 So, I'm going to talk about LoRaVan and so, by the way, this is not seen as an alternative to the normal internet that we see today, okay? LoRa cannot do that. This is something that could work in case you don't have any sort of communication mechanism, even if the telecoms fail to deliver the network. This could be an alternative because it gives you the freedom to set up your own network and it's fully independent and it doesn't need internet. Yeah, me, I'm an embedded systems engineer. I'm also an RF engineer. So, embedded system engineer for those who don't know who is an embedded system engineer, it's like we are the guys who make those tiny devices which listen to you in your home, like the Amazon Echoes. Yeah, so, we do the software for that and I'm also an RF engineer, which means I do radio planning. So, mostly I'm into electronics and I do basic electronic DIY hacking and that's how I came across LoRaVan and that's how I started working on it. So, my experience with LoRaVan was that we developed our own, we deployed the first public LoRaVan network in Kerala and we deployed it in Technofar campus, which is one of the largest technical place in Kerala and south part of India. And we deployed it in 2018 and we have crossed around 20 million packets through the network and it's a pretty interesting use case in Kerala because people are actually using LoRaVan in Kerala. And we also develop, I mean we are a small research group in Trivandrum, so we develop our own edge devices, the devices that I have carried over, it's a device that we have made ourselves and I'm also passionate about open source hardware and I believe it's a way forward, yeah. So, the device looks a bit untidy because I had to wrap it up because the moment you see an antenna, the airport authorities get worried and red light, yeah, those two things they are worried. Yeah, so before I start, let me give you an overview of the IoT ecosystem that we have in India. This is partly hyped and you might have already seen the slides, right, from NASCOM, they're expecting somewhere around 9 billion market, but I don't think it's anywhere close, so it's somewhere around 4 to 5 billion, but still that's a huge market that we're expecting. So the interesting part about IoT is that we are looking at smart utility sector as one of the main drivers and that is really important because it's important for the government as well as the consumers. So if you look at water meters, the consumers also want a smart water meter because they help you keep track of water consumption, keep track of leakage. For government also, it's important because it can monitor you and it can plan better for the next deployment, right? If they want better water connection, they can plan better. Right now, the government doesn't have data, it's just in Kerala of the record, I would say that we have around two-thirds water leaked. I mean, it's all getting leaked somewhere in some point, so there's a lot of leakage. So all those leakage can be tapped into if the government have these systems in place. So the government is really keen in implementing smart utility and it applies to all the utility sector including gas, electricity, all of that. And LoraVan is very important in the sector because it occupies a very tiny space. It's called the LPVAN space. It's called Lopover-Videry network. What it gives you is it gives you a really long network coverage but at the cost of bandwidth. So if you look at the other technologies on top, you can see Wi-Fi, 80.2.11, and you can see RFID, Bluetooth. So all of those technologies, they are really good at high bandwidth communication, but they do it at the cost of range. For a given radio power, you can only, this is actually physics, so for a given radio power, you have to lose bandwidth to gain range. So Lora observes a space where you have really good range and I'll talk about that in the later part of the slides. So it's not just about LoraVan. There are lots of technologies out there in the market which does LPVAN. So we have SIGFox, we have Bluetooth, we have InWave. So all of these technologies have their role to play. If you look at weightless, it's a very lightweight as it means. It is very lightweight but it doesn't give you the option to do a two-way communication straight away. So LoraVan gives you two-way communication, but it's not synchronous, which means I cannot do both transmission and reception, which Narraband IoT does. And Narraband IoT cannot go into mobile devices, like if you are in a car in a highway, you cannot use Narraband IoT, you have to use LTM1. So there are lots of technologies and they have their own use cases, but LoraVan is more about giving you longer coverage with very little energy, which means you have really good battery life. Yeah, so this is something which I get questions a lot. So why do we need LoraVan? Because we already have NPIOT coming up. It's because LoraVan takes around one-fifth of the power in typical use cases, around one-fourth of the power, in typical use cases, the math. Yeah, and