Getting Started with IoT: A Deep Dive into Myths and Realities

Hello. And welcome to today’s session where we are going to discuss IoT myth and realities. So I am Nicolas Lesconnec if you wanna know everything about me and my career, you can just hit pause right now. Otherwise, we’ll speed fast with that. So I’m in charge of partnerships at Soracom. I’ll be happy to get in touch if you want to discuss anything related to IoT and partnerships. So who are we at Soracom? We are a Japanese company, founded a few years ago in 2015 by AWS veterans and our goal is to help the IoT revolution, and we are providing connectivity services. And right now, we are powering a bit over five million IoT devices and serving several thousand customers. So let’s jump right into today’s topic right now. What is IoT? So let’s start with the official definition from the ITU. So the IoT can be viewed as a global infrastructure for information. enabling advanced services by interconnecting things based on existing, enabling interoperable information and communication technologies or otherwise known as ICT. So we could end the presentation right here Oh, let’s try to go beyond this long and detailed definition. What’s behind this IoT buzzword? This IoT thing has been around for, let’s say, ten years. Even more than that, if you consider the formerly this out is probably known as M2M machine to machine. And you may think that, okay, you know already everything about IoT, you attended all, you heard it all, let’s set the record straight. The first thing is, usually, people saying to me that the no IoT will speak to me about The useless crap, let’s say, frankly, that’s around, a lot of useless IoT devices, I was not allowed to put any images here illustrating that with concrete, products. So all you have is a little comic about coffee machines and what the points of connecting them. If you want to, have a look and discovering what kind of Very useful. I don’t see what devices have been around, and it’s appeared very quickly for most of them. You can check out online. You find Southern of examples, You have a Twitter account for the internet of that’s go that was a lot of those useless Let’s put a chip in it, solutions that in the end doesn’t break anything. But today, what I want to think about is the actual IoT internet of things solutions. This is an old analogy that I’m using a lot. But I still think it fits, which is the whole iceberg thing. Considering that those shiny consumer IOT gadgets that I just mentioned without naming anyone, Are the tip of the iceberg. So, yes, you’ve got plenty of consumer IoT solutions around. Great. And are very valuable and are very successful. I’m not denying it, but what I’m saying is that this is the, actually, the tip of the iceberg. And the majority of what you get in the IoT world are B2B solutions, usually what I would call boring, B2B use cases, boring in a sense that they are not all that exciting. Their value is not in positing the last new thing. Their value is what they are going to bring. Two businesses. Be it by bringing a whole new revolution or incremental improvements. But things that are actually going to improve and to add the value to business operations in a lot of different activity sectors. And what we are moving towards, I would say, is from this technology centric revolution, a lot having the last new thing, basically, wanting to do IoT to having something that is a tool, methodically, being used for your activity, your business, your solution. Same as we had a few years ago, I don’t want to count. Which is being used to all. You’ve got this personal computer revolution that is now everywhere. Now a very limited number of businesses that could operate without those computers that were not ubiquitous forty years ago, but you just forget about them now. Then you got the internet revolution. Exactly the thing. Most of you today could not, let’s say, survive the working week without internet access. You don’t even think about how a re how much of a revolution it has been or it is. For your use cases. It’s just here. Then you have the mobile devices, smartphones, tablet, you name it, and That’s where I’m putting the IoT. And maybe I could already update this, chart by adding this all, AI conversational AI that is, moving like a tornado, across business operations, in the recent, few months. So my point here is basically IoT is not just only a technology thing for technologists looking into the new technical thing if you if you make. It’s more about bringing something new, a new way to do things, a new way to access information just as you had with all those previous, technology revolutions. And the more you forget about this technology in the end, the better because you’re just thinking about what’s it bringing to you. As you don’t see that much about how your computer is working or what’s the magic in is inside. It just the tool that you are using to achieve something. While we are in the analogy, so I did the other thing, now and if another one that you may have heard a lot, even a bit more, even a bit too much, if you’ve been in the IoT space for a while, which is this ‘data is the new oil’ catch phrase. So that’s an old thing, not especially tailored for IoT. That’s something that has been, say, by, Clive in 2006. Thinking about that I’m not in this context of IoT, but worldwide and how it’s changing, the landscape in terms of platforms and so on. But it’s really fit to what IOT is bringing. So reading the full quote, it’s not only data that you, it’s that it’s valuable, but If refined, it cannot really be used. Same as crude oil. It has to be changed into something else. Gas, plastic chemicals. To create a valuable entity that then drives profitable activity. Exactly the same way with data. It needs to be broken down refine, analyze, and then it can reveal its value or bring value by enabling a lot