Expanding LPWAN Deployment Models: Enabling Both Large-Scale and Segmented Micro-Mesh Networks

Hello, and welcome to this morning's Tech Talk. My name is Chad Steider, and I am a senior manager of product marketing here at Silicon Labs. And today, we are going to talk about Wi-SUN. And we're going to look at Wi-SUN more today from an architectural level and what we're seeing emerging with the standard in the market today on a couple different fronts. So with me today, I will have Sami Kaislasuo and Mark Beacham, who will join me at the end as part of a panel to discuss this. Looking at the beginning, first, what is Wi-SUN? Why it matters for infrastructure IoT, the Silicon Labs Wi-SUN platform, micro mesh versus massive mesh, how you choose your architecture, and then live Q&A with the panelists. 
  
So, with that, let's look at Wi-SUN as an architecture. So when we look at IoT, we're moving from connecting devices to operating networks. So there are three tiers that dictate the set of requirements, and Wi-SUN is really built for this third tier. So the first tier being networks that are meant to operate a few devices, so tends to 50, 100 nodes max. And those can get away with direct connections, star topologies, and this is really what set the IoT world apart before we started seeing the introduction of a lot of these mesh topologies. Second tier is local mesh. And again, these are networks that tens to hundreds of nodes really geared towards localized coverage, homes, buildings, campuses, smaller sites. 
  
And then, the third tier for Wi-SUN is really that massive mesh, and that's really where Wi-SUN has found its footprint in the market up to this point. So, infrastructure scale IoT, talking about cities, utilities, regional grids, and there are Wi-SUN networks that are out supporting hundreds of thousands of nodes that are deployed today. So it really has found a lot of footprint in the massive mesh space, but we're seeing some use cases really evolve to expand it downstream into this tier two space, the local mesh space that we'll talk about as we go through the presentation. So as an overview of Wi-SUN. So it's an IPv6-based architecture built on 6LoWPAN with Ripple routing. Every node has a real internet address. 
  
And in networks, when we talk about the scales that some of these Wi-SUN networks are deployed on, having that individual IP addressability really is a key to how we've been successful in that space. It is a self-forming, self-healing mesh with multi-hop routing across large areas. This is important when we're talking about devices that are deployed, and then oftentimes don't have a human interact with them very frequently. They need to be able to recover if they fall off the network, recover if their routing path gets lost or obstructed, and that's one of the reasons why Wi-SUN has been so successful. It is a standards-based IEEE 802.15.4 spec. It uses standard FIPS 140-2 G5 certificate-based authentication. 
  
And because of the standards-based nature, it prevents vendor lock-in. So we're one of a few silicon vendors that support Wi-SUN. There are multiple stack options out there in the market. And it really allows you flexibility to choose not only the best chipset for yourself, but also gives you the flexibility to migrate as you need to. And with Wi-SUN PHN 1.1, it really introduced multiple node types. So, we have border routers in every network that are coordinating and routing the messages through the network. We have fully functioned nodes, which in Wi-SUN, those are the nodes that are routing messages to others. They can support OFDM or FSK modulations, higher data rates. Typically, those full function nodes are line-powered devices because they do have the overhead of routing and everything that goes along with the mesh formation. 
  
And then we have low-power limited function nodes. And these nodes are typically battery-powered devices, but they don't have to be. They're devices that typically are operating at lower data rates, and they don't have any of the routing overhead that you see with the fully functioning nodes. So they're paired with an FFN, and then wake up, check for messages, and then go back to sleep in a lot of cases. Infrastructure IoT, when we look at it from the big picture, it really needs multi-vendor interoperability by design, it needs to support scale, and it needs to have guaranteed longevity. And when you look at the standards that are out in the market today, Wi-SUN is really the only alternative that... 
  
or the only network topology and network standard that does this, right? So I mentioned on the earlier slide, multi-vendor by design. Because of the standards-based nature of the network, the fact that it runs the IPv6 stack, there are multiple vendors that can support this, and it really allows designers to choose the best chipset and stack for their application. On the scale side, Wi-SUN is very proven in the market. The mesh topology and routing scales horizontally. You can add nodes and other border routers without having to re-architect the entire network. And again, we have networks that are in the field, hundreds of thousands of nodes supported today that have been in the field for a really long time. 
  