this is important because we have seen in our use cases as well. So we have deployed Lora Soil Moucher Sensor and we have seen that over a course of one year, the battery life depreciation is around 10% and the device is transmitting for every one hour. And it's dynamic, so if there is a larger change, it will transmit in like 10 minutes, but usually it's an average of one hour and the packet size is somewhere around 30 to 35 bytes. So it's taking very little battery power and that's very important for these devices. And it also gives you really good range, by the way. So the 71,000 kilometer record, by the way, it's not connected in that. It just stopped from a satellite. So it's a hackaday post. You can go to that link and you can see it for yourself where they tried out Lora. Lora and LoraVan is different, by the way. I'll talk about that in later part. So they tried the Lora modulation scheme and they were able to reach out several thousand kilometers, 71,000 kilometers, and this is a world record, space record, yeah. And then we have LoraVan. Here, the experiment was conducted by things from the community and they were able to do 702 kilometers in Earth. So that's using hot air balloons because you cannot do that on the ground because the coverage of earth will come in the picture. So it's using the hot air balloons. And these are really theoretical values, but it is possible. And it was not possible before if you take the case of battery, if you take battery life into the system, it was not possible. And the network is really hackable, but hackable by which I mean the LoraVan Mac, for those network engineers, you might be aware of what is a Mac. So Mac is how the frames are formatted and Mac essentially forms a data link and the network layer. And it gives you frames and packets. So the Mac is hackable, so people have started building their own versions of it. And things network is a very big community and it's a community which runs LoraVan network across the world. And all you need to run a things network is a gateway and a gateway, something like this, which costs around somewhere from 1000 rupees to 10,000 rupees. So it doesn't cost too much. And if you have a gateway and you plug it to the internet, you have a things network. And it is spreading across like wildfire. So we have lots of communities across the world. So in India, we have 41 communities and that's huge, right? In Trivandrum, we have around 10 gateways full time up working, getting data and transmitting data. So I believe it's the same in Delhi and in Bangalore. And sky is not the limit. So people have actually started launching satellites into space now. These are basic satellites. And interestingly, the hardware that goes into satellite is same as this one. The same chip, the same controller. They just encase it in a enclosure so that it doesn't get affected by the solar radiation. So it's really hackable. The network is hacker friendly. And there's a telegram group as well. They're actually transmitting the packets to telegram group. So you can actually see it for yourself. So these packets are actually transmitted from a large orbit because these PICO or nanosatellites, these cannot go to a geostation orbit. They do it in a large orbit, which is typically around 80 kilometers. So if you have a low rev end device, you can actually use the satellite to hop a packet, typically in the range of like one to two bytes, over space to a ground station. And these doesn't cost much. It's just a matter of the enclosure. The enclosure might cost you around $5,000, but the hardware is like 1,000 rupees. And this works on battery, not on solar power. And the stakeholders that I have come across, so there are lots of stakeholders in India, by the way. So on the left, you might see startups. The elevator technologies is one startup that I've been working with. They do smart parking. And they are doing it in Green City project. StartLogic is a company which does smart agriculture. And Tronkart is a company which does smart water meters. So startups are really important in this case because of the very reason that we need local innovation. So in the case of Kerala, the water conception, the typical water conception in terms of monetary cost is around $500 to $600 per year, the average consumption. You cannot have a $5,000 smart water meter there. It doesn't give you the return of the investment. So the startups are supposed to come in, and they are supposed to innovate. And they are doing it. So there are lots of companies doing this. And there are communities like Lora Alliance, Things Network, and there's a project called Lora Server, which promotes communities, by the way. So these are communities which happen because of the open nature of the network. And then you have Centra, Tata Communications, IC4s. So all those agencies trying to push the boundary or to bring about a larger segment of the audience to start using Lora. And then we have companies like SEMTAC, Rackwalla, Skirling. These companies, they make the hardware possible. SEMTAC