The 5G Energy Gap - A Fatal Flaw for 5G Deployment
IEEE Future Networks Podcasts with the Experts
An IEEE Future Directions Digital Studios Production
The 5G Energy Gap – A Fatal Flaw for 5G Deployment?
As massive 5G networks are being deployed in full speed in 2020, there is a potentially fatal flaw lurking in the supporting infrastructure, the utility distribution network. The 5G Energy Gap describes the uncertain ability of the utility grid to meet load energy requirements of potentially billions of devices while maintaining grid reliability. The Internet of Things, Industrial Internet of Things, edge computing and other technologies and network trends increase the issue exponentially.
Energy Harvesting (EH) solutions can supplement or even mitigate the multitude of tiny power requirements of systems where it matters most, at the edge. Scavenging every form of physical, ambient energy from the surrounding environment, EH spares the utility grid and power plants, and is a critical factor in addressing the 5G Energy Gap,
Subject Matter Expert
Chair, INGR Energy Efficiency Working Group
Brian Walker: Welcome to the IEEE Future Networks Podcast Series, Podcast with the Experts, an IEEE Future Directions Digital Studio Production. In this episode, we hear from Brian Zahnstecher, Chair of the Energy Efficiency Working Group for the International Network Generations Roadmap, and Principle of PowerRox. Brian speaks to a potentially fatal flaw lurking in the infrastructure supporting 5G deployment, the utility distribution network. The 5G Energy Gap describes the uncertain ability of the utility grid to meet load energy requirements for potentially billions of devices while maintaining grid reliability. That said, Energy Harvesting solutions can supplement or even mitigate the multitude of tiny power requirements of systems where it matters the most at the Edge. Welcome, Brian. Thank you for taking time to contribute to the IEEE Future Networks Podcast Series. Can you start by telling us about the 5G Energy Gap issue, and why it's bringing a perfect storm to the juncture of telecoms and power?
Brian Zahnstecher: What I observe is a disconnect in the industry, particularly as all these massive 5G networks are being deployed and the disconnect is that all these tiny power devices and things that are supposed to be out there, the many billions, or even the trillions things, have a much higher what are called power cost factor, or essentially what is a multiplier on how much energy needs to be generated to support all these Edge devices. And so even though a lot of these things are supposed to be micro-power devices that when you talk about a factor of five to six orders of magnitude of energy that has to be generated to support them, and then you multiply that by the sheer scale of these number of devices, that can be supported even by locally within a region or whatever, a base station or whatever you want to call it, then it starts to really add up to be a highly disproportional demand on the actual utility grid and the power plant to support the load of all these perceptively almost negligible amount of tiny power devices that are on the edge of the network. And that disconnect between the tiny amount of power that they're consuming, you know, battery powered devices and all at the Edge, versus how much has to be generated, is what I'm referring to as the 5G Energy Gap. Now there's also this question about the perfect storm, and where that really comes from is that -- well, one nice thing about this issue is that a potential solution, right, is to actually supplement tiny bits of power on these tiny devices at the edge. And so that's where Energy Harvesting comes in and that has been a major focus area of mine for the past six or so years. And that's been kind of seen as more of an emerging type of nascent technology. And because of that finding its justification in the mainstream and real applications has been challenging from both a technical fit and a cost benefit type of analysis. But now here with the 5G Energy Gap, we have a potential real issue and now all of a sudden Energy Harvesting is a great potential solution to that gap and that issue. And so that's why I say kind of the perfect storm is the deployment of 5G, the massive uptake of IoT, and IIoT, the Industrial Internet of Things, and all these wearables and wireless sensor networks and little doohickeys and whatnot, combined with energy harvesting, and that opportunity is what I actually refer to as the perfect storm, as I see it today.
Brian Walker: So, would you say 5G is the first network deployment to impact the utility grids ability to meet load energy requirements while maintaining grid reliability?