And from a longevity standpoint, Wi-SUN supports OTA natively. They have multi-rate PHYs and certificate-based security really designed to outlast the deployment. And those multi-rate PHYs really enable you to scale up or scale down the data rates to each of those nodes as needed. So you can operate low bandwidth as needed, you can scale up as needed. And as the needs of your application change, you can adjust what PHY within the Wi-SUN ecosystem you're supporting. Now, when we look at how Wi-SUN compares to the alternatives, in these standards. So, I already talked about how Wi-SUN addresses multi-vendor interop, the network scale, and the operating costs, but let's look at some of these other technologies that typically align in this space. 
  
Proprietary mesh, it's really vendor defined. You have by the proprietary nature of these networks, they're defined by the network operator. Scale, hundreds to thousands. You could argue here that there is a little bit bigger scale that proprietary mesh can support as needed, and they have pretty low operating costs. Now, when you get into the cellular space, you have carrier-based interoperability. These networks do scale to millions, but the operating costs associated with each individual node are much higher than what you see with a Wi-SUN deployment. And as we get into the two massive and micro mesh network infrastructures later in the deck, you'll see you can accomplish a lot with a few border routers. 
  
And if you use those border routers as the only place with a cellular connection, your operating costs scale down pretty significantly. And then when you look at proprietary LPWAN, and in this case, I'm talking specifically about those networks like LoRa, Sigfox, some of those that are more star topology based, where they really fall flat is on the scale side. A lot of those networks have real limitations to the number of nodes you can connect to a single router. So now you're having to buy expensive border routers as you want to scale your network, and as you want to move from local deployments to regional deployments, or regional deployments to citywide deployments, you really run into some scalability issues and data rate issues. 
  
In a lot of cases, you may end up having a node that's only checking in with the gateway one or two times a day, which may be okay for some applications, but for a lot of others, you need a little more two-way communication than that. Now, taking a second and focusing on the Silicon Labs Wi-SUN platform, we really look at it from a total system level. So one platform, four layers, two SoCs that support Wi-SUN. Now, starting at the bottom level with the silicon, we have our FG25 and FG28 platforms, each optimized for Wi-SUN, but a little different part of the Wi-SUN network. We have the level two, which is our SDKs and tools. We have our Simplicity Studio ecosystem. We have standard wireless SDKs, analysis and profiling, all the reference designs that you need to go from silicon to device. 
  
And then, when you go from the SDK and tools to the next level, which is the Wi-SUN stack, we have our own in-house certified Wi-SUN stack. And one of the big things that differentiates us from our competition in that space is we have a lab in Rennes, France, where we have 1,000 nodes constantly running, testing out the robustness of not only our stack, but our silicon, how everything works together, how changing different parameters within your network configuration changes the behavior of those devices. So not only are we going through certification of the stack, but we're doing really robust testing at the network level to make sure that the expectation that you want to see in the field is what you're actually going to get. 
  
Now, diving a little deeper into FG25, this is a part that is geared for the infrastructure piece of Wi-SUN. It's up to almost two megs of flash, 512K of RAM. It supports all of the Wi-SUN PHAN 1.1 stacks, whether that's 1.1 core, which is really just an extension of PHAN 1.0, or you want to start adding in the high throughput PHYs or the LFN capability. You can do that all in FG25 with a lot of memory footprint and overhead. This has both FSK and OFDM PHYs, so we support up to the full 2.4 megabit per second that's defined in Wi-SUN, even though the part can do up to 3.6. We also support mode switch, and we have a unique feature in our parts for concurrent detection, which allows you to detect both FSK and OFDM messages at the same time. 
  