actually makes a chip, the RF chip, and Rack and Kirling, they make gateways as well as nodes. And Kirling makes gateways. So this is the ecosystem. And let me drive deeper into LoraVan. So LoraVan is essentially Lora plus LPVan. So what is LoraVan? Lora plus LPVan. So we have the Lora physical layer, which is a radio layer which takes care of converting the detailed data into the electromagnetic data. So that is a physical layer. And then we have the MAC layer, which forms a data link and then a network layer. And this is a wide area network. And it is an open network standard by the Lora Alliance. And there are regional parameters for each country. That is all released by the Lora Alliance through community participation. And then it's a start-of-test network by which I mean all the devices are supposed to talk to some other device. And that device is supposed to carry the data to some other device. So it's a start-of-stars network. It operates in a license-free band. In currently, in India, it's around 865 MHz. It starts from 865 MHz to 867 MHz. It is a license-free band by which I mean it's a free band to use, but it's regulated. So there are regulations laid out by WPC on how to use this band. And that is really important that we take that into account as well. So this is a network. So we have edge devices here, which sends data, transmits the data across to a Concentrator or a gateway. And after gateway, it's all IP-enabled. This is important. So it's all IP-enabled. And the thing about Lora is you don't need internet. You can do all of this over internet as well. That is what we were doing. So you can have devices on the edge. And you can have Concentrators. So all the devices talk to all the Concentrators. And all the Concentrators talk to a central server, which is a network server. And from the gateway, it's all IP-enabled. And everything can pass through a, usually what we do is we have a Wi-Fi back-hole or it's usually a wired connection. And then we have an application server, which takes data from a network server based on metadata. It will route the data to respective applications. And it is worth noting that all of the data, from the edge to the server, is all encrypted. So even the metadata is encrypted. So it gives you a lot of security features. So this is the radio and the Mac layer combined. On top of that, obviously, you have the application layer. So this is LoraVan in one page. So on the bottom, what we have is a Lora modulation. And on the top, what we have is a LoraLens Mac. So in India, it's IN865 to H6700. So what the radio does? So you must have all used FM radio in a car. So FM radio is a very basic form of analog communication where you modulate a signal, a carrier, based on an input. It could be a voice in this FM radio case, or it could be some data. And if it's voice, it's usually analog, which means it's an analog modulation. So you can modulate. There are several ways of modulation. You can do amplitude modulation where you have a carrier and you modulate it based on the amplitude of the data. Or you can also do, I mean, where a carrier's amplitude is modified based on the input. You have frequency modulation where the carrier's frequency is changed based on the input. So we use mostly FM, right? We don't use AM anymore. Mostly we use FM. And when it comes to digital communication, this FM is actually known as frequency shift gain. Because in digital communication, we don't need large sets of data. It's usually binary. It's 1 over 0. So you have one set of frequency for a binary 0. You have one set of frequency for binary 1. So you have a carrier. You modulate the carrier's frequency based on 1, 9, 0. And you get a modulator result. And this is frequency modulation. With Lora, it's very different. Because in Lora, you don't have a single frequency. Instead, we have a long set of frequencies. You have a start frequency and you have an end frequency, and it's called a chirp. And it's actually modulated based on a PN sequence. PN sequence is a pseudo-random sequence. So the sequence can go up and down. And if it's an up sequence, it's called an up chirp. If it's a down sequence, it's called a down chirp. So with chirps, the advantage here is that you have really good blocking immunity. So with frequency modulation, what we have is if you can block one frequency, you can maybe block the transmission. But with Lora, you have this excellent blocking immunity because you don't use a specific frequency because you are actually using an entire set of frequencies to define the up chirps and down chirps. And this is a chirp. This is how a chirp looks like. This is an up chirp, and we have down chirps here. In India, the up links and down links are supposed to happen in three channels. And this is mandatory for all the devices to have these three channels enabled. You can have more channels in between H65 to H67 megahertz band. And that depends upon your network provider. Or if you're running the network, that depends upon you to facilitate the additional channels. But the alliance