 And now, can you give a big round of applause to Alastair MacDonald for his talk, Deploying a Lora Van Network. Thank you. First, I've never really had a good day so far, and everything going well. I thought it was going well, and then I spent five minutes trying to find a knife that had him in my mouth. So, I'm just trying to lower the expectations of this talk. I'd just like to show you something. This is really boring. I'm going to press this. And, and, ta-da! In fact, we should see tomorrow. There we go. Actually, it's not very impressive, but this is actually going through, well, several cells. This will work over at the station. We've done some checking, and this will actually work over the station. This will work, well, I'd say two kilometers to five kilometers. It will work a lot further than your average Wi-Fi signal. It transmits much data, and it's called Lora Van. It's part of a low, an LP Van, low-power wide-area network, where we go, low-power wide-area network, and it's basically for connecting internet of things together. There's other versions as well. I'm going to try and speed through this. We've got time for questions, so if you want to know about other options. But this is, this is a free and open one to use. ChemTech is a company who's developed some of the, some of the technology involved, and they, they lease it out. But most of the infrastructure that we're using is open source. In fact, Lora is the, is encoding is the technology, and Lora Van is a system that moves things around, around the world. Where is the best place to stand? I'm going to stand there. It's not going to, yeah, right, okay. Well, I should get my notes out. As I said, I spent five minutes trying to find a fork that was in my mouth. There we go. So, what's going on here? This is the device, the, the low cost, you can get lots of different things. This is a Haltech module, which cost me 12 pounds. It's, it's a little bit rubbish, if I'm honest. It's got a little, little round antenna on there. The reason why is because the antenna that was on the board didn't work, so they just modded it and put it on. And I don't want to jump ahead at the next talk, but we're going to tell you how we had a very similar system problem with the, the badges. And it's, it's a bit, bit rubbish, and don't ever try and sold the pins on these, because you will break the screen. It's just that reliable at breaking it. But this is the device, which is over here. When I press the button, it's, it goes out into the ether, and it's picked up by stations. In fact, I've got a picture of the one that we've got at the other end of the site. Here. So if you've seen that, that's the one we brought along to, to pick up. There's another five, by the way, I'm not saying that we're geeks at all, but there are five stations all over this thing, and one is probably enough to fill, to cut with the entire site. So what's happening? This is sending out. It's sending it to all the things that can be picked up. It then sends it across to a network server. Now we're using a service called the Things Network. It's a three, one, it, it's getting paid for, should be paid for by selling things. I think they're having to look again at their business model, but it's free and open. You can link your station into their server, and then everybody can use this. So if you had one of these, you could be using our, our station. And you can feed multiple ones as well. And I should explain, I got into this because it's an interesting technology. And I was talking to someone from one of the digital catapults. And then something came across his desk and said, about rolling out to the Things Connected Network. Now I should point out, I'm not talking on behalf of the Things Connected here, but I'm not going to have a go because I'm being very nice. And they've been trying to break the catch-22. It's nobody's producing devices for Lora One because there's no network. And nobody's putting together a network because nobody's producing devices for Lora One. So they're saying, let's put a network in place. Good idea. Very good idea. The, the server bits, they bought a white box version. In fact, they are now moving over to the Things Network as well. So it, all the stations that they've got up already either are or very soon will be feeding me the Things Network. Anyway, so we've got one server here, you know, we feed multiple ones. And then we send it across to our server. And all my server does is it receives the thing and then sends back the time. So you can't see it, but on there it's just got a little thing saying what the time is, and that's just come back from the server because it is two-way most of the time. Right. I'm going to go over here. Finally, to look at my notes. As I said, five minutes trying to find a knife that was in my mouth. Ah, here we are. Right. What are the key features of this? Well, it's, it's called Lora because it's long range. That's not so important to me. It means you don't need as many stations. But if you're in a city or something, buildings get in the way and stuff, you, you do. What we found is that, that height is key. I mean, line of, near line of sight is best. And the higher it is, the better it is. So yeah, if you can put a station high up, you're going to get more, more coverage. The key thing, the key thing I sell about this is it's, it is low power. It is, you can have one of these in an embedded device somewhere. It will call back say once a day or something or when it's needed and it uses next no power. You can buy batteries for 10 years. I'm a bit dubious, but you