So, in a lot of these networks, you have nodes that are operating on both modulation schemes, and typically they have to send a switching packet to let the router know that they want to switch PHYs. And then there's a negotiation process that happens, and then both devices switch PHYs and communicate. With FG25, we eliminate the need for that step. We actually can detect FSK and OFDM at the same time, really allowing those multi-modulation type networks... to function as best as they can. And then on the security side, we've really leaned in with Secure Vault. So we have Secure Vault High within FG25, which really gives you hardware root of trust, key management, designed for those long deployments where you can design for the security that you need today, while also giving you some headroom to add some additional security as devices are in the field. 
  
Now, FG25 is really geared for that dual-band edge or those more FSK central nodes. So it has sub-GHz Fi's for long range and to add the Wi-SUN capability, but FG28 also supports Bluetooth. So, you can have Wi-SUN operating and use that Bluetooth connection for local configuration, for troubleshooting, anything that you need tech to device. It simplifies your overall architecture by eliminating dual chip designs. It improves the user experience with mobile commissioning and control, and it really allows you to be ready for the evolving needs of these true end devices in the network. Now, there are sub-GHz only versions of FG28, and that's really where we're seeing a lot of traction in the Wi-SUN space as well, because maybe it's a sensor node that you don't need to worry about that Bluetooth connectivity. 
  
Well, you can design a single platform that maybe I want to add Bluetooth down the road. FG28 can do both. So you can design for FSK only, and then either change to a dual-band part or turn on the Bluetooth part at another point in your design and really give you the full capability. Now when we look at the different architectures within Wi-SUN, and we're going to start by saying it's the same architecture or same protocol. It really is the same underlying network protocols and everything to go along with it, but it is a little different network architecture. And when we look at micro mesh, we're talking tens to hundreds to high hundreds of nodes in a network, really localized to buildings, campuses, sites, typically driven by a single border router. 
  
Massive mesh, we're talking about wide area. You need the resilience of multiple border routers. It's really built for infrastructure. And you have multiple border routers, redundant entry points. This is really built more for true critical infrastructure, where micro mesh may be something a little more localized. So diving a little deeper into micro mesh itself. This is an emerging use case for Wi-SUN that we're seeing in the market. The need for IP addressable 15.4 devices on sub-GHz is growing pretty significantly. We're seeing the need for better object penetration. We're seeing the need for more robust networks and everything that goes along with normal sub-GHz connectivity, but people want to be able to talk to each individual node. 
  
They want to be able to make sure that the network doesn't go down. And a good example of this is if you think EV charging stations or solar fields where we're not talking about thousands of nodes. We're talking about a relatively small number of nodes, but they're relatively critical. They can be spaced out pretty significantly. And historically, you've either had to rely on wired networks or cellular to each of those endpoints, and it becomes challenging from a scale and OpEx standpoint. So, we've seen a lot of traction for Wi-SUN to replace those cellular connectivity or replace those connectivity types within these networks, and really create these local micro meshes within the bigger picture. 
  
Massive mesh, this is where Wi-SUN has really shined for a long time, and we're still seeing a lot of growth in this space. Distributed infrastructure with thousands or tens of thousands of nodes, multiple entry points, multiple border routers. And these are for long lifecycle mission critical applications. But what we're seeing here, though, is as we're expanding and building out these massive mesh Wi-SUN networks across the globe, what you're seeing are pockets of micro mesh pop up within these massive installations. So you could have, if I go back and look at the building side of things that I showed on the previous slide, you could have a full micro mesh built up around an EV charging network on a college campus that's using Wi-SUN, and then it becomes part of a larger infrastructure deployment that's looking at solar and how we're delivering power to those ecosystems. 
  
So while they are a little different ecosystem and a little different architecture, they do fit together in the emerging landscape that we're seeing today. And when you look at how you choose this architecture, you want to look at, is a single border router sufficient? What happens if this network goes down for 10 minutes, 20 minutes, an hour? Is it something critical that needs to be able to communicate? Or maybe a single border router is sufficient. You want the flexibility and standards-based interoperability of today, but you want to look to the future and say, "Okay, maybe I do want to be able to join that infrastructure scale deployment and down the road at some point and really lean into what's coming in some of those bigger installations. 
  
