling as a new channel of selling in photo editor you dealer. But it has to interact with all those endocrine systems so that the whole reading experiences in that. So that's why that are very distinct. Three building blocks here, devices, IoT, platform, and an enterprise or external systems. 7. Technology Makeup: Energy is not a single technology. It's a, it's a veil of solving problems. It is, it is an ecosystem of a lot of technologies that is able to solve certain kinds of problems. So in that sense, we haven't done a lot of different functions are comprised of IoT and IoT systems. But you realize those functions to build an IoT system is using technologies and many of them that are involved in building an IoT solution. Let's see some of them here. Embedded systems. Embedded systems are about a device engineering part, which means it is about sensors. How would we program them? How we build a software using goes like embedded or assembly language or lower level tools to make sure our data is captured, aggregator and sometimes even analyzed when we talk about edge analytics on the device side. So lot of embedded technologies are involved as part of an Internet of Things already or part of IOT actually. And second is the communication systems that are stored. One is what kind of acknowledging is used for communication. Something like which protocols there to use. It could be it could be a different set of protocols between devices engagement, and then the difference between the gateways and IoT platform. So devices and gateways generally dark and goes out lower level protocols, something like BLE, Bluetooth Low Energy, or B or Z. R Mark was full of ten, but these do not involve an IP stacks. So they had an IP-based protocol. Generally. On the other hand, a communication between k2 and IoT platform Ingalls IP-based protocol because the communication happens over the next book. So here we're talking about protocols like HTTP, MQTT. Protocol is typically, these are the data that I used in IoT communication. And they're open verticals. Any kind of 32s are libraries of consumer libraries can be used on both sides to make sure that communication happens. The second part of communication is technology. And the sands you've been using, cellular could be using WiFi it. You could be using the wireline. Any kind of connectivity between devices and in Arabic. The third kind, which is very important is the cloud computing. Because unless you're building a system and IoT system, which is based on on-premise. On-premise technologies are datacenters and you're picking up open-source components to build some small programs. Most of the IRB production level IOT implementation happened over cloud. Something like AWS, Azure, Google G predicts odd, even. Rto, deep dark phones are like cloud-based. Is a very core part of an IoT platform. The third kind is the analytics that are two parts to it. Again, one is the streaming analytics to find out alarms and alerts from the streaming data. And this is more like what we call as hot path analytics. And second is according to Bhutan, which is when we aggregate a lot of data in the IOT platform, we are then able to run certain and God attempts to find out influences on the data that is gathered, gathered from the device to make some decisions and make some inferences. And then process the data card. And it could be and ecosystems. It could be some, some machine learning, ill-gotten town. There could be artificial intelligence, natural language processing, and many bonds to this analytics, to analytics, this is abroad. But a lot of these things can happen within the framework of analytics. The next part is the integration part. By indignation, I mean all those technologies that are used to integrate an IoT platform for third party systems like external systems for example, better API or the traffic API are, are are some third party that gives you data to enhance the data collected from devices. And also the enterprises that can give you reference data. Make sense of their device it altogether. And his indignation is all in sort of the same process, data or data to other enterprise system, maybe for long-term storage or for other systems to use. These integrations couldn't. Generally they involve API has a message for intermediate layer to. The next part is the mobility, the mobile application. Generally, the provider group provide reports and dashboards. It could be on the data that is gathered from the devices. We always employ argument dashboards to monitor and other devices and to configure their devices and go do the software upgrade or fair. But all of these options are available as a stop services. Mobile services, better. Sum all. The use cases of IOT are all mobile. For example, connected blood pressure monitors, for example. They communicate over Bluetooth. And these Bluetooth enabled UDP meters can be connected to a mobile app, which everybody has mobile apps. And mobile here acts as a gateway to mobile applications. Get could be end-user applications are they could be intermediaries like gateways to collect data and all worded platform so that this data can be gathered. And finally, the UI UX kind of technology that could be for web apps. It could be far overlaps, it could be other dashboard. This is a very, very critical piece as well because most of the users of an IoT system would not directly interact with the devices, but they will do it using using different form factors. And a UI. Ux technologies are very important to make sure it has a very interactive and intuitive kind of water. So all of these technologies and make up what we call as intended of tanks. 