can buy them for 20 years because you just, you just put the thing to sleep. If you're monitoring river levels, for example, you just get it to wake up every few hours and, and ping, send the data off, go back to sleep again. You can even have it so it only wakes up when there's an issue or something. Car parking is another one that some people are interested in because you can monitor when a car drives in. You send your signal out, say, ah, a car's parked when it drives away. You send it out to say it's gone and you don't need to run the radio the rest of the time. So that, that's its, its key advantage. What you're paying for that is it's only very small bits of data. Depending on what error correction you use, this, this, this, the encoding, you can either send about 55 bytes up to 220 but that's only if you've got really good coverage. There's a thing called a spread factor which I'll just mention in a moment which, which affects that. Right. What else have we got? Oh, it talks about module before, reliability. It's not promised to be reliable. I have to say we've, we've done various bits of mapping of where the signals should be good and bad. And if you run a simulation, I'll mention the software later on, it works really, really well. I've, I've been going around with this. I don't know why I've not been arrested for something yet but I'm going around the street with this looking at what's what and mapping out where, where the signal's good. In fact, there's, there's other people, there's definitely been people going outside at the camp doing it. In fact, I might even have it on the next slide. This was Friday. This was Friday when we had the station up and this is from a service called TTN Mapper. It's an app you run on your phone to show, to log where you are and then you have a station and just send a signal out and it works out how strong the signal is. Now it's a bit of a screw because red is good and I find that really screws in my head, especially with the station things. Do I have a cursor? Yeah, with the station things that blue is good. So colors maybe could do some work. But that was, that was Friday. Somebody had clearly gone around and that's our station there and I won't zoom in. It's a captured image. But somebody's obviously, oh, there's a station there. If it was you, come across and say hi, I want to know. In fact, if you have one of the other five stations on site, come and say hi anyway because I'd love to know what you're doing. In fact, there you go. This was taken last night where it sort of predicted where most of the stations are or were. Three one last night and three three were off. Also somebody, I don't know why I took that slide out. Somebody's sending stuff out every few seconds as it were because we, in fact, let's have a look at what's going on at this exact moment in time. This is the monitor for the station I showed you the picture of before and we've received 11,808 messages and only sent 84 because whoever's sending it is sending out, probably doing mapping or just experimenting. I don't know. I don't care. It's just data going through the network. Very small bits and you can see there the log. So if I press my button, press me button. There, that's mine. Just just been sent off. Okay, and because I'm not the strongest and quickest signal, this hasn't sent the signal back. The signal that came back from here has come back from another station. Right, okay. Now we're going to try and do something silly. Bit of audience participation. This is going to go horribly wrong. The way the encoding, the modulation works is a form of frequency shifting coding. I'm finding that. I know what I'm talking about now. Instead of just going beep, it goes boop, boop, boop, boop, kind of thing. And we can demonstrate why this is good by making beeps and doing that sound. So in a moment, we're going to do two things. I'm going to put my hands up so the sound doesn't turn the microphone off. And everybody, I want you to go beep, I'll just go beepy noises, okay? Right, let's just do a quick, quick try. I'm going to beep, beep, beep, beep. Good, right. Now I'm going to go beep, beep, beep as well and see if you can hear my beep over you. Okay, right. Hear me? This, by the way, we did try this on a smaller scale and it didn't work, so it's probably not going to work now. But I'm optimistic. I'm optimistic, right, here we go. Okay, beep. Now hopefully some of the fronts will have heard my sound over the top of everybody else because it's noticeable. And anyway, that's the way it works. You've got eight channels on the station that do that. More strikingly strange to quote, it's a ninth channel on there which just doesn't use that encoding. Sorry, this is outside the scope of this talk. If you want to know more about oddities, then catch me afterwards. Where are we? Right, so yeah. And that is another map from Mapper. This is a piece of, this is output from a radio mobile. Oh, what's the, I forgot the chap's name. VE2DBE is his call sign and he's produced a software for mapping out radio footprints and stuff. It's actually really, really useful. And I did that for here. And as you can see, the entire campsite is covered and this is how I knew we'd be able to pick it up at the station. And say these are accurate. I've been testing them out at home. We've got, what is it, Malcolm's in here who's been doing a lot of, well, I was going to say work or play, whatever way you look. He's cycling around with his bicycle doing mapping. It's brilliant. So yeah, that's what we were predicting. That's actually really useful. If you're