what it has laid down is every device, every edge device needs to support all these channels. The three channels for up link and down link, and one dedicated channel for down link in the RX2 window. And there is one parameter that's interesting. It's called a spreading factor. And spreading factor is basically the rate at which you code the data. So with the highest spreading factor, you code it at a lower rate, which means you transmit it at a lower rate. It's very similar to what? Very similar to lecture. So if you are, if you're attending a lecture, if the lecturer is speaking on a faster pace, the receiver is not going to understand anything, right? So if it's a lower pace, the receiver might get something out of it. So that's how Lora works. So in Lora, you can do faster and lower phases, like from SF7 to SF12. SF12 means it's the lowest coding rate, which means it needs a lot of time for the data to go from the center to the receiver. With SF7, it's quite fast. It takes no time. And Lora, so in any radio communication, the range or the quality of a signal is measured in terms of link budget. With Lora, you have around 156 dB of link budget. And that is huge. So in theoretical terms, that should be somewhere on 1,000 kilometers, 1,000 to 100 kilometers. And that's huge with Lora, man. And when you compare it to SIGB and technologies, SIGB is usually around 140, 145 dB. And Lora, man, it's much more than that. So in an ideal scenario, we should see somewhere around one kilometer for every six dB of loss. So Lora, man, in any case, will give you more range than any of the technologies that you might see. So what is the Lora, man, Mac? So Lora, man, Mac is a layer that defines how the frames should be made and how the packets are to be made. So a packet is something, it's a very basic thing which has a frame and an address. So in an OSI layer, that would be the detailing layer and the network layer. Yeah, it's an open standard by Lora Alliance and it's a Star of Stars network. So what Mac lays down is how you should communicate and the duty cycle of the communication. So the Lora, man, Mac gives you three options for a transmission, which could be class A, class B or class C. Class A means you are supposed to send data and receive only when you have sent the data in the first place. You cannot receive unless you send the data. You send the data and you have two receiver windows. You can listen the data on the receiver windows or you can send another data to again start listening. In class B, in addition to the receiver channels I've mentioned, you have time synchronized channels for you to receive data. And in class three, the device is mostly on listening mode and occasionally sends data. And this is important because the project that I'm talking about lies on class C communication where devices are actually listening more often than sending data. So why do we need Lora, man? Yeah, so one of the reasons would be that it needs a little network provisioning. So you don't, you can do it yourself. You don't want someone to come and it install it. So it costs nothing to build a gateway. If you are an academician, you can buy it from academy accounts because most of these devices are in the hardware lab. So then it needs very little local provisioning by which I mean the provisioning that we need to do in the edge devices. So these devices, the edge devices, there is nothing that goes into these devices. It's just a configuration file. And that too can be automated if you have a CryptoKey inside device. So, and also it gives you the option for roaming unlike other technologies. So in Delhi I've seen there is Tata communications and there is Centra. So if you want actually can hop between the networks but I don't think Tata gives you that option right now but it should, the technology supports it. And then there's security which is end to end secure which I have mentioned before. And there is power consumption which is really low. Okay. So how do you start messaging using LoraVan? Okay. Everyone good? So do I need to reiterate any of the topics? Okay. Okay. So how to start messaging? So LoraVan network, you need a network. You already have a network. In Delhi you have plenty of LoraVan networks. You need a Bluetooth enabled device which could be a smartphone and an application which can transmit the data from your phone to the edge device. And a server application is basically an application that which can route the packets. And that's it. If you have all of this you can actually use LoraVan messaging. So yeah, this is architecture. So what you need is a phone. You need, then what you need is, sorry, first what you need is a network. You already have a network. Then you need a node that supports Lora as well as Bluetooth. Okay. And then you can make an application in, if you can, there are plenty of applications in Play Store. What we use is some application which does Bluetooth and it gives you UART or Bluetooth. Okay, so it's basically UART, communication or Bluetooth. So I send data using, so I don't have