Now, on the massive mesh side, thousands or hundred thousands of nodes, you have the ability to distribute border routers across a wide area, things that are mission-critical. You need security and OTA headroom. Now, the takeaway here is no matter which architecture you choose, no matter which way you go with your installation, Silicon Labs has the SoC platform, the software, and the stacks to support all of these installations. So we can give customers confidence that when they deploy, they don't have to worry about, well, what happens if this changes or that changes? You can update to the latest standard, you can update the parts, you can change your data rate and PHY. You can really lean into the needs of your network and as those needs change. 
  
Now, this is my last slide, then we're going to jump into the panel. What to remember for today. Wi-SUN isn't just a protocol, it's a scalable network architecture. It allows you to choose the topology and the scale that you need based on the application that you're trying to deploy to. Only Wi-SUN combines interoperability, scale, and longevity in one single stack. The same protocol covers micro mesh and massive mesh. It's really just the same skills, just a different deployment architecture. And FG25 and FG28 give you the silicon range for both ends of the spectrum. Now, if you need the high data rate, high bandwidth OFDM capability, that we typically see on the massive side, then use FG25 as those main router nodes, and maybe hang some FG28-based sensors off the edge of that network. 
  
If you need simple, low power, smaller meshes, use FG28 and FSK and just build on that. So with that, I'm going to bring Mark and Sami back in, or going to bring them in for the first time, and then we'll ask some questions about where they're seeing these changes in their markets and everything. So Mark and Sami, welcome. Thanks. Thank you. So I guess we'll start with you, Mark. When you look at the applications where you're seeing Wi-SUN, what are the biggest challenges that you see as people go from scaling to hundreds of nodes to thousands of nodes? And how are you seeing people adjust for that? Yeah. So I cover mainly the smart metering space. So if you look at Wi-SUN, that's typically in electric meters because it is a high power, high bandwidth kind of technology. 
  
So specifically, a lot of companies or a lot of people that are going after these bids for the smart electric meters, it's pretty easy to put together a room full of 100 nodes. Maybe a little bit harder for 1,000 nodes. You have one border router, maybe two, and you can really showcase, "Hey, I have 99.9% uptime," or whatever that is. But once you actually start deploying in the field and you don't control maybe where those border routers go, you don't really control the environment that they go in, and we see a lot of challenges. And a lot of these utility contracts to do hundreds of thousands of nodes that are out there, they come with certain service level agreements. So you have to, at these massive scales, meet these pretty tough requirements to have 90-plus percent uptime, be able to talk to any or ping any of the 100,000 potential nodes within milliseconds. 
  
And so that gets really, really difficult. And so if your underlying technology is not really robust in terms of self-forming, self-healing, being able to get around a tall building if it detects that that's not the best path, that can be a huge burden. And also, if you don't plan well enough, the critical mass density needed for a massive mesh network in terms of border routers and nodes and all the hops you need, it can be pretty painful in the field to deploy that many nodes. Cool. Thanks. And Sami, I talked a little bit about some of those emerging use cases for micro mesh. What are you seeing when you're out talking to customers as those maybe key three or four that are really embracing the micro mesh architecture? 
  
I always like to start with the example of EV charging in a parking garage, which might be underground or where other technologies like cellular is just not available, or Wi-Fi doesn't penetrate through concrete. So you need basically mesh, multi-hop mesh, with sub-gig to be able to connect multiple devices. So, that's where Wi-SUN really shines. You can have around the corner implementation where you have 20 to 50, even to 100 nodes spread around a garage, but still, this kind of a local mesh or micro mesh can withstand that. And the throughput requirements usually aren't that hard. It's more like you need to have a robust communication over multiple hops. So that's one example. Then, street lighting is one where there's been several different technologies used in the past, and Zigbee, for example. 
  
But truly, a standardized sub-gig is something that we're seeing penetrate the market more and more. And there as well, it doesn't necessarily need it to connect thousands or whole citywide into a network. It might be enough to have it in a micro mesh type of a connectivity to ensure that the local connectivity is running as expected. And then where I could see micro mesh coming to play are where there's no cellular coverage, like agricultural use cases. So you have one point of contact where you might get to internet, but you now extend that to further in your irrigation system or these kind of applications so that as well is where micro mesh shows its power. And usually the node account still stays relatively limited. 
  