of things. That’s very similar to me to this old oil extraction, refinery, transport, exploration that you have in the end. So what does it mean? Is that IoT, basically, is here to help digitize, the physical world. Let’s say kicking all that. A lot of things have been done for a while. Good enough. So, in this example, is it better to just someone kicking tires. It’s a it’s a good a good mechanism to have an idea if but a flat tire or not. That may not be the most accurate one, but you you get the idea. A lot of things are still done out of all the habits, human checks. We all know we’ll have issues with that. We make mistakes. We misrepresent, on top also not the reality of what’s happening in the activity. Then you have to gather all those scattered, data points. That are coming from someone writing something on a notepad, someone, phoning in a report, It’s delayed. You have to consolidate all of that fragmentation of the data. You end up with Excel monsters that Actually, most of you have suffered with or are still suffering with at the moment where you’ve got a lot of data that you somehow need to make make sense of. It can work each work most of the time, but that’s where IoT can bring us something. Is making sure that we are moving to this, let’s say, all the way of collecting aggregating, analyzing data to something more structured. So digitizing the physical world, meaning that you will Through technology, measure something. Whatever makes sense. It could be a position. The temperature, voltage, a level of noise, of detecting events, something has been moving, a door has been opened. Time has passed to its time to send another data point and so on. And then you will get information, which is more trustable, more consistent, easier to process and with this more precise view of the context, and easier to build analysis, that should bring you to a more in or a more inform the better informed decision making. So what What for? What what could that bring? You may have seen a lot of things about the smart city, smart roads, smart grid or smart city. Yes, IoT will contribute or is contributing to have, to be correct, a smarter world, even maybe a smart world, but let’s say smarter to get start this great great illustration, by the way, has been sourced from Libelium, company doing IoT Solutions, lot of, great devices. So, check them out if you’re interested. And I took that illustration from them because it gives a a good view about vertically everything could use or could be impacted by IoT, be it maritime operations, infrastructure monitoring, garbage collection, parking monitoring, wildlife or what, of white donuts monitoring of events industry, City life, building management, road management, traffic management, you name it. You’ll have a lot of business vertical or human activities or anything, but it’s in our physical world that’s going to be either impacted or will get value from IOT solution. So you’ll get a lot of examples on that. Again, hit pause if you want to spend time reading each of those items. But not everything has to be smart. So smart CD is is powered by a T smartphone, powered by a TNTS one and so on, but it’s Not just about that. You’ll have also a lot of dumb devices in the sense of sensors things that we not need to have this, let a local brain that will just collect data or detect events pass them over to something that we should we could consider like a decentralized brain, which is going to make sense of this aggregated data and then maybe trigger, change of operations, change of behaviors either to the same devices, just send them an order, shut down the this valve, raise an alarm, or maybe it’s going to just push information or orders through other parts of your business because based on the information that I don’t know. This shipment has been delayed or is properly damaged, because of what we are detecting, then let’s put the information to the to the team or to the factory that was supposed to receive that so they can just they can adjust what they are doing right now. So in all, it can contribute to a literal smart factory, smart operation. The devices themselves can be smart. They can have CPU, whatever you need to have high, compute locally. You have even AI on the edge, what you call edge AI is basically having these artificial intelligence on the edge, on the device itself, but in the loss of situation, the devices by themselves, as them as they get, they are just here to be the interface between this physical world and let’s say digitized, world. This is, the combination of all of them that will this hide somehow effect that will make, something smarter and improve the smartness, let’s say, of your operation. Because don’t think I see. I again, and I will maybe repeat a bit too much, today. Your devices themselves your IoT devices do not need to be smart. They can be, but in a lot of case, they are redone. Maybe all you ask from them is to tell you where is your trader right now? Or where is it? Where was it when it was last seen? Where is your shipments? How hot is your fridge or your refrigerated unit. That’s the basic information, a temperature. How bad is the air in the classroom that’s c o two monitoring? Again, that’s a basic information. You don’t need the device to be smart. Maybe the device itself doesn’t know that the hair is good or bad. It will just tell you the number that it’s monitoring. And up to you, to interpret it to your application, to interpret, and to take action, and maybe make a LED blink send a notification so windows gets open, and we get a better classroom here. So what’s in an IoT device? That’s a whole rabbit hole. I love technology. As you may have seen, if you if you read my bio, at the beginning. So I love technology. I love getting deep into those, technical, topics. But it’s somehow handless in the case of IoT because inside, thing connected or communicating