The Future of IoT Networks
David, thank you so much for coming today. To get us started, why don’t you tell me about yourself, your current position and a bit of brief history?
Sure. I am one of the technical co-founders and a co-CTO at Everactive. I’m also a professor at the University of Michigan. My research there focuses on ultra-low-power wireless integrated circuits and low-power radios. That’s actually some of the technology that we spun out of UM and into the company back in 2012. Ben Calhoun, my co-founder, is a Professor at the University of Virginia working on digital circuits and self-powered systems. That rounds out the rest of the technology that we spun out from the University of Virginia back in 2012.
Can you give us a little bit of history of the origin of IoT networks? Where did it start? What was the first step?
IoT networks first got started in niche applications, and probably in mostly industrial applications that we’ve never heard of. These include monitoring or control solutions, maybe wireless, maybe not, but a solution that you would buy from a single company, and they would come and install it in your facility, and it would do that one thing, and it would talk over their proprietary network, which might take advantage of some standards behind the scenes. But, for the most part, it would be a closed solution designed for one purpose. Examples are industrial control systems, HVAC, and security. So that’s kind of how this got started. Many times when we talk to older companies that build IoT devices today, they’ll tell you about how they started doing IoT back before it was cool. From a footprint perspective, it looked a lot like today’s IoT networks and IoT solutions, so that really helped lay the groundwork and get things started. Then as it grew, that’s where you get into benefits from things like standardizing on the wireless communication or platforms to help aggregate devices. There really hasn’t been a convergence on one single standard that’s been adopted or accepted as the IoT network protocol. Others have emerged more recently to address certain needs that IoT devices might have, where older IoT networks or other IoT networks have fallen short. So even today, it’s still a bit frothy, so to speak, there’s still a number of IoT networks that exist today and no real convergence yet.
Currently, which are the most popular IoT networks (technology/protocols)?
It really depends on the vertical we’re talking about. You can call WiFi an IoT network, many IoT devices use it, and it’s a pervasive network today. That was not originally intended for the IoT, but has been adopted by certain IoT devices. We see this a bit more in the consumer space – but not exclusively – simply because WiFi is so widespread today, it’s so pervasive. Bluetooth is also very popular for consumer IoT products, especially ones that are only required to talk to a smartphone. Cellular networks are used a little bit, but not widespread. Some solutions have adopted cellular networks, but the cost and battery lifetime tends to be prohibitive for IoT applications. So cellular IoT networks tend to be more popular for plugged in devices, or in places where you really just only have cellular coverage and there are no other options. You’ve probably heard of LoRa. LoRa is a popular IoT network that’s slowly gaining traction. LoRaWAN is an open networking protocol. It has been adopted and rolled out by private companies across the US, and they’re building those up now. So that is an up and coming IoT network. Interestingly, there’s a startup company that is deploying LoRa over satellites to provide full coverage of the globe with LoRa.
So this is LoRa communication from the satellite to the IoT device on the earth?
Yes, bi-directional, LoRa communication from a satellite to a terrestrial device.
The number of devices that this could support would be limited, so this would likely be for remote applications where you might be out in the middle of an ocean or somewhere where there’s not even cellular coverage. Amazon Sidewalk is another IoT network starting more in the consumer space, but Amazon is now looking to penetrate more industrial spaces too. Many of the Amazon Echo devices support sidewalks today. And they’re now selling dedicated gateways to stand up additional sidewalk hotspots. And then, of course, Evernet, Everactive’s IoT Network. Evernet is the only IoT network that supports batteryless devices. There are many advantages to our network that I think you don’t see in a lot of the commercial, standard or popular IoT Networks.
There is a company called Helium that uses LoRAWAN IoT network. They have been innovating on the business model side, offering incentives to hotspot operators to increase the coverage of their LoRA Network. What is your perspective on that?
Yes, that is the business model of having somebody, perhaps some good Samaritan, stand up a gateway on their property, like at their home or office, and register it with the network to receive some kickback from the IoT network to allow the LoRaWAN network to backhaul to the internet using their broadband internet service. In the case of Helium they use a LoRa hotspot, and the devices that have subscribed to Helium’s network, can then attach to that hotspot and connect to the internet riding on this good Samaritan’s internet connection. The business model there is to provide some kickback for those who stand up those gateways. And then the advantage for Helium in that example, is that they now have an additional gateway and additional coverage and could support more users on their IoT network.