Cool. Thanks. So now this next question is going to be for both of you. You guys interact with a lot of customers. You see a lot of network deployments go from concept to actually in the field. What would you guys say are some of the biggest mistakes teams make when choosing a network architecture? I can take that first. I think going from the test deployment to full-scale field, you have to design for your worst-case outcome. A lot of times you can design in your testing to say, "Hey, we seem to meet everything." But in the field, it's going to stress your sensitivity, it's going to stress your range, it's going to stress your known immunity, your multi-path, all that stuff. So you have to design well above what your test environment is currently going for in terms of the throughput that you want or the range. 
  
Also, there's a big push for this either network management kind of thing or tools to help you go from a small network to a very large network. So that's what I'm seeing within that space, but I'll ask Sami the same question. So one thing that usually makes sense when you're selecting a first deployment, you prioritize the simplicity. You might select cellular just because it's very easy to deploy, but in the long term, that might not be the right decision after all because we've seen already some cellular networks being shut down during the lifetime of a deployment. So with better planning, better fit-for-purpose selection and the architectural selection as well, you could have prolonged the lifetime of that deployment. 
  
So being very mindful of how long that network needs to be out there, the architecture, what scalability it needs to have longer term, instead of just selecting what seems to be the easiest in the beginning. Yeah, and I would add to that, I think there are a lot of times when we think about these large-scale deployments, we get fixated on long point-to-point range, right? So you see standards that are out there advertising tens of kilometers or longer point-to-point range, when in reality, a lot of these network topologies that we've talked about today, you have enough nodes to really build out a true mesh. And when you think about how you can optimize RF performance and network throughput and everything in a mesh versus those maybe really long point-to-point connections, I think that really helps when you're planning your overall deployment, right? 
  
You might need to cover 10 kilometers of total range within your network, but if you have four or five nodes at least within that span, now your point-to-point range shrinks pretty significantly and is pretty easy to address. Mark, a question to you. How do you think routing and coordination really impact real-world performance and reliability? I know you're really invested in what's happening in the India metering space. Yeah. So, how are you seeing a standard like Wi-SUN address some of those- Yeah... performance and reliability challenges? It's hugely important, right? So with the Wi-SUN standard being that whole self-forming, self-mesh kind of thing, especially let's talk about India, right? 
  
You have extremely dense populations, you have suburban populations, and then you have, we'll call them villages, the low density, but also you have the terrain, right? There's a lot of mountainous valleys and things like that in India, and you can't really control the density and the terrain that's there. And so if you don't have a super robust coordination of, "Hey, this meter is way far away and hard to get to, you'll find it or route back," then you're going to have dark spots, right? And so that's where massive mesh sort of meets micro mesh, right? In a city, you have this massive density. You could have 1,000 or 2,000 ratios per border router, but then as you get into the lower density areas, that ratio starts to go down, and you use these micro mesh pockets to get out of these valleys or to get out of these villages for your network. 
  
So the fact that Wi-SUN is both good in the high-density areas and now good in the low-density areas, you don't need to have two different solutions. In the past, people would do full cellular in the low density and Wi-SUN in the high density. But now that you can get everything back, do a micro mesh in the low density and a massive mesh in the high density with one solution, with some clever in-meter gateway or border router planning, you can just have probably the most comprehensive bid to go for some of these massive utilities because these utilities cover all these regions and they want to go with one partner that can just turnkey all of them instead of piece-wise it about. Yeah, I think that leads... 
  
Sami, did you have something you wanted to add to that? Okay. No, cool. I think that kind of leads me into the next question for both of you guys. Where do you see the potential for massive mesh and micro mesh converging over time, and how does deploying something like Wi-SUN for those micro meshes kind of help future-proof you as you look at what's coming in this infrastructure IoT space? Maybe if I'll start. So, I can see that as these larger mesh networks get rolled out, they can provide a really nice backhaul for, let's say, more localized services, whether it's EV charging at a parking lot or whether it's some kind of campus area micro mesh. So, kind of thinking it in a way that you will have a citywide network, whether it's metering, street lighting, offering that backhaul to more simplified sensing use cases or kind of extension of these micro meshes, whatever the use cases be. 
  