thing, So I’m sorry. I’m using those words interchangeably. IOT devices thing connected communicating things because that’s what it is, and I’m using thing on purpose. It’s Maybe not a a full object that you interact with, or smart, whatever. It’s, I think, could be a chair, a table, whatever. Which is communicating and sending information. But basically, if you zoom out or if you think about these IoT devices, well, or this IoT solution. What you’ll have is something to sense the environment. Again, temperature and movement, CO two, in here, really basic information, the basic sensor. You have the power source, of course, either your thing is plugged or is powered by batteries. You’ll get either a CPU or the microcontroller I think that’s that’s where you have your local logic, your embedded software, depending on how dark or how smart it is. It will change this level of complexity. The kind of code you can work you can write are using a full blown operating system are you walking up with a MCU? Are you, writing low level code, that kind of things? Again, on each of those layers, so in the new f communication, You’ve got the antenna for the data to get out. You’ve got this connectivity layer. So this data that is sent by the thing can be received. And processed by a centralized brain. Okay. But each of those layers can be again open and split into a dozen or half a dozen layers each way, and you’ll get full expertise. And it’s very odd. Not saying impossible very odd to have real true expertise over this full spectrum from the sensor to the cloud side and the computing and the decision making part of things or the dashboard. Whatever you are using at the opposite end of that, IT spectrum. So each of them are all all the can of worms are rabbit hole that you could get into and so on. What I just want to say here is basically understand those layers understand where you have expertise where you can yourself or your company bring add value and where you should, seek help. Don’t try to build again a new sensor. Look for the one that is the most fit to your purpose. It probably or maybe not exist on the market or close enough. Again, but if that’s your expertise, then you can dive into and offer something new. But if you’re an expert in sensors, chances are that you are not going to be experts in cloud computing. Same you will have or you would require expertise to write your embedded software, but you require something else again to have a proper antenna. So all this clever thing that you built in your embedded software can be actually sent out there. So to sum it up, lots of components in an IoT device. Each of them requiring Very precise and different skill sets and with the whole set of constraints that you’d need to take into account. The saying is that the chain is as strong as the weakest link and that’s very often the case. It’s very easy with IoT to build something, which is great on paper, great idea, but one of those layers has not been taken into consideration seriously enough. You didn’t have the skill set for reason or another, and you end up getting something that if it doesn’t work, or require lots of overhead, maintenance. And in the end, tanks your project because it’s not efficient enough, and you bleed time, you bleed money, you’ll bleed the effort, trying to fix that. So just take that into your account, and something that might be missing here is security, Again, thing often, goes or white hopefully that the S in IoT stand, for security, because of bad blood, bad history of people not taking security seriously enough in the IoT world. Apologies, Again, I won’t, dive into this specific rabbit hole of IoT security. One thing to really take into consideration or have in mind is that Security is not a feature. It’s not one of those layers. You cannot just add security on top or at bottom whatever. You cannot just add security as a feature. It’s a process. It’s a evolving process. It’s not something that you could just simply turn on or off. It’s not either I have security or I don’t have. It’s risk arbitration, meaning that you need to do, risk assessment, risk management, to make sure that you understand that each part of your device, you have good understanding of what is the attack surface, what is the risk associated, what is, the probability of something happening at that stage, what is the associated impact, the intensity of something that could happen. And again, chain is as strong as the weakest link. So you have to take charge about the security of the device side of things. What what happen if someone gets the hand on my device? They get they get physical access. Do we have a the do we have a debug port that I let on my production device that makes it very easy for someone to, to get into. Do we have, anti tampering mechanism in place at the hardware level? Again, we have been connected or co connected and all communicating devices. So what could happen in case of remote access? Could that remote access mean that my whole fleet is compromised? Your usual mistake is adding a shared set of credentials or private keys that are installed on each device, and and if someone gets access to one of them, they can compromise your whole fleet. What happened in the past is devices being turned as attack vectors to third parties, being part of large botnets like, what happened a few years ago with the, mirai botnet that was, based on thousands and thousands of devices such as security cameras, badly protected that were infected and turned into attack vectors. So in that case, in your risk arbiter risk arbitration, The thing to consider is what can happen to myself. How bad can it be for my own operation if my fleet of devices are compromised? But what are the kind of damages it could do to other people or other company businesses if my devices are turned into attack vectors. So you have this security on the device. Again, that