5G is the secret sauce to enable IoT
“Cooling out, cooling out, waiting for my lady / Got the special sauce that’s why she’s my baby” – G. Love & Special Sauce, Baby’s Got Sauce
John Donovan spoke at length about the promise of 5G wireless and the implications for IoT, with implementation by AT&T begining as early as 2019. The technology is fast, and implications for everything from home electronics to autonomous drones and robots are huge.
First, some details about 5G, courtesy of the Internationl Telecommunications Union (http://www.itu.int/en/mediacentre/Pages/2017-PR04.aspx):
-
It’s fast. Download speeds are 20 Gbps, 20x faster than the current peak rate of 1 Gbps with LTE. Importantly for IoT, Upload speeds are 10 Gbps peak.
It’s dense. 5G spec calls for support for up to 1 million connections per square kilometer.
It’s low-latency. Spec considers max latency to be 4 ms, with a subset of connections considered ultra-reliable low latency communications (URLLC) communicating at 1 ms. This is vs 20 ms for LTE. Latency is designed for a trade-off with power requirements, so certain items will respond more slowly to consume less power
Imagine a scenario: you are purchasing a smart lock for your front door. First, you need to connect it to your home wifi and install a password or something that allows it to communicate with your existing network. Second, you need some sort of advanced firewall from a provider like Fortinet to prevent malware from your connected HVAC system allowing people into your house, because the folks who make door hardware typically aren’t the top-of-the-line network security providers (https://www.fortinet.com/). Finally, you use it for a few months before a new version comes out with newer hardware that processes better and, I don’t know, let’s you assign temporary keys to your dog walker. So you throw it away or put it on a different door and buy a new one.
With 5G, the entire situation is different. When you open the lock, it’s already connected to the 5G network. You tell the company who owns the lock that you purchased it, and it knows to work with your phone. No processing is done on the lock itself, so any decisions the lock has to make or upgrades to the software take place on a GPU farm where power is inexpensive, so it never needs to be upgraded. Physical equipment becomes much more like apps, where upgrades can be pushed to the centralized server farms because latency is so low that computation doesn’t need to take place on the equipment itself.
Qualcomm expects up to $12.3 T in new goods and services to be generated by the IoT, and an increase in productivity of up to 89%
(https://www.qualcomm.com/invention/5g/economy). URLLC designation would allow autonomous drones coupled with high-speed decision-making software at a centralized server to work continuously. It’s difficult to imagine the possibilities, but it is clear that 5G coupled with IoT will be a key enabler of the fourth industrial revolution.
Works Cited
https://www.nist.gov/sites/default/files/documents/itl/antd/Vince_Park.pdf
http://www.ni.com/pdf/company/en/Trend_Watch_5G.pdf
https://qz.com/179980/the-plans-for-5g-to-power-the-internet-of-things/
https://en.wikipedia.org/wiki/Fourth_Industrial_Revolution
4 comments on “5G is the secret sauce to enable IoT”
Comments are closed.
MS&E 238A
That’s for summarizing the wonders of 5G for the world of IoT. This is a very critical technology which has to be more reliable and robust to be adapted to wide use. I am wondering what kind of solutions can enable a constant 5G connection? Also with the vast amounts mobile devices and machines are being mounted with internet connection sensors, will the consumers in the end required to pay more fees for data because we simply have tons of products connected online?
Thank you! 5G trials have been underway at some Tier 1 carriers for well over a year, and early indication is that the already high expectations will be met or exceeded. As John Donovan noted, interoperability assurance within the wireless industry is standards dependent, and when those are codified, the timeframes for various market launches will become clear. The hierarchy of infrastructure that John showed in his presentation is also fast becoming a broad reality, as local network density enhancement provided via small cells is underway alongside the every increasing capability in the handset (UE, or user equipment, in industry terminology). Video transmission in particular will consume bandwidth, and 4 Pi geometry sensors/cameras on a vehicle will be a reality. Camera resolution has a big effect, too, as we know from 4K TV. In my opinion, 5G will really remove a constraint on innovation and creativity will bring interesting, unthought of applications. I can’t wait to see it.
As a member of the IEEE committee for 802.11 and the Wi-Fi Alliance, we discuss 802.11 standards and other wireless topics on their forums frequently. Personally, I agree that 5G will revolutionize the IoT market, but I also believe that AT&T’s John Donovan has a vested interest in growing his organization through the 5G service subscriptions. While that’s not a bad thing, continuous 5G connectivity to an IoT device is not always the best approach to M2M and M2C connectivity.