Brian Zahnstecher: So, not directly to the best of my knowledge. So, if you're saying like, "Oh, did we see this with 4G LTE? Or 3G, or previous generations of deployments?" And I would say, no, because-- and the main reason is that nothing before was enabling just the sheer number of things, and that would be touching the edge of the network that the network had to support. So, you know, before we're talking at most-- we're talking like, you know, smartphones and things and especially in the 4G, you know, 3G to 4G LTE era, and their streaming requirements and whatnot. But not supporting a whole bunch of these tiny power things that other-- especially lower power networks may have been supporting before. So, they're even in the cellular space, there are things like LTE CAT-1, there's one called CAT-M, and NBIoT or Narrow-Band IoT, specific protocols to address from a data perspective to address all these low-power, low-bandwidth, high-latency devices. But we've never had an occasion or application or deployment of saying, ‘well, we could have 10,000 sensors or nodes or endpoints within a single spot, serviced by a single base station,’ and that's really the difference here. So, there's no, I would say, direct precedent for this Energy Gap that we talked about, but what I always see as a very good analogy is look at the deployment of PV, right, of photovoltaics and solar in Germany, I think we're talking maybe roughly ten years ago or so, where the government said, "We'll subsidize this, we really need to make a concerted effort to ensure there's high penetration of PV within the country and the consumers is a good thing to do." And what they found out is they put the stuff out there too fast, before the energy grid was able to handle it. So, in other words, you have a grid that was designed for unidirectional distribution and all of a sudden you're putting all these solar points at the end point that are now turning into a bidirectional grid to feed some of that stuff back. And because it got deployed so fast before they were really kind of prepared for it, or foresaw the impact of what happens with doing it too fast at scale on a grid that was not necessarily designed for it, is they saw massive rolling blackouts and all kinds of grid stability issues. So, to me, that's kind of the closest analogy and case study in something like this that we can learn from, and that there is precedence for. The only difference, and I believe the reason that the connection between the two hasn't been widely recognized, is again to the point about now we're talking about tiny power devices that maybe at first blush seem like a relatively negligible amount of energy as a pool even collectively, but again, when you take that power cost factor multiplier into account, that's what makes it kind of a new animal in this regard.
Brian Walker: Can you describe the flow of power from end-to-end, as well as the dynamics of the Energy Gap in 5G Deployment?
Brian Zahnstecher: Sure. So, I like to think of everything in terms of sources and loads in terms of energy. And so, this sort of applies to something that I would refer to as the power value chain, which basically follows energy from generation all the way through to the end load. So, as an example, let's say in the 5G network, an example of the blocks in this chain would be starting from generation, which is the power plant. And it doesn't matter what it is, whether it's coal-fired plant, or a renewable one, or whatever, and then that outputs some energy, which goes through a utility distribution network, which makes its way into either buildings, or let's say in the example of your edge device, your smartphone or your little wireless sensor network thing, that's serviced by a base station. So, the energy flow goes through the utility grid into the base station, and through its power amplifier and other blocks, to then output a transmitted wireless signal from its output antenna, which is then received wirelessly by the device at the edge. And then, of course, consumed in that load. So, that is the end-to-end power value chain for, say, an edge device like that on a wireless network. Now, if we said, for instance, what about a server in a data center that maybe we wanted to characterize-- the load in that is like a CPU or a memory, or something in a system that sits in a data center that's ultimately consuming the energy. So, from there it's the same blocks in the front side, you know, goes from the power plant through distribution into the building, but now within the building, it may go through some energy storage, some backups, something like battery system, UPS, something like that, uninterruptible power supply, and then it goes to the IT equipment racks where that typically AC power or voltage is then converted by some kind of front end power supply or bulk power supply or AC to DC or rectifier or silver box, whatever buzzword people like to use, and that's converted into DC that's used by the system and then there's usually a series of DC to DC regulators that will then convert from higher voltages to lower voltages that are requirements for all the system loads in ASICS such as CPU and memory and FPGAs and things like that. So, and in that case there are more conversion stages, voltage regulation converter stages and therefore every stage of whether it's conversion or distribution or whatever has some loss associated with it, and so, that would be the power value chain to go from say a CPU on a server, back to the power plant.
Brian Walker: So, do you believe the utility industry's evolution to a smarter grid will address this challenge?