can be broken down into plenty of different things to take care of. Then you’ve got the security of your your data. And after that, when it’s on the move, when the data is in motion from your device, to your, what I call the central brain, early on, most probably, your cloud application. And then what happened what is your security level on your data at rest once it’s stored in your database or in your home system? Because that’s kind of a typical mistake to spend plenty of effort of taking charge of the physical security on the device level. Making sure you have hardware secure elements, anti tampering mechanism, safe management of your device memory, You take you encrypt whatever, data flows that you have to make sure that nobody can listen in the middle and intercept your data. And then ends up being installed in an unsecure database with admin admin as credentials, for example. And it makes it very easy to anyone to just access or even alter, the data of your whole system. And lastly, maybe it’s the business application itself. Who gets to access the business application that interacts with the data? How good is it protected? Again, is it available online to anyone with, default credentials that kind of thing? I’ll finish by mentioning the elephant in the room, which is this nasty picture that is on the right, I have been trying to find a better version of that one for years now, but I’m still using that one in my slide decks. Because I I find it’s a good illustration of what I’m trying say in regard to IoT security. But if in that case, just imagine this is a very important, isolated, control room whatever is inside. It doesn’t matter if you put the best built in class, world class locks. Everything is great. Highly secure. Access control is top notch. You name it. But if your actual door is made of cardboard, attackers intruders won’t lose, the time trying to fight your world class lock. They just push through the door and get in. And let’s apply the same to an IoT system. Think of it as a system. And again, as a security, that’s something that is never, done and finished. If you don’t take sufficient care of one part of this whole chain from the physical device to the business application through data transmission and data storage, then do not consider your entire system to be secure. That’s really a whole thing that needs to be well covered. And of course, from the outset of your project, I said this earlier, but I’ll say it again. You cannot just slap security on top of your IoT system. It needs to be a consideration from day one and again a risk arbitration. Not every not all IoT system in the world need, world level on. Sorry for my English. But not all solution needs military grade or whatever level of securities. Again, risk arbitration because What are the risks for your for you, your company, what are the possible liabilities you open yourself to? And put in front of what it costs to have top security physical in motion at rest and so on. Cost in time, in money, in energy, sometimes, and so on and so on. So this is what I was about this, side comments about security. But another common kind of mistakes that is off is still often done in the IoT world is focusing on technology for the sake of technology. So good example of that are communication protocols. I’ve been working in the world of communication protocols for a while now, and you see that people that get all excited by their own. All encompassing are that protocol is the best. You need to use NB-IoT. NB-IoT in the future, or you need to use satellite based communication because it should be ubiquitous. You need to use Zigbee because that’s the one thing, rate performance and whatnot. So I won’t go into each of those blocks. You’ll have different means of communication, different protocols, I’m not comparing similar things in that picture, but just give you an idea of there are a lot of them. NB-IoT — should I use NB-IoT? Should I use LTE? Should I use LTE? Should I use this or that? That should never be the first question or the first decision in an IoT project. The questions you need to address are, what do you actually need What do you need? And and with emphasis on the need, because very often what I’ve seen is people focusing on what they’d like. I want live information about everything in my system. I want to know where are every one of my shipments on my truck real time? But do my business really need that real time? I’m not sure. And maybe it’s too expensive. And maybe having a report every hour, or whenever something something some of the shipment stop or depart is enough. It’s really about not the technology, but really about the business need and what’s going to actually bring something to it. And then, are you able to fit this IoT project, this IoT solution that you are trying to build into business operation? At the end of at the end of the day, if it doesn’t fit, into business operation, and by of it means if you don’t adjust your business operations to adapt to this, new solution that you’re pushing, then it’s going to be ignored, not going to be used, and you’ll end up with yet another useless IoT project. That was maybe great. Maybe the device that you built was amazing, but If the data that it’s sending or the orders it’s supposed to receive are not linked to your actual business systems. It’s going to end up useless. So the key thing is about the specific use case and thinking hard about What do I need to do? What can I afford to do? Because it still need to make economic sense in the end. And then you can enjoy and get deep into all those technical rabbit holes and getting into the right technology and fine tuning everything. But first and foremost, what do you actually need? What are you trying to build? Why are you trying to build it? And what is it supposed to achieve? And how is it going to fit into the growing scheme of things? So