the entire system. I couldn't plug in my system. Okay, so I couldn't plug in my system so I couldn't demonstrate it because the packet forwarder runs on my system. So what I do is like, I have this node installed on my home. What this node does is it takes in data and it has a temporary buffer on it because I'm not supposed to send data every second because there's a lute cycle limitation. So what this node does is it takes in data and sends every minute to a gateway which is okay. And this gateway has a packet forwarder on it and this forwarder takes the data to a network server and the network server would then pass forward to the packet to an application server. Okay, and thankfully this is all of this runs in this box. This is our project. It's called Lora in a box which runs all of that in single box. This is nothing but a concentrator plus a Raspberry Pi. So after you do the messaging, the device would transmit the data to the network and you write rules in the network using rules engines. It could be noted and noted what it does is it look up into the lookup table which we have written and it will forward the packets to the edge device and see the devices, all of the devices are in listening mode. So if you send the data, the device is supposed to receive it. That is what is expected but you cannot send data always. That is the whole idea of this. And the moment that edge device device gets the data, the data is forwarded through UART to a mobile application. So I'm an embedded system engineer so I only know UART, I don't know Android so I haven't developed an application. So it's very simple. You just need a device address, enter the device address and a message. So typically what we do is around 100 bytes per minute and that should be enough for basic communication. We are not talking about UTF-8 or local phones or anything. You're talking about basic ASCII that Latin one. So where you have very basic symbols. I believe that's enough for an international town. You don't need local language. Yeah, and concluding remarks, I would say it's an independent network for short messages over long range and it gives you the end-to-end security. It's a license-free spectrum. You don't need to ask anyone to use it. It's law-caused. You can start from, a gate-based can start from 1000 rupees or less than 1000 rupees to somewhere around 50,000 rupees. So it depends upon what quality you want. Already available in the market. And the point I want to remind you would be fair use. So the network is open. The spectrum is open. So we need to make sure that it's used fairly. Be just to the network. We have seen cases where people have started transmitting data every second. And if you do that, what you'll do is you'll crowd the spectrum and you'll end up creating interference for some other device. So remember that it's a free license, free spectrum. So we need to make sure that there's a fair usage policy. So the network vendors like Centra and Tata Communications, they have their own fair usage policy. But if I run my network, there is no fair usage policy after implementing our network. So we don't have it. So for testing, we run, even for our cases, we do one second transmission, but we're not supposed to do that. But yeah, for testing purposes, yes. But if you plan to deploy it on a network or for communication, please make sure that the edge devices and the network comply with the fair usage policy. Yeah. And I need some help with the project. So we don't have good packet forwarders. So it's pretty basic packet forwarding using Nordred. And I don't think it's really scalable, actually. So what I need is someone who can do the mobile application, a fancy application, like Firefly. But something which is a real lightweight, something under one KB, if possible. I don't know, one KB world, under one MP. Yeah. And then I need a packet forwarder because I would love to discuss this over dinner. The packet forwarder is nothing but a data storage system and a rules engine. And this right now is all written in Nordred. We need something better. And yeah, so I can make these devices. I can make this 100,000 rupees. So these are pretty cheap to make. Also, the gateways, we can make it really cheap. And yeah, that's it. So before we open it up for questions, I'm just gonna say that I'm gonna do a flash talk on this and how to build your own pages. So what you're essentially building is paging technologies, right? I don't know how many of you here have seen an actual pager work. In the US, at least they're still there. They have their own networks. You can actually create one with this. I'm gonna do a flash talk on that. Stating that and questions. Four questions. He'll be here after this and you can ask him. Okay, so to start with, you have two devices over there. One is your gateway, one is your radio. The gateway needs an internet connection. It doesn't. It can work. The gateway then is also a radio transmitter. Yeah, the gateway has, what it does is it takes the packets and through radio obviously. And this is not a