Yeah, I'll echo that as well. Once you get what we call kind of the canopy network, which is either metering or street lighting, then all these other use cases where the problem with networking certain things is infrastructure, right? If I want to monitor trash cans on a city block, I don't want to rent poles or sign cellular contracts or go deploy my own gateways because that upfront cost is just so high. And so it's like, hey, if I can reuse this canopy network and I'll just create a micro mesh of my trash cans and then use the backhaul of the streetlights to get that back to tell me when I need to go send someone out to go replace the, pick up trash, then that's a way to do it. So that's kind of the big convergence there. 
  
And like I mentioned, even in its own metering or streetlight network, you have different densities. And so you can kind of toy around. The nice thing about these flexible meshes is you can have a border router that's comfortable with 2,000 nodes, and you can have a border router that's comfortable with 20 nodes, and it's just flexible, and it'll form and kind of get that data back, which is really nice from a convergence. Yeah, and I think, Mark, you hit on the canopy network, right? I think these canopy networks or infrastructure networks are really what's unlocking a lot of these micro mesh and even battery-powered applications that in a way we haven't seen before, right? If you think about something like water meters, historically, they could only talk to other water meters and had to rely on the ecosystem that was set up for them. 
  
And even it was happening in electric meters as part of AMI 1.0 and some of the early AMI deployments. But now that we're moving to more standards-based, these devices don't have to be what's building out the infrastructure, right? They can just join an already existing infrastructure and unlock some use cases that maybe were not part of IoT before. Yeah. And one thing that's wonderful that was interesting is things like school campuses, right? What was happening in the past is you'd have one building, and it would kind of be this self-contained network, and everything would go to the cloud, and then you have another building, and it's self-contained. And now if you have this canopy Wi-SUN or some longer-range network over, you can actually get these buildings to start talking to each other more locally instead of piping everything. 
  
That was always the problem. You couldn't get any message from one building to the other because your 2.4 gig or even the sub-gig, it wasn't the network power. But if you have a few streetlights between the buildings, you can say, "Hey, you have an emergency over here. Let me lock the doors on this building kind of in real time or unlock the doors because we detected a fire somewhere." So it's kind of just combining everything, kind of gives you that real-time local control, which is nice. Yep. Yeah, and we got a question in the Q&A that I'll go ahead and answer real quick. So I talked about this misconception that you need super long point-to-point range for a lot of these networks. And one of the things I didn't talk about in that was the bandwidth advantages of moving to something like mesh, right? 
  
If you look at some of these super long-range point-to-point networks where everything is connecting back to a single gateway, that means as you scale devices, each device gets less and less bandwidth, right? You get less and less time on air because you're scaling from that standpoint. When you start deploying these mesh networks, you're not relying on any single node's individual kind of window that it can operate in. You now can kind of send messages from one device to another and wait for an opening and kind of route those nodes intelligently so to lower an overall network latency and increase overall performance. Last question for you guys, for Mark and Sami. Where do you see the biggest growth opportunities for Wi-SUN in the next three to five years? 
  
So, I know, Mark, you said on the metering side. Sami, you're more on kind of the industrial agriculture distributed energy side. But if you had to kind of put a finger in the air and say, in three years, we're going there, where are you seeing that growth coming from? I'd say street lighting for sure. That's kind of already starting. But beyond that, maybe it's that distributed energy, solar fields, whether it's solar trackers or even kind of the optimizers and the inverters. So I'd say that's my second guess where the growth could come. Agriculture is much more conservative, so that might take a little bit longer time to penetrate there. But I'd say that would be the third one following Yeah, for me, again, mainly focused on the metering space. 
  