a few factors, and that’s going to be the last part of the, today’s video. A few factors that you need to consider to have a successful roll out of an IoT project. And go beyond something that go that works well in the lab or well on paper and never gets used in the end. Staying in the infamous proof of concept or PoC limbo that a lot of IoT project over the last years have ended up in. And really, all in all, it’s about business consideration. IoT as a set of technologies should be used to serve those business considerations. You should not try to force your technology of choice into the business because it’s most probably not going to work. So this checklist is about what are your goals, what are your limitations, Where are you planning to deploy your IoT things? Are they going to be deployed in a location? If you know in advance and you control, So are they going to stay stationary? In a factory, in a building, and then you can use local communications and so on and so on, or are they going into areas you don’t control, meaning that in terms of communication, if I get back to those connectivity issues, you need to use operating networks because you cannot just deploy your own gateway, in the wild, and maybe you don’t even know in advance where your thing are going to end up going or moving across. How do you need your deployment to scale in terms of How much and how fast? Are you looking to have a few dozen devices, and it’s going to say it like that? Are you going to need tens of thousands hundred of thousands. Is it going to be a fast ramp up? Is it going to be a a steady growth. That kind of thing that is going to change maybe the way you do things and your ability to scale also on the cloud line of things, how are you going to scale those operations? Because you are not going to manage a dozen sensors the same way you are going to manage a million devices that may communicate at the same time and so on. Where where are you? Maybe not have passed on this other company at an organization. Where are you in terms of digital maturity, and that goes back to what I was saying about fitting into business operations because there is no point in building the Let’s say this shiny new version, this new connected or IOT, X, this IOT solution. If it cannot fit in your organization because you’re not there in terms of digital transformation. Third, and that’s something that maybe should be worked on first. And so how are you going to integrate your deployment with your processes operations and existing technology solution. Because most probably this IoT project you are building is not going to exist in a vacuum. It will need to fit into your organization. Either you need to adjust things, speaking about digital transformation, or So making room for this new, project, or you will need to make sure that this new piece of the puzzle will fit right in. And of course, consideration should you build your own or buy either globally Oh, at each part of this, multilayer’s, multilayer, vision that I shared earlier on, basically, Should you build again a new device, or will you find a device that will fit your needs? Maybe not a hundred percent, but maybe eighty five, ninety percent of the units at a great cost, availability is tomorrow morning, and maybe that’s the right choice for you to, to buy. So again, not getting technology first or technologists first, and making time to assess, this build up by, sorry, making time so you can take the right decision about building everything yourself, building part of it, or by buying an existing IoT solution, turnkey, already does everything or part of it part of those, tech technical bricks in your IoT deployment. That’s really a critical one because That’s an old adage in technology, but you may have the temptation to reinvent the wheel because existing wheels are not as great as you’d like them to be. But you’re going to most probably lose a lot of time trying to redo something incrementally better than what already exists. Instead of focusing on what is your key value and the key expertise you can bring in and focusing on why you are doing this. Instead of the tiny bits of the whole. And so speaking about use cases, we can get you a bit into details, today about what are the top consideration that you should have. One is the data volume itself. Because I’ve been I’ve been speaking about devices, sensing devices, but you’ll have you’ll have a wide range of ways to transmit data. That’s going in some of them are going to be very tiny amounts of data. As I said, the temperature of just a one or a zero basic information. Door is open. There is closed. Maybe you are going to send a full video stream that’s, always on camera. Maybe you are going to send a lot of telematics information and this will change the way you are going to, build your whole solution. In terms of technology choice. Local local computing needs, the kind of connectivity you can use. Of course, you are not going to use, LoRaWAN or Sigfox to stream video. That’s not going to fit. But if you are just sending a temperature measure, it’s good enough and not waste your time, your time using LTE to send the information about the temperatures. Fourteen degrees, for example, not even doing that. So think about that. And if you are using a few kilobytes, a few megabytes or gigabytes of that are every other week or every other month is going to change the wall architecture of your system. Another consideration in terms of the data that you’re, communicating, it the what I would call the refresh rate, basically, How frequently do you need the information? I give the example early on about ship shipping and monitoring and track operations and the need or not to have live information. In some cases, you will need a constant stream of information. Because as soon as something hits a given threshold, it’s going away from your usual operating