What is 5G? Part 1 – Evolution and the Next Generation
Welcome to our conversation series on 5G. 5G is the next generation network technology, promising exponentially higher bandwidth along with lower latency to support the vast range of consumer and commercial applications in the Internet of Things (IoT). For those developing data-intensive applications, it’s important to know when and how 5G will be deployed.
In this post, Harald Remmert, Digi Director of Engineering spoke with Scott Nelson, Founder of Reuleaux Technology. The “What is 5G?” topic grew into two related conversations. This post, Part 1, covers the evolution of the network generations leading up to 5G. Part 2 covers the adoption of 5G in consumer markets and the IoT.
In case you missed it, be sure to check out the first blog in our 5G series, “Venturing Into the Fog of 5G.”
What is 5G and How Did the “Gs” Evolve?
Scott: In this conversation we review past and present technologies that led up to where we are today and where the market is heading with 5G networks. Cellular technology continues to evolve and move toward full adoption of 5G, but this will take some time as the network build-out is a lengthy, time-consuming and expensive process.
At Digi International we are in the business of developing the supporting technologies for the commercial IoT and Industrial Internet of Things, or IIoT, which moves on a separate mission from the consumer market. My conversation with Harald reflects upon how 5G will be used in and develop with those separate paths.
Start at the Beginning – What was the First “G”?
Scott: So, Harald, today a lot of people are asking the question: What is 5G? Let’s talk through 5G in the G context. Maybe you could start at the beginning and run us through all the way to 5G so we understand what it is.
Harald: Sure. So, it started all with the first generation, 1G. This was back in the ’80s, with analog technology and moving on to 2G, which is the first digital GSM technology, also CDMA, 1xRTT. So, two competing technologies at the time, which in some areas of the world are still around.
Scott: Can you spell out acronyms like GSM and CDMA for readers?
Harald: Yes. GSM stands for Global System for Mobile, and CDMA stands for Code Division Multiple Access, which is a competing cell phone technology.
Scott: AT&T went GSM and Verizon went CDMA, right?
Harald: Right, most of the cellular world went the GSM fourte. Verizon and Sprint went with CDMA.
Scott: And Qualcomm owned CDMA, did they not?
Harald: That’s correct. Not to be confused with Wide CDMA or WCDMA, which is the technology that then was later used in 3G, and that was used jointly by pretty much all parties.
Scott: All parties? So, we’re at 2G and then…
Harald: Yes, 2G. And initially, 1G and 2G were primarily driven by handsets. So, being able to communicate over phones. 2G also offered data communication at very low speeds, but that was more for the enthusiast at the time.
Scott: Was texting first possible in 2G?
Harald: That is correct, and then also very slow data communications as well – 9600 baud, so very, very low speed.
Scott: We still have customers using our formerly 2G products, right? And they’re having to move due to the coming network shutdowns?
Harald: Yes, and we’re helping them transition from 2G and also 3G technology, onto 4G LTE and beyond.
Scott: While not labeled as such, 2G was effectively an IoT technology at the time?
Harald: Absolutely. I remember in the mid-2000’s when Digi introduced one of the first cellular routers with a 2G modem. At the time modems were super expensive and pretty big. Since then costs have come down and also the size has come down significantly.
Scott: Okay. So, back to 3G.
Harald: 3G was primarily driven by handsets again, and was really the first technology that had reasonable data speeds, into the Megabit/s range. It started out with around 1 Megabit, at its peak reached up to 30 to 40 Megabit/s. That was pretty zippy for a variety of applications both on smartphones as well as for the IoT ecosystem. Then there was an interim stepping stone between 3G and 4G. Some carriers called it 3.5G, others called it 4G already. So, this 3.5G or the initial 4G non-LTE was pretty much the higher speed variants of 3G technology.
Scott: Like 3G+?
Harald: Yes, 3G+. This was HSPA, which stands for High Speed Packet Access. It was the same technology or the same generation, but with a little marketing twist.
What is a Generation in Network Technology?
Scott: Shifting gears, let’s talk about what defines a generation when we’re talking about these technologies. How does one know when one’s truly moved from one generation to another?
Harald: The International Telecommunications Union (ITU) and its partners define the requirements and timeline for future mobile communication system. The 3rd Generation Partnership Project (3GPP) then takes these requirements and writes specifications that are bundled in a series of releases. Around every decade the ITU defines a new generation to reflect technology advancements and to service application and industry needs.
For example, the first 4G LTE release that was implementing ITU’s requirements for IMT-Advanced mobile communication systems was Release 8 (2008). This release was followed by several releases that built upon the initial release and provided more functionality. Moving on to 5G, the first 5G release was Release 15 (2018), with Release 16 on the horizon for early 2020 and Release 17 targeted for 2021.
Scott: So in other words, within a generation, mobile communication systems are compatible. But when the generation changes, you have to upgrade your hardware, correct?
Scott: So CAT-M, which is part of LTE, was able to be deployed on existing LTE hardware?