8. Case Study 1 Connected Health Care: In these two lectures, I am going to talk about two case studies. One is about connected health care and the other one is about asset monitoring. The idea of bringing case studies to these, to this course is to show you how IoT systems play a significant role in solving problems. There's a whole gamut of problems that IoT can solve. But let's see. Let's start with looking at two use cases here. One is about remote monitoring of patients and other is condition monitoring of equipment. Let's go through the first one here, which is that a mode monitoring stations. Here we have a patient who has a medical equipment. Now the equipment can be small devices like nutrition pumps to large devices like hemodialysis machines. Most of these devices have profile. So more often than not, they connect with a gateway over Bluetooth and they're able to send data. So here we show the gateway as a mobile phone, which serves to focus, as I said, have fun as a gateway to acquire data from the connector medical equipment using BLE, Bluetooth, Low Energy. And it also acts as a, as an interface or a user interface to use mobile to see his parameters or the history of the data that has been collected. Some analytics surrounded in an emergency can connect with the health worker directory. So once the data is acquired by this gateway, it sends it to an IoT platform over HTTP, HTTP, or HTTPS. Now secured, an IoT platform is able to gather that data. If it does some analytics on Twitter and then it hears this data or what an API. A health worker. So health worker is able to use a healthier support system, which can be a user interface. On an existing system that dough under the US can show the readings of the patient. And if you really see all these arrows are bidirectional, what it means is the health care worker can then change the parameters of the connected medical equipment remotely using the service in the IOT platform known as command and control. So command-and-control is able to send data are, are the new setting for the medical equipment. To say change. The level of nutrition. Duty to inform the patient to make some settings are to and from the patient to connect to the, to a doctor or a hospital based on the parameters that he has just monitored. This inflammation goes back to the gateway or rather in case of HTTP, the gate drip, a particular queue or an end point. And the IoT platform where it can see if there's a command for it. And then it informs the medical equipment to do the required setting or show a pop-up on the mobile app to inform the patient that I write. Thing to be done. Could, could be, could be monitoring a more global regularly. Our genes don't change the nutrition levels. So it could be any of the commands that the health worker wants to do, that the medical equipment. So if you see in this end-to-end journey, there are two critical elements. One is the gateway, which acts as an intermediary to get the data from medical equipment and pass it on to IoT platform. Iot platform is something that does some analytics on that data and shared the data with the backend system, the healthcare system, which the healthcare worker is used to seeing ready can monitor a particular patient. Iot platform alter, takes care of the Golan and control as we just discussed, on the sort of data flows in the reverse direction are done through an IT platform. So I wanted to show you now the utilise setting. How does an IoT platform play a major role in this end-to-end remote monitoring. 9. Case Study 2 Asset Monitoring: Hi, In case study, we are going to look at asset monitoring as a case study on a particular use case we are going to look at is condition monitoring of a refrigerator that retail store. Let's see how it works. We have a smart refrigerator that we have to monitor. We need to monitor the temperature of it. So we have a gateway. Typically refrigerators are do not come with a gateway that is built in. So they need some sort of a gateway which, which can gather data is and is able to convert the protocol from the one that refrigerator is using to an IP-based so that it can send data over the network. We have an IoT platform here. The gateway sends telemetry or sensor data from the refrigerator to an IoT platform. Now, IoT platform has multiple functions here. It ingests that could be a continuous stream of data are the gateway can be configured to centred at certain intervals, like every few minutes to the IoT platform. Now at the gateway can ingest the data from the refrigerator every few seconds. But based on what we want, we can ask the gateway to Senate only after a few minutes. Now, IoT platform ingest that stream of data and applying certain rules. For example, be warned that if iterator to always maintain five degree C of temperature, for example. And now if it goes below five degree C for a certain amount of time, we need to raise an alarm. So this kind of a rule is said in our rules. A one-step rule is applied. It may send an alert, it may generate an alert. Rich and IoT platform uses a notification hub to send it to a interested business applications. So Notification Hub typically has the capability to send them SMS, send an email, or raise an alarm to a business application so that some technician can look into it. As we see further. And Arctic platform can send this alert. You were ignitions Admin App that education can monitor what needs to be done based on it. On the reverse. If you have a control signal coming from an admin app, which is received by command-and-control function of the IoT platform. It then sends the control signal back to the gateway. Gateway is able to take care of the same setting, sending the command to the smart refrigerator. So Gateway has one gateway can be connected to multiple refrigerators in a retail store. And it is able to get the device ID, for example, from a modern control function, device ID, and what kind of action that needs to be taken on that particular device. So gateways able to connect to the right refrigerator. And the refrigerator has to have that much intelligence will receive that signal and change the settings. So that's how an end-to-end smart refrigerator function works. Now IoT platform here shows only the minimum functional components needed to perform this activity. Although there are many more functions of IoT platform that you keep looking at. Once we see the IoT reference architecture Howard the walls. But I've kept the minimum functions here so that we're able to understand how the end-to-end smart refrigeration works. So I hope you understand how condition monitoring of refrigerator works and how gateway and IoT platform are key components in an end-to-end architecture. 