What is working well, and providing value with the existing IoT networks?
If you walk the halls at CES, you see all kinds of new IoT gadgets out there. So I think, what’s working well in my opinion today, is that there is a lot of interest and a lot of growth in the IoT space. I don’t know if that’s necessarily directly driven from the presence of these IoT networks, because the space is frothy and what any given IoT device might use for communication varies based on what the device does, or on what customers it’s trying to reach. But I think there is definitely a lot of growth in that space and that is certainly evident by what we’ve seen both from CES, but also from new companies and new products that have emerged in other verticals.
The second part of your question was value and I think there are two sides to that coin. There are some IoT devices that have clearly proven that they are solving a real problem for their customers, and there’s real value for those solutions. Cheaper access to home security is a great example. Everactive’s steam trap monitoring service is another great example of delivering real value in the industrial space. On the other side of the coin, there are IoT devices that tend to be more gimmicky in terms of what they do, or maybe they’re solving a problem that doesn’t really exist, and it’s a much more gimmicky type of solution. I won’t name any here, but these tend to be obvious when you see them.
What I really appreciate are the solutions that go after real problems, and are providing real value. And again, going back to Everactive’s products as an example of that, I think that our products have done just that. We have identified real problems in the industrial environment that are costing our customers real money, but also, increasing their footprint on the environment. By adopting our solutions, they can save money and also cut back on emissions and waste in their facilities. So I think there’s certainly a lot of traction, there’s a lot of need, and there’s a lot of pull for solutions that can deliver that type of real value.
What are the challenges that the existing IoT networks are facing?
Scale is one. If you’ve ever set up an IoT device on your own network, whether it’s WiFi or Bluetooth, you’ve gone through a series of steps to get the device connected, paired and authenticated, and attached to your account, and usually multiple other steps. We tolerate this for one or two IoT devices or maybe a handful of them at a time, but imagine rolling out 1,000 that way or 10,000 IoT devices. At this scale is where you’ll see a bottleneck, not only in the barriers, or the friction to attaching those devices to the network, but also in the amount of IoT devices that the networks can support. A lot of the incumbent wireless networks are not set up to support tens of 1,000s, or hundreds of 1000s of IoT devices on them in a local area. They usually use a protocol that is more geared towards high bandwidth and streaming data to a smaller number of devices, where a small number of devices might be 10 or 100. But on the scale of the IoT, 10 or 100 is barely scratching the surface. So one challenge is reducing the friction of connecting devices or associating devices with a network. And another challenge is the number of devices that the network can support. And both of those play into the scale problem.
A separate challenge is on throughput and latency. Think of wireless networks as operating within a fixed resource, which is the wireless spectrum. This is a scarce resource, meaning it’s not growing in size. So you need to carve the available spectrum up among all of the users of that fixed bandwidth. There are different ways to carve this up. For example, the number of IoT devices attached to the network, and overall total throughput of the network, can trade off with each other. But there’s another not obvious trade off in this space, which is the latency to communicate with any one device. If you’re willing to tolerate long latencies, to communicate with any one device, then you can usually trade off latency to support a large number of devices and build a pretty efficient solution that way – spectrally efficient, I mean. But quite often, for many IoT applications you don’t want to have to wait to talk to any one of your devices. Imaging if that latency could be up to an hour. Or a day. You might want shorter latency than that, and that puts pressure on those other two elements: the overall throughput and the number of IoT devices that you can support.
So how can existing IoT networks do both: how can the support both the devices that can tolerate an hour or more of latency, and those IoT devices where you want to be able to ping them and check in every minute or second or millisecond while sharing the same wireless spectrum? I see that as another big challenge of existing IoT networks.
That’s a good overview of the different constraints and how they’re all playing to try to solve it. So that leads me to another question and maybe related to those challenges you’re talking about. What is driving the need for new generations of IoT networks?