typical gateway that you see. Typical gateways are like you don't have network server running on the gateway itself. This has a network server and the app server running on the same device. So essentially what then you have there is that that's a box that can either transmit over Lora to the next gateway or it can transmit over the internet. Yeah. But you can also, you can use the internet. Yes, so we're talking about a shutdown situation where there is no network infrastructure in the sense of Wi-Fi or internet or anything. People are only dependent on mobile data and mobile data is gone. Okay. And so now it becomes then about saying that you need to be within Bluetooth range of one of these boxes to be able to transmit messages on the Lora network. Not these boxes, these devices. Yes, but that device needs to be in range of that box. So what is the distance that they can be apart? So this device right now is around, is an indoor device. So if you're installing in one block this should be enough for all the communication. So what this device does is it takes Bluetooth data and transmit over Lora. That's it. So that sounds like something that's going on in your kitchen or in your pocket. And this is going to be at some neighborhood junction. Yes. Okay. So how far apart can these two get? Yeah, so I've seen, I've shown in the slide slide. You can- No, actually 70,000 is in test conditions. Yeah. Okay. A protester on the street trying to use this, how far away can they get from one of those boxes before the transmission stops working? So this is an indoor gateway. If you have an outdoor unit, which is just like you install a better antenna and you install it outdoor, you can do at least four kilometers. Okay, so then it's completely feasible that say university network goes and installs a bunch of these boxes around their campus and students walking around with those things in their pockets continue to remain connected, albeit only text, no images, because it's extremely low bandwidth. Yeah, but you can also do images, over text. But you're talking about 100 bytes per minute. So it's extremely slow. Yeah, but there is an application called, in FROID, I believe it's image SMS. So we use that, you can actually send basic images. Yeah, it is extremely low quality images. Yeah, yeah. It's a very slow transmission. Yeah. Yeah, you can do up to five kbps, but- But it does not look like, this is what some journalists can use to file a report from the field, you know? Because you cannot send paragraphs of text. It's not an alternative to internet. As I was saying, paging systems. And by the way, what we do is usually, we connect the gateways using a Wi-Fi backhaul. And these are like, that may cost you some, because the ubiquity, long range Wi-Fi devices, somewhere around five to 10,000 rupees. So if you have that, we have done it for like eight to 10 kilometers. So trying to figure out what the vulnerability might be. Again, imagine we're in a network shutdown situation, and I'm carrying one of these devices. I thought I understood you to say that someone else, like a malevolent person, could swap the network by having their own edge devices. Edge devices need some amount of provisioning, by the way. Either you can have the keys managed using the network server and the application server, or what you can do is you can install a key, as a crypto key from Microchip. You can install a key, and it gives you a sequence of numbers, and that sequence can be matched in the network. You can put an authorization system on the network. You can flood the radio spectrum, but you cannot flood the gateway. Okay, any more questions? If I flood the network, the radio system, I can stop this transmission. Yes, but it's very hard compared to other technologies, because it uses a chip spectrum. So it's really hard to match, block all the spectrum. Okay, so intentionally flooding is one part, but how many people can have such devices? And at one place, because you're saying that you're sending out a message every minute, that means about 60 people together, there'll be a message approximately going out every second. So how many devices can be in the same vicinity, and still not be jamming each other? So I've taken 60 minutes as something, like a figure that ideally 60 minutes per, 16, one message per 60 minute. And for that, we approximate it at somewhere around 900, 2000 devices. 900, 2000 devices. Yeah, 1000, we can do 1000 devices. 16 minutes, yeah. One last final question. Okay, then we, in this chart, regarding the lookup table. So whoever owns the lookup table will be able to route where the messaging has to go. That's a vulnerability, yes. That's a vulnerability, yeah. Because network engineers have to work on this. I'm a radio engineer, so. Okay, so there is already one regulatory sort of paper by try to regulate this at some level, on machine to machine network communications, where they intensively talk about Lora as well. So do check out that paper, we'll end it here, and actually it will be available for you to talk more.