Brian Zahnstecher: Yeah, I think somewhat. Certainly, every day more hooks are put in there for enabling what I refer to as intelligent power management. So that means both hardware and software hooks are enabling features in equipment that allow not only the recording of information and telemetry data, that can be aggregated up all the way from that load level, from a subsystem level, so, say from a CPU or data storage or networking stuff within the server and to the system level, to know how much the piece of equipment or the radio unit or whatever is consuming, up to even say a last level, or a regional level, or data center level. And so, it's one thing to just feed that information back up kind of as a one-way telemetry indicator, but then the other great thing is now we can take that information, perform analytic analysis on it, and then we can use it to do all kinds of great things. And that can range from enabling say a business case analysis, for instance, maybe you just want to optimize your utilization based on the dynamic price of the real time energy markets, so, knowing how much you're consuming and where regionally you're consuming it may-- and knowing how the price of energy is changing in the near term may cause you to shift some of that load from one place another as really a cross optimization of OPEX, of operating expense. But you may also take some of those analytics and realize that you can more intelligently manage the power within systems, so, you feed that information back into the system to make decisions and changes in real time the way perhaps the power is allocated either within a building or a data center or even within a system, to make sure that you're optimally maximizing your energy efficiency from the system level all the way up to perhaps the regional level, if you're-- there's a lot of energy involved.
Brian Walker: You say there are critical points in the 5G network that are ideal for techniques to optimize energy use. Where are those points?
Brian Zahnstecher: I usually look for the opportunities within the network to kind of break them down of all the network constituents and then see and prioritize them by what's the lowest hanging fruit. One, where can you have the most impact? And two, just which block in that network has the biggest impact to the global energy footprint, right? And so luckily, both those things happen to fall within the same block in the network, in my opinion, at this point, and that falls in the base stations. So, today, if you look at the global energy footprint for all ICT or telco or communication networks, however you want to refer to it, that full-power pie-- and that includes everything that touches the network from data centers to even UE’s, user equipment, which is essentially like your smartphone and everything on the edge, if you add all of that up, 60 to 80 percent of all that energy is consumed right in the base station due to the mostly horrible efficiency of the linear power amplifier in that base station. So, that says that of all the blocks in this network end-to-end, and in that power value chain, that far and away the largest consumer is the obvious place to start for focusing on improving energy efficiency in those. Now, what's also very fortunate for us, particularly as we look forward to 5G and more of a migration to small cells, and heterogenous networks or HetNets, is that as cells get smaller, their opportunity to employ intelligent power management techniques becomes better and more amenable. So, as a general rule of thumb, the smaller a device or system or whatever is, the lower its power budget will be, or its energy footprint, and, therefore, makes it more manageable. So like today, you know, we work on mostly on macro cell model, where there's a big tower that consumes many, many kilowatts and it services a greater area and number of users, and whatnot. But because of that, you can't just turn power on and off very quickly on those things, and typically it never fully goes off. But there's kind of known traffic patterns and they essentially try and adjust the energy that's being consumed in there, in that base station based on known traffic patterns so, if it's a time when it's used more, they turn up the power, and in the middle of the night, when it's not used as much, they try and turn it down. But, when you go to smaller cells and things that are talking about how the power budget is on the order of 100 watts or less, as the cells get smaller, ideally someday even getting into things like picocells, and all that, where these small cells will run on ones of watts, that you can actually turn those on and off within say milliseconds, or tens of milliseconds. So, there's a lot of great features in the kind of the 5G technical spec, which is actually the 3GPP standard, that will enable more intelligent power management and the ability to shut things on and off quicker and ultimately at the end of the day, I always say there's nothing more efficient than something that's off. And then the second most efficient thing is something that's operating at its optimal point on its efficiency load curves. So that's why I think it’s, not only the biggest consumer, but also one of the best opportunities. And as I mentioned previously, you know, the ability to apply Energy Harvesting techniques to supplement energy budgets of devices at the edge should also be hugely enabling to reducing the burden on the base station, and therefore mitigates a lot of that very high power cost factor energy requirement. And the only other thing I'd like to mention of that is for opportunities for savings within the base station is, as I mentioned before, the linear power amplifier, the LPA, is typically the worst part, the lowest efficiency part of that base station. So, the reason base stations in general consume 60 - 80 percent of the global energy footprint, is because the-- what is typically a linear power amplifier that's used -- has efficiencies at best, you're lucky if you're in the double digits. So, you know, we're talking about 10, 12 percent efficient. So, that means for your wireless transition, you know, 80 to 90 percent typically-- well, actually more like 90, 90-plus percent, of that energy is just being lost to heat in the base station, and so therefore, just-- even just a maniacal focus on improving the efficiency of power amplifiers in the base station is the overall single best place to focus on for improvement, has the most room for improvement, and also would have the most impact on the global footprint. And to that end, I know there are some folks who are really focused on that, including some that are actually involved with our initiative and even our Energy Efficiency Working Group to really move the needle in that regard.