wanting to put sites up, use this. Again, catch me afterwards. I'll put my contact details on the last slide and I'll direct you to this and you can get it to predict where things are gonna go. It is really useful. Right, that's a blank slide. What's gone wrong? Okay, I'll put the wrong one on. I don't care. I don't like slides anyway. Right, well, this was useful. I've done that. I'm gonna skip over security because I disagree with a lot of people on the security. I don't think it's that secure. Some people do. The one thing I will say is that there's actually encrypted twice. See, put that down. It's encrypted between the device and the network and between the device and the application, which should be your application for historic reasons, really. So when I press that button, it picks the data it's gonna send, it encrypts it, encrypts it again, and it sends it out and it picks it up and it goes onto the network server. Every network server you feed and every network server that has received this message decrypts it or tries to decrypt it. If it can't decrypt it, it decides it's not for it. And if it can decrypt it, then it says, ah, I understand what this is. This is useful. I should pass this on to the application I'm supposed to pass it on to. But the original model was that the network server wouldn't know what data was going through it. It would just be relaying it through. So it would then pass that encrypted onto the application server. That would then decrypt it and then you would have your data. What really happens nowadays is those last two bits have merged together and you've got your own common API to get it to your actual server. So for example, things networked by defaults, they use MQTT. So when I send something, it sends me an MQTT message and I can send one back. I'm then converting that into a JSON thing. And again, outside the scope of the talk, I can waffle on for hours. You probably noticed. Right. So that's the reason why we've got the two keys, although it's just more things to type in. But you need to activate. You need to get onto the network. You can do it the old way where you put in your device identifier, which comes from the application. This is just a short number that will say the number of your device on the thing. It's not unique, but it's unique at that point in time. And you have two keys, the two keys for the encryption and a counter, which starts at zero. And this is to stop what's called a replay attack. So if somebody picks up your signal and can capture it, they can't send exactly the same thing again. I mean, as they were using it, I don't recommend it for unlocking doors, but say you use it for unlocking a door, they could capture that, send the same thing, unlock the door. You wouldn't want that. So what it does is it has a counter which upgrades, counts up every time. And it's really annoying if you reset it because you've then got to go into the system and say, the account is going to be zero again or something. I strongly recommend using over-the-air activation. This, actually under the bonnet, it still needs those numbers. But what you have is a unique identifier for the application. A unique identifier for the device. And those are unique. Those are globally unique that nobody else has and you have a nice long key. And you think, there's more stuff. I've got to type in more. Why would I want to do that? It's just more secure. And it sends that across to the station. That goes across to the network. The network then encrypts, well, works out those keys, sends them back in a secure way. And then you have those numbers on, but every time you reset it, every time I reset that module, again, it will send out that information again. The counter will start at zero and it's a very secure cryptographic thing. Right, so one word of caution as well. If you're going to use a lot of devices, the byte order of the keys varies and on some systems they even swap over depending whether it's the application key or the, it's just a mess, okay? If it's not activating, try reversing some keys around. Most of the sites have a nice reverse round to do it and trial and error is the best way of doing it. Okay, how are we doing for time? We'll go nice and fast. Right, this is good. Well, it sort of said data limits. There's various bits of licensing around the world. In Europe, we're running on 868. I think we're only allowed to use 1% of the spectrum. The things network have a restriction on top of that. They don't really enforce it at the minute, but they probably will soon. Work on the assumption that you're going to have to send, but you've got up to 30 seconds worth of air time a day total. So if you use what's the frequency shift, sorry, FS 12, which is the really reliable slow one, then it will take a lot of time to send. So you're only going to get a small bit of data maybe once or twice a day. If you use a lower FS value, then you can send a lot more frequently, but of course it's less reliable. So if you're further away from the thing, you're going to be able to get less data through. Right, yeah, sorry, the other limits. Yeah, as I said, I was working with the digital catapults and as of when I started writing this talk about a month ago, there were no official sites up. We had brilliant coverage across Newcastle, but we did this by basically going to friends with amateur radio masks and tall buildings and saying, can we put this in here? And they said, yeah, because officially, we needed to get people to sign pictures, sign papers, we needed to get some sign off from IT people who were worried about