Harald: That’s correct. It could be deployed on existing LTE infrastructure with a software update.
Scott: Okay. So clearly, 5G is going to require new hardware. Maybe that’s the way to think about it. Every generation requires the network operators to go out and touch all their towers to put in new hardware.
Harald: Yes, that’s right.
Scott: And we can say that every few releases we tend to get to that point?
Harald: Yeah, that’s right. I like to think in generation syndicates. So, when you think about it, so 5G, give or take, is really going to ramp in 2020, right? 4G was really ramping in 2010, and before that 3G was ramping up around 2000s. Then you can count back.
Scott: That’s interesting. So every generation is also approximately a decade…
Harald: Yes, roughly a decade.
Scott: It’s interesting to know because a lot of people have been caught off-guard by the 2G-3G shutdown, and the technology upon which they’re presently operating is effectively 20 years old. And so they’ve got products that live that long.
Harald: So every decade, we have a new generation, right? And the previous generations overlap with that to some degree. A lot of times the switch to the new generation is driven by the operators to reuse, repurpose the spectrum. The new generations have more spectral efficiency, which means you can transmit data faster and more effectively over the network. And with more devices coming online and faster speeds required by the users both on consumer phones as well as on the IoT side, carriers have no choice but have to find additional spectrum, or they have to basically shut off one technology to reuse the spectrum of that technology.
About LTE Technology, the Cats and NB-IoT
Scott: Let’s come back to the evolution and talk about 4G. To a purist like you, true 4G is LTE, right?
Harald: LTE stands for Long-Term Evolution, and that’s really what it is – an evolution. Initial LTE devices were category (CAT) 3, which means the device can transfer up to 100 Megabits per second in the downlink, and 50 Megabits/s in the uplink. That was really the first widely adopted LTE category. And since then, we had some other marketing terms to make this a little easier. LTE-Advanced is CAT 6 and up, LTE-Advanced Pro is CAT 11 or 12 and up, and then there’s 5G.
Scott: LTE actually contains the CAT-1, CAT-M, and NB-IoT releases, correct?
Harald: That is correct, yes. Interestingly CAT-1 was also defined very early on. So this was within 3GPP Release 8, but was not deployed until much later because the initial main use case for LTE was still phones. But then later the ecosystem found that, hey, we need a less expensive radio. And for some applications that used to operate well with 2G speeds, even 3G speeds, 100 megabits was overkill. So, we needed to find a way to provide a more cost-effective solution that appeals to a wider range of products. And at that point, really the technology roles split into a high speed path and a lower end path. CAT1 was the first one, and then followed in later releases with CAT-M and NB-IoT.
Scott: And these networks which we at Digi refer to as the IoT or end-to-end networks, are characterized by much lower data rates and much lower power, so they’re appropriate for machines that don’t have much to say and then they’re easier from a power and management point of view.
Harald: Yes. And power specifically on the battery life. And so these are devices that can be battery-operated, and run for an extended period of time, so years.
What is 5G? And How Will It Deploy?
Scott: Okay, now, bring us home. What is 5G? So 5G is the next generation. What’s new about it? What is it?
Harald: It’s the next generation. Previous generations focused on speed initially, so 2G, 3G, 4G LTE, and then within LTE also we start to see diversification to support a huge number of devices for IoT. Now, 5G is the fifth generation and that adds another dimension to it, which is lower latency. So it extends the other dimensions into higher speeds and a higher number of devices, but it also expands, adds another dimension, which is lower latency. And this latency is really key for new applications around AR/VR, so you can have your goggles on.
Scott: AR is Augmented reality and VR is virtual reality.
Harald: Yes. And lower latency for vehicle applications will be critical, for example in autonomous vehicles communicating amongst themselves and with the infrastructure. So, driving the latency down helps with those communication requirements, right? Otherwise, if you have a line of cars, everything is kind of happening in 20, 30-millisecond latency, right? By the time you’re on the 10th car, right, so you have a latency that’s really long.
Scott: Yeah, you have accidents.
Harald: And you have accidents, exactly. Also for IoT, so for industrial IoT, this is really interesting because now you can have one communication infrastructure wirelessly that can be used for industrial automation. Industrial automation is very time-sensitive.
Scott: What is a good example of an industrial IoT application for 5G.
Harald: For example, you could send a signal when a motor or a conveyor needs to stop, right? That has to be split second and sometimes even single-digit millisecond timing, and 5G will be able to provide that. Now, with 5G there is a new infrastructure, so you see new network cores being built in parallel with the 4G core. You see new antennas going up – in some cases in the same antenna sites – and in a lot of cases, there are new antenna sites.
Scott: My understanding is 5G will be much more densely deployed?
Harald: That is correct. For the carriers, it’s actually an interesting migration. So initially, they’re really focused on building out the 4G network and providing coverage, and then at some point they start to shift gears and they start to work on densification of the 4G LTE network, and then they also use that. As they were seeing 5G coming, they made some forward investments to also put 5G-capable technology in so that they can use that and more easily flip a switch, and fire that up.