conditions, then you need to act. But in some cases, everything the only thing you need is a proof of life every other week, every other day, or maybe, every other month. It’s just okay. This remote pump of whatever is still operating. All these, remote warning sign is still up, and that’s good enough. You don’t need more than that. You don’t need, a video view of this equipment. What you need is a cheap thing that is going to send information every now and then. So and then you can add all kind of granularities in between. Hourly updates, daily updates, every fifteen minutes. What, you name it. But again, that’s very important because that’s going to change a lot of things in terms of connectivity technologies that you’re going to choose, but also in energy management. Because of course, something that we is going to collect and transmits constantly is going to use way more energy than something that just send a very basic information every other day. So that’s going to change again the way you think about your device and about its deployment, maybe you’ll need an energy source. So Of course, where is your data sent from? Is another consideration to have in mind? Stationary, moving, you are not going to do the same thing or you are going to build differently for gas meters that most probably will remain stationary. If they start moving, you have a whole new set of issues, but it’s going to be way different be between those stationary devices. So utilities, building management, that kind of thing, those devices will remain stationary on the factory floor or moving assets. Are you monitoring the whereabouts of your spare parts across factories? Those assets are moving or you’re monitoring your fleet of cars. They are moving either in a semi closed environments in a city or in a region or they are moving globally internationally, that’s going to change a lot. Again, stationary moving, control and control. So it could be a closed loop in in logistics, or it could be unknown locations, Again, that’s going to change the way you think about your project and the way you build the technology solution. And then in term of the radio transmission, if you use wireless transmission is is your data being sent from Hodo? So you have little constraints in terms of propagation of your signal of initial attenuation and so on. Are you going to have a sensor which is deep indoor in a building or even underground. I was speaking about utilities. A lot of cases, your me meters are going to be underground. And then you let high attenuation of your signal, which means that you need either to boost your signal. If you agree if regulation allows you to, which means more energy, or it means that your signal won’t travel as well, and then you’d need a higher density of receivers. So gateways. Base stations, why whenever you call it. So that’s, third consideration is actually very important. So we already have How much volume? How how much data do you need to send? How often? And where are you speaking from? And as I mentioned, transmission distance. Most of the time, you should not really care about that. That’s your radio benchmark, that is communicated by connectivity specialist such as hers or others, what you care about is that the data that you need to transmit from your device is received as it should. But in a ballpark, big families of communication technologies can be split about unit communication in a personal range. So, for example, if your thing your communicating thing is a device which is a attached to a person, and that will stay with this person. So thinking this watch, for example, you can bet that you would have this kind of gateways somehow that are you going to use BLE because my smartphone is going to be a great gateway for my personal devices. But that wouldn’t be a right technology solution if I were using to monitor animals that shouldn’t have a smartphone or even people that don’t have a high, rate of technology adoption, thinking about elderly care, environments, for example, and then you maybe you need to set local gateway to receive this information. But basically, will this data be sent from a few meters, from the source, and then you are going to use ble Wi Fi, smartphone protocols Or are you going to work on something which is on unknown? I was going to say long distance to unknown distances. And then you’ll rely on operating network responsibility because you don’t really know where you would need to and maybe you cannot set those gateways yourself and then you are going to use operating network from your local operators, most probably. But as I said, the concern shouldn’t be that this protocol can send data at ten kilometers, fifty kilometers. You don’t really care. What you care about is how much is it going to cost you to use this technology. Maybe it’s great that it sent the information so far because it means you need to have only a few a limited number of gateways. So lower cost of deployment, lower cost of operations. That’s why you’d care about is how much it’s going to cost per iteration wise. So maybe either in your home operations, let’s say you deploy your lower one gateways, in your subscription because you are using either an operator or on operated CFO network. And the other concern is how much is going to cost you from an energy point of view because do you need, to use a high, power output To send this data very far, and in that case, your batteries could be depleted very fast. Are you able to send at with only a few milliwatts? That’s, again, will make a whole lot of differences in how your designer solution. So how much data, how often How how far and how fast do you need the information? Pretty often you’ll start your project thinking all your customer if you’re building something for a customer, we’ll tell you they need the information right now, and it would be real time. Real time is a very precise meaning, but