security. And to be honest, I totally understand that. It was next to impossible to do. So things are getting better. Now sites are going live, this is brilliant. But my advice to you, if you're wanting to do something like this, try the softly, softly approach. Don't try and get people to sign the life away or agree access. A lot of people then start thinking it's like a mobile network. I'm going to make thousands of pounds by leasing this site out. It's like, yeah, in five years time, you might get next to nothing. It's just, let's say, catch 22, there's no money in a minute, and there may never be. I'm just playing around with it. Oh, the other package forwards, this is an interesting issue we had. One building, we had two stations. I'm one of the founders of Make Space Newcastle and we get on really well with the building manager. So we had a station at the top of the building, but we were in the basement and it was a bit iffy coming through six, seven floors. It did it in some places, but we installed another one in the basement and they were both on the same IP address and it started to go horribly wrong. Most of this infrastructure runs on UDP passing back across. There's a new version now, which is a TCP connection. I believe it's MQTT and TCP. There's a third one as well, but let's not get into that because I don't understand it all. And my advice is it's more effort but use the newer version. In fact, cursing was, I'm glad I made notes. Jack cursing has produced a nice bit of a thing based on the original Semtec software and that's the best packet forwarder to run on your stations. Since we've done that, we can actually connect two stations on the same IP address. For reference, the bug seems to be not in the thing. It appears to be on the network end, but anyway, that's a strongly recommend that. I'm skipping over security. Well, I think I've gone through my list. I'm sorry if that's a bit hidden myth, but hopefully we'll move on to questions now and I'm expecting some good ones. Any questions? Thank you. You're doing that, that's a simple thing. Oh, yes. We've got this cute mic, can you be a hand up and I'll clap at you? Oh, yeah, yeah. Thank you. You mentioned there are multiple speeds and bandwidths available for traffic to travel through the network. So I'm going to have to stand in front of the speaker because I can't hear you from it. Don't do that. This is going to be a problem. Right, OK, good for you. OK, so you mentioned there are different speeds of data transit through the network. Is there a mechanism for the radio part of it to back off if there's difficulty transmitting the data and receiving it? No, it's in very small blocks. I mean, the stations broadcast the data, but it's all managed through the, oh, not on PowerPoint, sorry. So the stations are fairly dumb. And when they receive the thing, they will send it out. And when they receive something from the server, sorry, they will just send it out. The early packet forwarders, they received it and sent it out blind. And the timing is critical. What I didn't go into is the response that comes back. There's two windows. There's a one-second window and there's a two-second window. The things network uses a one-second window, which is a pain because a lot of sites that have 3G is not quite fast enough, because it needs to receive the message, send it to the server. That needs to process it. It sends it back to the station and sends it out. And if you're feeding multiple networks, then it's the last one that came back is in the buffer. It could only remember one. So if you've got a lot going on from multiple networks, then one network can basically put their message over the top of another message and you've lost a message. This is improved. And in fact, on the thing I showed live, it actually said that it's been scheduled. That's because it's been sent across and it's ready to go. But no, I don't think there's any backing off. And quite often, it is being used one way anyway. So I know in Oxford area, there's a flood mapping initiative around this. What's the best other use cases for LoraWAN sensor or from an end user use case deployments that is a question because I go to a lot of meetings of trying to sell the benefits of doing this thing. And I tend to answer that depending on the audience. So for example, when I go to a gas network, it's like, well, you can monitor your switches open and stuff. And I think they're really keen on that because every time they monitor anything, they spend a fortune on a GSM and getting power to it and everything, this battery. So something like that. The car park example is one that a lot of city councils I think it's a bit of a wishy-washy answer, but I don't think really know what we can do with it. And this is another reason why I want to get more stations up to break that catch 22. So yeah, I've got random ideas. I mean, I want to be monitoring the time away metro system. Now, they are getting a new, well, they've got a new signaling system and they're supposed to be making that data available. But just having a light beam through the track or something to send data back would be great because then I could have an extra five minutes in bed just if I knew the train's running late, you know? So yeah, I'm sorry, wishy-washy answer, but to be decided, those are some examples. I've just remembered one example that we might be doing is I mentioned before the workshops in the basement, we might be putting a doorbell under back and using it. This is not the best technology for it. It's just a fun thing to do. Has anybody done a risk