Scott: I’ve heard, Harald, 50 meters squared. I’ve had nodes on a 50-meter grid. Is it really that dense or is that only in certain situations? When you think about the density, what do you envision?
Harald: So, with cellular networks today, it’s actually an overlay of networks. You have your macro cells on your macro networks. Those are the ones that are operating at a very low frequency and they’re covering a long range. There you might have a cell tower every few miles, right? Then you have the small cells, so there are different flavors of that, but let’s use small cell as a generic term. Those are covering a much smaller area, and they can be up to a 50-meter square radius.
Now, these small cells typically operate at higher frequencies, and that is also one main difference between 4G LTE and 5G. 4G LTE was typically operating in what is called the low bands – frequencies between 1 and 2 GHz – or even below 1 GHz and 2 GHz. So, these frequencies penetrate walls very well, right? They reach. But it’s a very heavily used spectrum.
What’s now getting more popular with small cells is the mid-band. That is the frequency range from 2 GHz all the way to 6 GHz, and that whole range from sub-1 GHz to 6 GHz is also called sub-6. So, when you hear terms in 5G, they talk about sub-6 and they might talk about millimeter wave (mmWave).
5G mmWave is using frequencies in the 24, 28, and 39 GHz spectrum, so really, really high frequencies. And the higher the frequency, the shorter the distance a radio wave can reach. That’s when you hear about having a small cell or radio on each light pole, for example.
Scott: So I read in one of the things that you wrote recently is that one of the effects of that density will be reliability, and the reliability is going to go up two or three yards of magnitude. Is there more than just the density? Basically, you can be hooked up to many towers at the same time. What else creates reliability?
Harald: Well, part of density is obviously that you need to have a cell tower in range to be connected in the first place.
Scott: So you can see several towers at any point in time?
Harald: Exactly. If you switch towers it’s rather seamless, but there might still be a small transition time, right? So what really helps the reliability is a feature called Coordinated Multipoint (CoMP).
Harald: In the past, each cell tower basically was operating independently, and there was no communication between them on the radio side. And so radio could only be connected to one cell tower at a time. With coordinated multipoint, there’s a coordination between those going and you actually have the radio connected to multiple towers. So that if all of a sudden you enter a tunnel, right, or there’s a high rise building and you drive into that range, if the one cell tower all of a sudden disappears, the other cell tower can pick up right away and so there’s no latency.
Scott: For anyone who’s ever driven through Wisconsin, these are called tower handoffs and they often don’t happen, right?
Harald: Right, and then you have a drop.
Scott: Okay. So with 5G we can expect lower latency, higher reliability, higher bandwidth, higher frequencies. Let me come back to something that I think is interesting. I’m a physicist, so I love the word “spectrum.” Right? You use the word “spectrum” a lot and then we talk about frequencies, 6 gigahertz, sub-6 and above 6. The way I think about spectrum is a bit like television networks. So if the NBC network only had blue light and the CBS network only had red light and the ABC network only got green light, they could only broadcast in that part of the spectrum. This helps to visualize spectrum. I have to think about the whole radio spectrum in a histogram, basically, and then I take chunks of it. And you literally buy a chunk, right?
Harald: That’s correct, yeah.
Scott: So, tell me how this works. The 3GPP says, “Here’s the part of the spectrum that is going to support this release,” and in this case, it’s from sub-6 to 39, did you say, or 29?
Scott: Thirty-nine gigahertz. Then how do they allocate…I mean, they auction off, you hear about these auctioning off the spectrum. So, how do the carriers kind of buy their various spectrum pieces?
Harald: A very good question. It’s a global question, right? The ITU hosts the World Radio Conference (WRC) every four years and they look at all spectrum globally and decide what spectrum should be made available. And then, by the way, they also define what future generations should look like. So, they’re more the visionary, whereas the 3GPP is more the executive for that. And so they say that, “Oh okay, so we need to have these many nodes per square kilometer, we need to have these speeds,” and so on, right?
Scott: They figure all that out.
Harald: Yes. And they talk about spectrum more in generalities. They say, “Okay, so there’s, let’s say, 24 GHz, right? So, let’s look at that.” And then they distribute this out to the entities like the FCC in the U.S., and then the FCC works together with the government to auction that off. So a carrier, as a mobile network operator, can bid on either full spectrum or certain chunks of the spectrum and then if they win that bid, it’s billions of dollars and the spectrum is good for decades. They own that spectrum, and can use it for their service.
Scott: And nobody else can?
Harald: And nobody else can. And that is also called license spectrum.
Scott: License spectrum? That’s a good point. So then presumably, they might sell 5.95 gigahertz to Verizon and 6.05 gigahertz to some range. So they’re in the neighborhood but they have very specific frequencies at which they operate?
Scott: Thank you for all of these insights. I always enjoy our conversations. We’ll continue our series next by talking about the present state of 5G network build-out and what to expect for 5G deployment timelines.