I would say one driver for new networks is the limitations of the old IoT networks. Bluetooth was originally designed for streaming audio, and the standard has evolved to support lower bandwidth IoT devices that don’t need to stream data, perhaps a temperature sensor, and it just needs to check in periodically. But originally, that network was not designed for those types of devices. It was designed for audio headsets, streaming audio data. This required the ability to stream data to a small number of connected devices – you don’t need to pair 10,000 audio headsets to one phone. So you have incumbent wireless networks that were designed for certain applications that just don’t necessarily scale to the new applications that we now have in mind or envision for the IoT, which needs to be designed to support the challenges we’ve talked about already: Scaling, low latency, removing the battery, or reducing the power consumption. So you can get longer or infinite lifetimes for these devices, and be able to deploy them anywhere. And enabling this kind of pervasive monitoring.
Let’s switch gears here and start thinking a little bit of where this is going. What are the trends that you have been observing in the development of future IoT networks? What’s top of mind, top of the list?
Coexistence is a big one. We just talked about incumbent wireless users or wireless networks, and there are a number of them that already use these scarce spectrum resources. So how can the incumbent wireless networks coexist with the new IoT networks that are sharing the same spectrum. One way we are addressing this is by being part of the 3GPP Partnership, which is defining the cellular networks and the cellular data standards for the cellular networks.
And can you briefly explain what 3GPP is for those that don’t know?
3GPP is the international standards body that defines the standards and the protocols for the worldwide cellular communication network. So the reason your phone works here in the US, and then you can get on a flight to Japan, you get off the plane and your phone still works; is because 3GPP has defined a standard that has been adopted across the entire world for cell phones. They don’t all talk at the same frequency, each country will typically have their own bands that they open up for cellular communication, but the messages and protocols that they use, all comply with this 3GPP standard as defined by the 3GPP partnership. 3GPP is a large group of member companies, probably measured in the many hundreds to 1,000s. Members include companies, from nonprofits, governments and other stakeholders that are interested in defining what the worldwide telecom standard looks like.
For example, one of the things that they want to do when they roll out a new technology like 5G is not break 4G in the process. So there’s some amount of backwards compatibility here, and that’s what I mean, when I say coexistence. 4G and 5G obviously have to co-exist in the same spectrum. Some of the trends of future IoT networks is being able to operate in the presence of existing or incumbent users.
Other trends are ones that we’ve really talked about, which is scaling to a large number of devices and, and trying to produce a better trade off between the number of devices you can support, the throughput or data rate of those devices, and the latency to communicate to those devices. So again, those three elements are critical, and different applications will want to prioritize a different set of those three elements. So the trend is to design IoT Networks that are able to effectively and efficiently scale to support different applications that have different needs along that three point spectrum. And then overlaying all of that is power consumption, and to design IoT networks that can support ultra low power devices. For example, we’ve seen LoRaWAN adopting technologies to reduce the power of the device. WiFi, has adopted a wakeup message protocol to help reduce the power consumption of IoT devices connected over Wi Fi networks. 3GPP, Likewise, has also adopted wakeup messaging to allow IoT devices to go into a sleep state where they can be woken up with lower latency, but also with low power. And I hear Bluetooth is working on something similar, although it’s not ready yet. And then coming back to Evernet, which is really the core of our network, has an always-on receiver on every single one of our IoT devices, that’s always listening. And what that allows Evernet to do is wake those devices up with very low latency while also supporting large scale and high throughput for each one of those devices.
Where will you first expect to see the impact on new applications or use cases that could be enabled by the IoT network that solves the challenges we have discussed?
For me, honestly, it’s more of a question of where it won’t have an impact? Think about what the cell phone has done. The cell phone has pretty much impacted every vertical at this point, right? it’s probably hard to point to any business, any vertical, or any sector that hasn’t been impacted by the cell phone. That was the last major computer revolution in mobility. Now, we’re looking at IoT as the next computer revolution. Really what’s going to happen here is that computing is going to become embedded in a lot of the everyday objects we interact with, and to a larger extent, the objects we don’t.
We’ve talked to customers in the construction space, who are interested in using IoT for augmented reality to improve safety and accuracy, also tracking equipment, and tracking materials. Think about constructing a large skyscraper in the downtown area of a city, where you don’t have the luxury of having a field to store all the construction materials and equipment on-site. So all your materials are stored somewhere else off site, and it’s a big logistical nightmare to get those materials to the right spot at the right time, and being able to find them when you go looking for them to eliminate lost time. This is a great application for asset tracking and activity monitoring.