According Google, and connectivity analysis of their Waymo project (Self-Driving Car) Google Developers believe that continuous connects to the cloud or cellular network present a couple of problems. First, the cost associated to continuous 5G connectivity and Second, there are several implied security threats that can occur when the vehicle is constantly connected or relies on a 5 G connection. For Example, what were to happen if someone jammed the cellular signal. (https://waymo.com/tech/)
Waymo itself, is completely autonomous in that all of the compute need for the vehicle is self-contained within the vehicle and it only sends or receives traffic to the cloud when it needs to report hazards or log events that other vehicles need to be aware of. (https://youtu.be/fbWeKhAPMig)
5G Technology and IoT
In April of 2015, Alexander Hellman’s wrote an interesting article entitled, “Why IoT Needs 5G.” In his article, he asserted that 5G engineers expect that it will be able to handle about 1000 times more mobile data than today’s cellular systems. It will also become the backbone of the Internet of Things (IoT), connecting fixed and mobile devices becoming part of a new industrial and economic revolution. (http://spectrum.ieee.org/tech-talk/computing/networks/5g-taking-stock)
In a discussion with FCC Chairman Tom Wheeler, he stated in an interview that, “This technology could theoretically dramatically increase wireless broadband speeds and throughput – up to 10 gigabits per second” and “The possibilities of 5G are very intriguing.
The technology is certainly intriguing, but even more intriguing is what it means for the future of communications.”
The Next Generation Mobile Networks Alliance defines the following requirements that a 5G standard should fulfill:
Data rates of tens of megabits per second for tens of thousands of users
Data rates of 100 megabits per second for metropolitan areas
1 Gb per second simultaneously to many workers on the same office floor
Several hundreds of thousands of simultaneous connections for wireless sensors
Spectral efficiency significantly enhanced compared to 4G
Coverage improved
Signaling efficiency enhanced
Latency reduced significantly compared to LTE
(http://ieeexplore.ieee.org/document/6815890/?section=abstract)
(http://www.techrepublic.com/article/does-the-world-really-need-5g)
After reading your article, I was very intrigued by something you said, “With 5G, the entire situation is different. When you open the lock, it’s already connected to the 5G network. You tell the company who owns the lock that you purchased it, and it knows to work with your phone. No processing is done on the lock itself, so any decisions the lock has to make or upgrades to the software take place on a GPU farm where power is inexpensive, so it never needs to be upgraded. Physical equipment becomes much more like apps, where upgrades can be pushed to the centralized server farms because latency is so low that computation doesn’t need to take place on the equipment itself.”
If we take the example you cited and look at the August Smart Lock as an example, it operates entirely off of the application. The application simply triggers a motor to lock and unlock the door according to a schedule, virtual smart key and can be assigned to a guest or on demand.
The following picture shows what the inside of a Smart Lock looks like.
(https://www.cnet.com/news/break-it-down-a-lock-with-brains/)
Just as a side note, though it’s a very innovative product, I believe this product is considerably overpriced. I made the same thing in my garage using some micro-electronics purchased on EBay for 15$. I made the outer and inner casing using a 3D printer. (https://3dprint.com/wp-content/uploads/2015/03/d22.png)
With regard to an application, I used a simple python script and that executing using HTTP. I am sure that it would not be difficult to create a nice looking App in Swift.
(https://www.google.com/#safe=active&q=Android+SDK)
(https://developer.apple.com/swift/)
Users who have LIKED this comment:
Christian, thank you very much for your detailed comments. I agree that 5G is only one side of the coin and that John, as an executive at AT&T, clearly has a dog in the fight. Re: M2M, there would need to be some on-board compute for safety reasons – and as John mentioned in his discussion cost could be significantly reduced versus a scenario where all of the compute was performed onboard.
Also agree that the smart lock is an imperfect analogy, and I believe it would have been stronger had I used a higher-cost example. There are a number of cloud robotics startups aiming to eventually provide drone operation, security robots, you name it, and perform the bulk of computations on the cloud. An advantage of this would be that it significantly reduces capital risk for customers – instead of feeling concern that their unit might be obsolete after a number of years, customers can be assured that the centralized servers will be upgraded as necessary to support operations. The unit cost of said robots would be significantly reduced, as each unit need only connect to the cloud. This is not the case today. The August smart lock, for instance, would be obsolete if phones stopped supporting bluetooth.