even if we extend that and consider that a few milliseconds is real time, Is it really that useful? Do you have that level of safety concerns is is that business critical to receive the information? So the last data point of the event detection within milliseconds. Or can your system afford to wait a few seconds? And in some cases, if you really look closely at the business case, you realize that Well, in fact, the only thing I I need is to get the information at some point. If it’s it’s five hours, It’s actually good enough because what I’m using it is to build intelligence that is going to be processed, and going to build my whole background of information. It’s not something where I will act real time. So that’s why you have solutions that, for example, are going to be installed very remote places, farm monitoring, cattle monitoring, forestry equipment monitoring that are going to use satellite networks, the new kind of satellite IoT solution where Yes. Your device will need to wait for a satellite to pass over to collect the data, and then you need the satellite to go over ground station to deliver the data on earth, and then for this data to take a few seconds to get you to your cloud application. So long story short, you have a few hours between measurements and reception. But in a lot of cases, it’s good enough. It’s not good enough if that’s, gas valve that you need to shut down right now because the leak has been detected. In that case, you need real time. And that’s a whole lot, again, all the other choices in terms of data protocols and communication technologies. And data resilience, what do I mean with the resilience is again getting beyond the, the want to get to the need. Is it that bad or that critical if you miss of if a few of your data transmission are lost? In some cases, yes, there is high criticality and getting a one hundred percent delivery rate. And bear in mind, especially using wireless, one hundred percent doesn’t exist interference will happen, fading will happen. So you could have replay mechanism. You could, reconstruct missing data based on what you know and what you can infer from previous and future communication. So in some cases, you can afford the ninety something percent delivery rate, even lower in some cases, and that will again change your technology choice and the way you design your software. So will you implement those replay mechanism or are they built in the protocols that you decide to use? Can you, on the receiving side, reconstruct or infer missing data points, and is this approximation accurate enough for your need. And in two cases, five or four gates is good enough. I was, talking about classroom here, for example. Let’s say you send the value every ten minutes. If you miss one of it, it won’t change a lot of things. What you want is to understand the trend and of course from time to time to improve open the windows and so on, but the key thing is to have this whole trend, and you can afford missing one. It’s not going to be critical. In the true meaning of it. So that’s again, about arbitration. If you will need to achieve a hundred percent delivery rate, then that will mean extra cost in the way you build your system and you build this resiliency. And again, arbitration between how much is it going to cost and how much do I really need it and how much can I afford to miss, a data transmission every now and last but not least, power management — I already spoke a bit about it in the other points because that’s what’s behind a lot of choices you are going to make with IOT building IOT solutions? And it will depend a lot about Where is your device? Does get access to free energy? For example, I spoke about a refrigerated unit, early on. You want to monitor the condition of your refrigerated container or reefer. You can plug your device into that. The energy you are going to draw is most probably not going to change a lot of things, but that means that, okay, You can afford having something that is maybe not highly optimized in terms of energy consumption. But if your device needs to be in the wide with no human intervention and no access to free energy for months or years. Then you need to get really deep into this, rabbit hole to say it, one last time of power management and the whole life cycle of your devices. Because that means most probably at some point you’ll want to replace the battery recharge the battery of your device if it’s not on the most one. In some cases, you should have, deploy and forget devices, but that’s generating, unnecessary electronic waste and so on, and that’s not something I would advocate for. So let’s say ninety nine percent of the time, you will need to recharge or replace batteries of your devices. And the cost of doing that can tank a whole project because If you did not design, your whole system, precisely enough in terms of of power consumption. You can end up with very bad surprise with the battery that is supposed to last for years and has actually lasted half of it, ten percent of it, and then you need to send people on the field to fix that, and it’s going to cost you a lot. Or you need to recall your devices again, huge amount of loss. So that’s really critical because every time you need to do something, That’s not a technique, technology related. It’s having someone physically plug put on the charging dock or opening and replacing, the batteries of the device. That’s human effort. That’s time. And of course, in the end, that’s money. So the best thing is to be able to build predictability as accurate as possible about the life cycle of your battery and feed that into the life cycle of the device itself. It’s like thinking about a tracking solution. For example, making sure that, okay, you know that every X rotations, you need to replace batteries in the devices before they are deplete, depleted, but that way You just have to do it in a