For the North American and Japan markets that are a little bit more established doing their second generation of meters, Wi-SUN's already got its hooks in there and taking off. The second step for that market would be the gas and water meters. Right now, they're all proprietary or cellular, MBOT, LoRa, stuff like that. If you can switch that to Wi-SUN LF and then hop on with the electric meters, you don't have to deploy infrastructure on towers. But I would say the biggest hockey stick would be the emerging economies that are doing their first round of smart meters. The emerging economies are skipping AMR and just going right to AMI 1.0. But they don't have the length of proprietary experience that North America or Japan has, so they're starting from nothing. 
  
It makes cellular attractive, but they're looking for some sort of scalable, standardized, long-range mesh solution that they can use for their meters, and I think that's where we'll see the potential hockey stick. So India's a big one. They're doing their RDSS. The Philippines, Indonesia, all these guys are having these government-mandated programs of let's connect, 90 million meters. What technology should we choose? And if one of those countries is like, "Okay, Wi-SUN's the standard we're going to choose," then that's a massive hockey stick. You're going to get a lot of metering companies trying to come in there and bid for that. And so I lied a little bit. I did have one other question then that I want to hit on that we haven't touched on yet. 
  
And Mark, you kind of hit it with that last answer. When the first deployments of metering and first deployments of some of this infrastructure came out, they were all proprietary, and now we're seeing a huge global push to standardization. So, I'd love to hear from both of you on what do you view as the big drive to standardization and, as we've seen in some other markets and some other industries, it seems like standardization is the way to go. When we start seeing that standardization, things start taking off. So I'd love to hear from the two of you on what's making that drive for standardization and what you're seeing and where you think it goes from as we move and scale these deployments. 
  
Yeah, I think the big driver is the multi-vendor or the flexibility. Specific example would be in Brazil, you had a big meter vendor that they hitched their wagon to and were going up, and then that meter vendor decided, "You know what? This market's a little too bloody. I don't want to play in it anymore." And so they pulled out, and it left the entire program in shambles because you couldn't replace their meter, you couldn't integrate with their network, they controlled everything top to bottom. And so with things like Wi-SUN, you can be interoperable down low and say, "Okay, you don't want to play in this, then I'll go over here." And from the government-owned utilities, it allows you to have flexibility of, "Hey, I can bring up a couple local vendors of these meters and choose from them and spread it out." So the vendor flexibility is there. 
  
And just on the proprietary side, it's great to be proprietary. You can be in control of your own destiny. That's great. But there's a massive maintenance burden to maintain your proprietary network, versus just going along with the standard and keeping up with it and all the new features that come out with it and just recertifying as you go on, instead of 20 years later, this proprietary standard that not many people understand anymore, or sorry, this proprietary protocol that not many people understand anymore, so. Yeah, and related to the having multiple vendors available for standardized technology, you can start finding off-the-shelf gateways or border routers or hopefully in the future also, sensors. 
  
So you don't have to build necessarily everything from scratch for your network. Compared to a proprietary network, you would need to basically build everything yourself. So I see the ecosystem building up much more nicer with a standardized technology. And even though standards move slowly, I think getting new features and getting new use cases supported, it's easier with a group of companies put together and implementing those together gives it a longer lifetime in the end, even though the standardization itself might take time, but it gives the future-proof in that sense as the standardization body or the alliance keeps the standards alive. Yeah, and I think Sami, you actually led me to something that I hadn't talked about yet. 
  
I think the emergence of off-the-shelf modules, off-the-shelf routers, off-the-shelf management tools is something that I also think is big in this space and in these deployments. And we've got a really good partner in Digi who has their Hive routers, their XP modules that so many people are familiar with that are using our silicon and our Wi-SUN stack to help in some of these Wi-SUN deployments. So for anybody listening that is looking for a scaling partner, is looking for somebody to provide those gateways, we've got Digi linked on our Wi-SUN website, so please go and check out their Wi-SUN offering. And with that, Mark and Sami, thank you so much for joining. Ayub, thank you for assisting in the Q&A online. 
  
I appreciate everybody attending today, and I think we've got two more tech talks next month that are both Bluetooth oriented, so one will be channel sounding and one will actually be with the Bluetooth SIG. So, please attend those two tech talks next month and thanks everybody for attending today.