Brian Walker: So, does power efficiency return power directly to the user?
Brian Zahnstecher: Not really. It's more in the way of mitigating, generating a lot of stuff. A lot of extra energy. So, in other words we talked about how if I have to source my energy requirement at the edge from the power plant or whatever, I have to pay that very high power cost factor that's ten to the fifth, ten to the sixth. But, if I simply can supplement some of that micropower at the edge, then at the device itself, I can reduce the burden on the base station from the amount of power it needs to transmit, and so therefore, my power cost factor at the edge is really almost nil, right? It's a factor of one, because you're self-generating it right there, and you're not paying for it upstream. So, by being able to do that, you're relieving that base station and everything upstream in the power value chain up to the grid and the power plant, and that's really the true value. Because for one, if what I'm hypothesizing about the 5G Energy Gap is true, then this isn't just about a convenience to the users, saving some load, it could be the difference between the utility grid being able to support the 5G network, or being perhaps be stabilized or at worst, going down. And so, there's a very indirect impact to everyone, whether they're using the network or not. I mean, imagine a scenario where because you deployed a whole bunch of little tiny sensor networks, and endpoints locally, that it causes the local utility grid to be destabilized or go down, and people lose power to their homes. That's not just talking about quality of service for your 5G connection, now it's, "Oh, I don't have power to the home, and now that's a really serious issue. So, even people who don't even know what 5G is, all they will know is they're getting impacted in a serious way because their power goes down. And so even if they're not aware of it, proactively addressing that issue and mitigating it with something like Energy Harvesting, does provide them that value and sparing them the hardship of losing power and all that. From a direct kind of power bill savings perspective, it's more like, again, supplementing that budget at the edge, it's just what you prevent from having to be generated, that the real benefit ripples through to the whole network. And hopefully people would also argue that there's a benefit of saving and mitigating major generation also ties directly into mitigating carbon generation, as well as all the other ancillary environmental, as well as business and technical, factors that go into that.
Brian Walker: Clearly this is a complex subject. Where can people go to learn more?
Brian Zahnstecher: Well, there's several resources. One great thing just around this 5G Energy Gap thing, in particular, there was an article that I wrote that I put in the IEEE Power Electronics Quarterly Magazine in the December issue, which is out there, and free and open to the public, which you can just go to Google and put in my last name and 5G Energy Gap, or something like that, it'll be easy to find. And I think more importantly within the IEEE Future Networks Initiative, there is a roadmapping effort and we've recently kicked off an Energy Efficiency Working Group, which is openly tasked with comprehensively putting together all this information, identifying the problems, the solutions, and all that, with a kind of three/five/ten-year roadmap outlook. So, that recently, that had just kicked off! As a matter of fact, this Friday will the first meeting of that Working Group. But initially, we have a whitepaper that we're working on putting together, that will be distributed by the initiative in the near future. So, if you also go to the International Network Generation Roadmap under the Future Networks Initiative, there's plenty of resources there, including the information about this Working Group, the first edition Roadmap materials, which are already out there, as well as the information on-- as these things come out like our whitepaper, and eventually, a full roadmap chapter in the second edition of the Roadmap, which I don't have a specific target date, but I would assume is roughly within a year's timeframe or so. And in addition to that, you know, I'm sure there's plenty of other things -- I give a lot of talks in this space, as well as do a lot of the colleagues that I work with in the Working Groups and initiatives that I've mentioned. And other than that, feel free to reach out to me! My content info is always out there, and you do a search in Google for my last name, especially anything to do with power or energy or whatever, and I'm pretty sure I'm the only Zahnstecher that will come up in that regard. Certainly, the only Brian Zahnstecher! And I'm always happy to field information about this, enable people with information, you know, evangelize and preach the awareness and points about this and energy harvesting, and I'm always happy to help get this important message out there, and internalize and understood by the industry.
Brian Walker: Thank you for listening to this edition of the IEEE Future Networks Podcast with the Experts. Discover more about the IEEE Future Networks Initiative, and inquire about participating in this effort, by visiting our web Portal at FutureNetworks.IEEE.org.