We’ve talked to customers in the smart agriculture space working in precision agriculture or indoor farming, where they monitor things like the condition of the soil, including its temperature, moisture, and acidity levels, and they also track more environmental variables like amount of rain, sun, etc.
We’ve talked to customers in the security space. Imagine a world merchandise is tagged with something that your phone can read to securely conduct a transaction. This could autonomously unlock when you purchase the item. There’s perhaps no more need for a cashier to broker that transaction, you can walk up to an item, purchase it, and it immediately becomes yours. And the device acts as a security device to prevent theft.
If these IoT devices are small enough and cheap enough, they could potentially even be integrated into the products that we buy, including things like clothing. Clothing comes to mind when I talk about walk up kiosks. Depending on the form factor of these IoT devices, they can be integrated directly into the clothes themselves.
I could keep going. I imagine a world where essentially, every vertical has been impacted by pervasive wireless battery-less computing IoT devices.
How do you see the role of government regulators in relation to the adoption of future IoT networks?
That’s a good question. One aspect is certainly that the government funds research in certain areas, and they’ve been pouring money into the IoT, because that is where the growth is. They also, to a large extent, control the access to the wireless spectrum, and that is a scarce resource that gets traded. You may have heard about some of the broadcast TV bands being opened up for communication now. So this is evolving over time, and the government has a role to play in that evolution, because they ultimately define what each band can get used for. So we might see more frequency bands that get opened up, maybe exclusively for IoT networks to cut down on some of the interference challenges that we were talking about earlier, like dealing with operating in the presence of incumbent users. Today cellular communication is done over private bands meaning the spectrum is purchased, it’s private, and only those cellular networks are allowed to communicate on those bands. We might see something similar to that for IoT Networks in the future. I think those are some of the roles that the government plays that will have a direct impact in to the future of IoT networks
On the other side of the coin, how do you see the role of community and employees as we’re evolving towards the future IoT networks? What do you think are the concerns or their interests?
As the IoT becomes more of a part of our society, it becomes almost like a psychological adoption of the IoT. There is some hesitancy to IoT devices as they become more pervasive, and the question of privacy and security comes into mind. I think there will be some intersection between the IoT technologies and society as a whole as morality and ethics come into play here. So there are some guidelines to ensure that everyone behaves as good citizens in using and accessing these new technologies.
The companies that are developing these technologies can’t put all our faith in humanity and us being good Samaritans and good citizens. Therefore, the right checks, security protocols, policies and practices to intervene when necessary will need to be in place by the companies that are developing this technology as a safeguard so the benefits outweigh the consequences.
What are the new IoT network standards that are developing? And what’s their focus?
We’ve talked about a number of them already, like LoRa currently deploying IoT networks, WiFi adding wakeup radios and Bluetooth is looking to add wakeup radios as well. These are networks that are being built out or being modified to try and accommodate the IoT.
Talking specifically about 3GPP, I think that’s an interesting one because there’s already global coverage by the cellular network. You can have a device that can roam across the globe pretty much and maintain connectivity to the cellular network. The cellular network is operated by multiple different operators or carriers, and there’s a lot of stakeholders making that cellular network work and function as a unit.
So if we look at what cellular networks are doing in the future, there is a new IoT protocol that is being introduced in 5G called Red Cap. It is initially targeting wearables and video cameras, because it offers a higher data rate than what the 4G IoT networks currently support. 4G IoT networks are more in the 100 kilobits per second, and Red Cap is going to target 10 megabits per second. So better for wearables and video cameras, but not IoT devices. So I think that’s interesting because it’s a trade off. We were talking earlier about throughput and the number of devices you can support and latency. But if we talk about the latency perspective, that really isn’t addressed by this first version of Red Cap. However, the 3GPP group is looking at adding wakeup to Red Cap, which again, is the 5G IoT protocol in order to address the latency trade off or to provide lower latency with lower power, but without compromising the throughput of those devices. So I think that’s a really interesting example because when you make a change that enables IoT devices and roll it out over the cellular network over the entire planet, now you can quickly spread IoT devices around the world, and what those devices we can do. If we try to intersect IoT devices that can communicate over those new and improved networks that are ubiquitous around the world you are opening the door for IoT to become part of global mobility and worldwide connectivity.
David, that was excellent. Appreciate the time.
Rafael, thank you for having me.