controlled environment when they when they come to base, basically, instead of having to go in the wild and fix things that are broken. And again, things to be considered is the environment of your device. The temperature it will operate on, are you going to be outdoor and very hot or very cold environment. It’s going to change the chemistry, in the end, and it’s going to have a huge impact. With your device and battery, withstand humidity, shocks, and then you get into certification. If you’re operating in harsh environments, you’ll get into a tech certification. That kind of thing that, again, will add complexity and cost to our project. And once more, on the topic of energy. So I’m not, an electrical engineer. I don’t I don’t understand a lot about electricity. If I’m being honest, But, basically, you need to, all batteries are not created equal. And you need to make sure you make the right choices of providers of battery technologies so your device will actually operate, optimally, between the, basically, the energy you will need for your continuous operation, but and then assessing the whole capacity of the battery, you need all the ability for your battery to deliver peak energy when you need to do something, either your device needs to collect physical information and very energy intensive. Could be gas monitoring that kind of stuff or communicating use a lot of energy and is pulling off. And in that case, you need to, have that in consideration as well. And last of those, consideration I listed today, is the wall. I already spoke about life cycle, but How is it going to do, be on your battery? What’s going to happen to your device along its life? And again, As often, I tend to frame that as a risk management approach, it’s not all white and black. It’s not on or off. You have to have this approach of listing the what ifs and being fully honest about them at each with each part of your solution and at each stage in the life, what if something bad happens here? What if this happened? How severe is that What’s the probability of that happening? Is it one in a million or is it highly likely to happen? And then in front of that, you will put the cost of prevention. And, basically, you have to do this assessment and to make the right decision about what are the mechanism you will need to put in place. So some example, I gave a little bit of power management, recharging, replacing, In some cases, you would just swap devices. I give the example of trackers. Maybe instead of getting into the effort of opening and replacing the batteries, we’ll just you will just remove the old one, slap a new one, making sure you update create a, identifiers in our system. And just deal with the old devices later on. But this way, the the track, the thing you’re monitoring can just, go on. Process maintenance, sort of partners. That’s just an example, but no magic here. You will need to think that your device will live in the wide. For a long time. And, basically, when you start deploying, it may feel like that’s the end of your IoT project. You made it to the end. You build a great device. You have perfect embedded software. Communication is top notch. You have the right, analysis process. You have business application, which is approved by the end users. Everybody is happy. And you can fit that at the end of your project. But, actually, same as with a software project that’s only the beginning of your issues is what’s going to happen in two months, in six months, in one year, two years, five years. You will need to upgrade your firmware. Maybe, again, maybe somebody else. Basic sensors won’t need that, but what’s in place already for you to do your device management remotely, firmware management, Can you do it remotely? Do you need to send someone on-site and to plug into the device to deliver an update? Can you afford to wait for the devices to come back to control hub, to have, local management of that, and same with battery? How much can you plan? How much can you predict? And what’s going to happen in terms of emergency action when something really needs to be fixed or updated? So that’s again thinking about not only the lab conditions, let’s say, but how is it going to behave in the wild? How is it going to behave at scale? And how is it going to behave in time? And taking into account that you only know so much about what’s going to happen. So again, risk management. What could happen if even if highly unlikely. Take time to make sure that we identified assess and make this, this decision, this risk management decision. So it’s really a critical part, part for me is what’s going to happen across the whole life cycle of your solution. And we are, reaching the end of today’s presentation. I hope it was, enjoyable for all of you. And if you have an IoT project, or you want to get into details about connectivity because we at Soracom are experts in connectivity solution be the connectivity itself, but associated platform services to help you manage your fleet of devices, manage security, manage all network level, transmission, cloud integration. Would be very happy to, to have a talk with you get into a full blown consultation if you’re deep into your project. So please reach out. You have the generic sales address at soracom.io here. Don’t be afraid, it’s sales, but that’s still people that want Only that will not just push you and try to close a deal with you in the following hour, but we’re here to hear about your actual needs, what you are looking for, and start, helping you build. Our motto is: you create and we connect. So we are here to help you do that and we’d be happy to, to have a chat with you. Thank you for listening. This webinar is part of a whole series of online content, US from us. Please take time to check out the other, videos that you have available on a different channels, and I’d be happy to see you soon for another session.