10. IoT Data Lifecycle: In this lecture, we're going to talk about the data flow across an IoT system. So if you see from left to right, there are five different phases. The grass which the data flows in its lifecycle across and IoT ecosystem. So if you see from the left that data is first captured from devices. It is then processed. Further, it is stored. And then some analysis or analytics happens over the stored data. And finally, it is visualized and in different formats. So let's, let's get a bit deeper into, into, into this flow and see what are the different elements that get involved in it. So initially, in the capture phase, we have n devices like connected guards or smart shopping carts. It could be industrial automation. Some connected, and it could be smart buildings, for example, there are different kinds of sensors that are bought into these devices and devices, or sometimes even the devices themselves can send data because they're able to capture their contextual information. And this data is then saying, do a gateway. Now here it is some sort of a generalization here because it is possible that the devices are themselves. We'll do send the data to an IoT platform, but many a times we have to use some sort of a telematics device or a gateway which captures the data, aggregates or does the protocol conversion and then sends it forward for possessing. Further. For possessing, it goes to an IoT platform where the first of interaction is a stream ingestion mechanism, where the continuous stream of data is coming, it is ingested. This data is generally normalized, some certain kind of business rules that applied to it or no processing. Data transformation happens. And after that, based on the kind of eta. And it is stored in a structured database or an unstructured database. These had mostly reference datas are elementary data, transactional data that gets stored and used here. But finally, this is generally a big data system that has a data lake. It could be an enterprise data hub or Data Lake dying of a place where the data is stored for long-term usage. Most of the IoT systems required that the store raw data as well are generally the ones who do dies processed that is also stored in big data stores like Cassandra or HDFS, often Rich via certain analytics rules over it. Something like MapReduce functions are machine learning and guarded arms or artificial intelligence based incident scenarios like like inland in use cases like predictive analytics for that matter, we need to apply certain machine learning algorithms there. You also may need to apply some natural language processing over it. So all of these complex processing happens during the analytics phase. The output of which goes to the visualization borders. It could be an admin border where you connect. Very review see what kind of device behaviour monitored within our devices are connected or devices that ally whether or not this analytics can also feed apps, can have air via random functionality where you visualize this data I'm and for the end-user consumption. So these, this is the lifecycle of the data flow. Typically the elementary and sands have kind of a data flow across all the phases of often for data manipulation across diode ecosystem. So this is, this is the data flow across the whole IoT platform from the devices to the visualization across all its phases. 11. IoT Architecture Ingestion and Processing: Hi. In this lecture, we're going to talk about nephrons. Iot architecture is a very interesting part of the course, and I think one of the most important aspects as well. So I wouldn't I would like you to understand it. We call it an unreferenced IRB architecture because it is technology agnostic. We're going to talk about all the functions that are in an IoT ecosystem. So let's see from left to right. From the left post, we talk about all the devices like connected cars, smart buildings, smart washing machines, connected healthcare, smart cities, and any devices where you can put sensors and extra eight out of them. As we discussed earlier as men, these devices are not generally able to send data through an API platform. So we need some sort of an IoT gateway, which has also gotten feel good with sometimes a basic functional which is to get the data from sensors aggregated. Sometimes change it so that you can send it on a different protocol like MQTT or S GDB. And then also applies somebody who got a tom. Sometimes even machine learning algorithms are possible at edge now. So this is IoT gateway is, are the educator does all of these functions. What we have not talked about earlier? We do not get Adolf from IoT gateways. You also get it from our sense of what it means is that a third party providers hardware, sensors, and extract data out of the n devices and send them to their own flowers. Now in that case, we can fetch data using their API and strong leverage, which is called device sends up yards. And there is another kind of a different data source that we have. So we may not be touching devices. That is what you might just be getting data from their APIs. All of this data gets ingested through the post wind of it. In an IoT platform. This is typically an MQTT broker. Or it could be any blind audit, could be APA guidelines that fetch the data from the ice fields. And basically it is about fetching the data are the gateways pushing the data into a message broker here. And IoT platform is further pushed. Downstream processing. Stream processing has multiple functions. One is to normalize the data because there are different data types, because it has gone from doing different device types. So we will have infinite speed and structures start to apply some rules on different data structures. We need to normalize them to a common data model. So stream possessing Onsen was due to a nominalization. And we apply some rules which comes under a function gone stream analytics. So those rules and applied to find any threshold bleachers are, are something like the temperature going down beyond a sort and find on a few level going down that less amount of odometer reading any dynamo business students to find alarming situations. As part of stream processing. Often register data gets stored in our database. So these data stores that are shown in the the diagram here, comprise logically or multiple databases that store elementary data, events, alarms, auditing data and reference data systems and augment that elementary data coming from devices. So all of this gets stored in data stores. It could be a mix of SQL and NoSQL database. Further, and this data is used by business services. Business services on the actual. This is for which you are creating the whole dark Sean, for example, encases smart parking. You might be getting data off the data from the devices telling you how many parking spot full at the moment. But a business service related with smart parking might need to find out how many are less common areas. So how many DNR? And show all that data in our visualization. So all the kind of business logic is applied in business services. These business services, from domain to domain. It may vary from smart retailer to smart buildings, smart to ease on business owners and their rest of most of the black form components remain the same. So data ingestion that talks to the gateway, business services are are two elements that keep on changing for every domain. 