assessment for this? And why does nobody attach wings to it? So you mentioned that you don't think Elora One is very secure. What level are the security vulnerabilities and how could you hack this system? I'm some, I think this comes from a different perspective that I don't think it's horribly insecure and if you're just sending boring sensor data back then you're putting out into the public domain anyway. I wouldn't worry about it. It's not going to be a trivial thing to the break, but it's using the two keys. It's you're putting your trust in other services and it just doesn't, I'm going to be honest, it just doesn't feel right. I can't go up and say I can mathematically prove this is really secure. It's not using symmetric key encryption. I'm sorry, it's basically there's no public-private key, anything like that. So if you grab, if you can get hold of a device, you can get keys out of it, stuff like that. It's just not as secure as I'd like it to be. So I can't go up to someone and say, yeah, this is a secure thing. What I can say is that if you're hosting a station on a network, I can't comment on the firmware on the device, but it reminds like the Amazon Dash was hacked by sound. So I'm not saying it's impossible, but I'm saying it's highly unlikely that the broadcast side is the way to break into that. It might open up with security holes, but not from that point of view. And the data going through your network is completely encrypted, so you don't know what it is. The only way it can be found out what it is is when it arrives at the server, when it tries to decrypt it. Fortunately, because the data is so small, you can feed multiple services and it's just a very small amount of data, so it won't even have to worry about feeding it over 3G or anything like that. And I was going to say something else then, but it's just gone, because as I said, five minutes, knife, mouth. No, I shall finish answering that there. And is that okay? Any other questions? You're just scared that somebody's going to throw the thing at you, yeah? Right, I'm very new to Laurel Ann. I'd like to set one up. Where do I start? That is a good question. Join up to the Things Network would be my advice. So thethingsnetwork.org. You may well find a community around you that you might find people there already. There's certainly in the North East there's a few. And you can message people, you can talk there. As for the equipment, I don't really know. The equipment, we've got the curling stuff. It's okay. It's expensive, and I wouldn't recommend that most hobbyists get it. And there will be people better placed to answer that question. What I will say is if you're talking about doing it really seriously, don't penny pinch. The stuff we've got, it's really nice, but if they're paid twice as much from the same supplier, they'd have got something like 3.6 times the coverage. And considering getting the sites was the main problem, it would have been money well spent. So don't always go for the cheapest. If you just want to experiment and get going, you shouldn't really, but you could just get an Arduino and a module and actually set up a single station site. And if you're doing it down in a basement like we were and stuff like that, then that's cool. That's fine. And Things Network don't ban that, so they might need to start cutting down on abuse a bit. But at the minute, you can put it on, it's fine. Ttmmapper.org as well, you can sort of see how people have been doing testing as well. That's crowdsourced. And you can see all the single cell ones because they show up as red. Now, some of them, like all the ones here, will have shown up as red when only one thing was used. But yeah, the Things Network.org. I think there's, oh, there's one over there. Hello. This might be a terrible question because I'm not even sure I understand what I'm gonna ask about, but I've been reading recently about 5G networks and it occurred to me that it seems very much like a lower one network architecture. So is there any parallels to draw or is this gonna be competing in a way? Yeah, this is an interesting one. In fact, the reason why I was talking to the person in the digital catapult was because we were talking about 5G. And a lot of people see it as a halfway step to 5G. I totally disagree with that. It is a technology that stands on its own. I think it has a place after 5G comes along. I think it's very more limited than most people think. It is for sending small bits of data. I get very angry when people try to deliver firmware through this. It's not designed, I know we're trying to do it, but it's not, that's the wrong technology. So yeah, I think it has its place. The 5G stuff, there's a lot of authentication involved. It's a lot more secure, but there's gatekeepers. So now I see them as parallel technologies. I see 5G being a big success. This may or may not be a success. I don't know, I hope it is, but it's something cool to play with at the minute. I do question some of the commercial ideas with these, especially considering it's in OSM space. It's unlicensed bandwidth. So this is competing with things like door openers and a lot of remote controls and things like that. So it's very, very hard to learn a guaranteed service on this with a service level agreement. So I don't see many consumer products using this in the near future, 5G maybe. But in terms of monitoring a switch or something, this is brilliant. Okay, are we out of time or? Never mind, grab me afterwards. I think I've been ripped into the badge talk, but after the badge talk, I'll be able to find me as well. Okay.