12. IoT Architecture Analytics, Visualization and Integration: Once the data gets stored, analytics comes into picture and shown here is quadruplets analytics. So what it means is, it is April, you're able to apply machine learning and got to towns predictive analytics, AI, artificial intelligence, any kind of analytics or you want to do on the data coming from devices is done in analytics. And, and further, this analyzed data is also stored in data stores. It could be the same Dido's towards order, could be big data stores like Cassandra, our HDFS, et cetera. All processed data is shared across side the IoT platform using APIs. Now business services also consists of those services which are actually APA backends. Api Gateway, autumn message bus on custom connectors are the elements that are part of an integration services that are shown on the far right. These services exposed the APIs in some way or the other. That is, the data that is captured by and processed by the IoT platform is shared across with other systems using integration services. Now Enterprise Services or an external system, as we discussed earlier, as well, represent systems, often enterprise like CRM or ERP system, salesforce, HR systems, be it old systems, kinda things. External systems could be read that API or traffic API are micro markets for example. So these are the kind of exactly the kind of external systems there can be connected using Integration Services. Visualization portal generally contains two kinds of elements. One is an admin border to control the devices, and another one is to show the reports and dashboards based on the analysis of data gathered. That is, the part of visualization could be in the form of a web app or mobile app, or even augmented reality or virtual reality kindness applications as further data something concert sentences at the bottom. Citizens which are kind of generic services like security, which includes Identity and Access Management, certificate management, logging, auditing, monitoring of the whole platform. So these kind of services are shared, services which are used across platforms. And generally, the infrastructure of a platform provides these kind of services. 13. IoT Architecture Device Management: What do we have? So far is one critical element here, that is device management. Device management is something that separates or data black lung from an IoT block. So if any mid-October Device Management has a few functions. Actually see your fold functions and be precise. One is device registration and registration. One cannot attach or device and expect the black Shuang due process its data unlisted devices registered with the black hole. There's a whole device provisioning process in which bias is registered with the platform. The platform makes sure that device gets. And then a handshake happens and the platform starts accepting the data farmer device. So that's, and that's how the device provisioning works. Similarly, Device registration processes onset. And secondly, you want to configure devices. Every non dense or device configuration is another function, or this part on device management. Thirdly, you need to have some sort of command and control, and command and control and make sure the reverse direction of data can be known from IoT platform to the device. So that it was communication happens because Alignment control, it is able to identify which devices I will connect to the devices, which devices are sand worker minds. It is kind of a state machine that will make sure the commander and the responses are received. And the responses I've gotten textualized and stored in the database. Function is the software upgrade or devices that is about upgrading the software of their devices. It is a complex function where it is again, achieved through Device Management. Finally, ahead of devices, IoT platform should always know how many devices are and how many are coming in the devices that are not responding for the last five minutes or ten minutes. So it can declare them as dead if it hasn't received a heartbeat from those devices every few seconds. So that information also needs to go through. We'll go to the enterprise systems or external systems via that I, ignitions can be in song and the troubleshooting. And so that is part of device management. Arlen, all, all these functions comprise all on offense IoT architecture. Now, every scenario is different. You may be using only some functions are fair. I'm not others. For example, if you are not doing any device interactions and you're only getting data from Device Cloud, human art, neat device management or Donald audiophile using any honored. You don't need any machine learning or AI guy knows. So there is no need of a cold path analytics block here. So depending on your situation, you will pick and choose what kind of functions you need for your IoT solution. I hope this is Maitreya. Is there any doubts? Feel free to write down the questions. I'll be happy to answer them, but this is something very important to understand. 14. Conclusion: Hi. So this completes our course on the fundamentals of IoT systems. In this course, we covered the core elements of an IoT system. Different data flows as they pass through different services and components in an IoT platform. We also discussed the building blocks of what makes up an IoT solution. Further, the altar discussed do case studies to see a real life examples of IoT. And we also study the chef friends IoT architecture, which is an end-to-end representation of all elements of IoT. Maimed Chrome devices to black farms to visualization. Our last pits of IoT. I hope you enjoyed this course and I hope you totally understand now the basic fundamental elements of IoT and the concepts of IRD. Further, I'll be launching, do more courses on the solution architecture of IoT and also advance architectural concepts in the Internet of Things space. So my best wishes to you as you onboard yourself in this journey towards more learning in the IoT space, which has limitless possibilities as you know. So good luck and my best wishes. Thank you.