As you’re tapping, scrolling, and swiping on your phone, you probably don’t give much thought to the fact that your apps are consuming electricity—just that they’re chipping away at your battery life.
Many of us use our phones all day, every day, rarely considering where the energy that runs them comes from, how exactly our apps consume battery life, and what happens each day as our smartphones touch dozens—or even hundreds—of other devices when we use them to connect. Because we can’t literally see our data traveling across the world, it’s easy to forget that our phones have an energy impact beyond the power they drink up when we charge them.
On its own, a smartphone has very little energy impact: estimates put it at less than a dollar of electricity per year on the high end, depending on who you ask. But the real cost is less obvious, encompassing the cell phone towers that keep you connected, the vast networks that connect them, the data centers that process your data, and so on.
Tools built into Apple’s Xcode and Google’s Android Studio within the last few years measure the energy impact of various code implementations, and warn developers in real time what the consequences of a particular choice will be. Code keeping the network connected but not sending anything, for example, now warns developers of the impact on the user’s smartphone battery. As consumers, both developers and not, we should be aware of this too.
Tracing your phone’s real energy impact isn’t a trivial task, but let’s take a quick look at the places your bits and bytes might go, and what it takes to get them there.
The mobile network
Because we can’t literally see our data traveling across the world, it’s easy to forget that our phones have an energy impact beyond the power they drink up when we charge them.
As your phone’s 4G LTE modem fires up and connects to the network, you’re consuming energy both on your device and at the receiving end: your carrier’s vast network of towers, which ensure you’re always online. Cell phone towers, or base stations, use large amounts of energy to dissipate signal over extended distances. This can make up a big chunk of your phone’s energy impact—though just how much depends on where you live.
In dense cities, where there are more base stations and higher subscriber counts, towers consume less power, but in remote locations where signal strength over longer distances is a priority, base stations can consume much higher amounts of energy as they prioritize range over speed.
Almost all of the energy impact from this process is incurred by your mobile provider at the base station, which is part of a vast network of these towers deployed across countries to ensure there are no gaps in coverage. That sprawl requires massive amounts of energy.
Despite that, network operators like Verizon, T-Mobile, and AT&T rarely used renewable energy for their cell towers until about five years ago, when almost all major U.S. mobile networks committed to using solar power and other methods of generating clean energy to boost capacity as customers increased pressure on corporations to audit their environmental impact. But most of these mobile providers didn’t release hard numbers on their energy impact until a bit more recently—they were initially incentivized to deliver ever-increasing speeds in a race to the bottom at the dawn of 3G and LTE, rather than to reduce their environmental impact.
AT&T, which uses 15.4 million megawatt hours of energy per year, is among the companies investing in alternative energy sources, focusing mainly on reducing its per-petabyte energy footprint (one petabyte, or PB, is 1,000 terabytes). Thanks to those investments, AT&T’s per-petabyte energy expenditure fell from 281 megawatt hours in 2012 to 139 megawatt hours in 2016.
In the years since the company started reporting, the energy efficiency of the hardware it uses to build its network has improved dramatically. That drop in energy consumption—a more than 50 percent reduction since it began intentionally investing in green energy—is a powerful reminder of how many pieces of the world your data touches, and how much work is left to be done.
It’s worth noting that in some places, it still seems more financially viable to use dirty energy sources, such as diesel, than it does to lay a power cable to remote locations, like at the top of a hill. (Dirty energy was the backbone of global networks until relatively recently.) But with renewable energy sources, such as solar power, becoming cheaper, and with the long-term business and political costs of dirty energy rising, some countries in the emerging mobile market are betting on solar, wind, and other clean energy sources as the way forward. India, for instance, had more than 400,000 base stations in 2013, but the majority of them used either dirty energy from the grid or relied on off-grid diesel generators. Seeing the situation elsewhere, India’s government mandated that renewable energy power 75 percent of rural towers by 2020.
Into some more networks
Mobile networks are just one hop on your data’s journey. As you fire up, say, Snapchat to take photos and send them, you’re touching a sprawling mobile network that uses electricity at an unprecedented scale to deliver your self-deleting photos.
Once your photo wings its way beyond your provider’s cell tower, it travels through the company’s vast switching network, out into interconnecting networks, and across a massive, convoluted set of devices to reach its destination: a cloud service like Amazon Web Services (AWS) or Microsoft Azure, where it’s processed for delivery and ultimately stored for the long term.
A single request could touch DNS servers, undersea cables between continents, content delivery networks, and make many, many more hops to get to the data center at the speed you’d expect.
To get to that end point successfully, your data touches dozens of servers, switches, routers, and underwater cables around the world. A single request, for example, could touch DNS servers, undersea cables between continents, content delivery networks, and make many, many more hops to get to the data center at the speed you’d expect. And once it’s delivered to the data center, the same data must travel across the globe again as it makes its way to the recipient.
Alcatel-Lucent, the French telecommunications giant acquired by Nokia in 2016, estimated in 2013 that 69 gigawatts of energy were used to power data and communication networks just to get those bits and bytes to their destinations, which is the equivalent of powering New York City 12 times over—and that was several years ago.
The average U.S. iPhone owner uses 1–2 gigabytes of LTE data per month, according to some studies. That data comes at an energy cost that’s opaque, because we can’t see it directly.
The American Council for an Energy-Efficient Economy estimated in a 2012 study that the energy impact of data in real terms is about 5.12 kilowatt hours per gigabyte of data consumed. It also uncovered where most of that cost is incurred: cloud providers.
According to that study, 48 percent of each gigabyte’s energy cost is incurred in data centers and point of presence networks, 38 percent at the end user, and just 14 percent in transit. Using data has real-world energy consequences—and cloud providers are starting to pay attention, with green energy now a key business investment.
AWS is one of the largest cloud hosting platforms in the world, raking in billions of dollars of revenue each year, with estimates of at least 1.3 million servers distributed across the globe. A single Amazon data center can use as much electricity as 25,000 homes—but the company now powers more than 40 percent of its servers with renewable energy. Depending on which Amazon data center your data packets ultimately head to—some are greener than others—it’s now likely that solar or wind energy powers them. Amazon’s renewable energy investments deliver 2.6 million megawatt hours of electricity, and more is on the horizon, with a commitment to using no less than 100 percent renewable energy in the future.
Apple, one of the world’s largest and most valuable companies, already gets more than 96 percent of its energy from renewable sources across its offices and data centers. Google estimated that it would reach 100 percent renewable energy by the end of 2017, with 2,600 megawatts of energy generated by wind and solar power—almost double that of Amazon’s already significant investments.
Microsoft, Facebook, and others have committed to doing the same, spurring cloud companies big and small into action. The same companies have pushed to make developers aware of their own energy impact while building apps, with the aforementioned new tools in Xcode, Android Studio, and other development environments measuring code’s efficiency in real time.
The impact of your personal rectangle
Consumers simply aren’t aware of what’s going on behind the scenes as they use their devices, because those devices abstract away the complexity of the networks behind the apps.
Because smartphones make it so simple to jump on a video call, fire off a photo, or upload an entire album, unlimited data plans are becoming the rule, rather than the exception. Needless to say, we’re using more data than ever.
The rise of 4G LTE—and soon 5G—unleashed a race for the fastest speeds and ever-increasing data caps, but paid little attention to the larger-scale energy impact mobile devices could have. And consumers, even if pushed, might not find it easy to grasp the scale of this energy impact, because they never see the sprawling data centers, switches, and network cables their data travels through. Consumers simply aren’t aware of what’s going on behind the scenes as they use their devices, because those devices abstract away the complexity of the networks behind the apps.
So it’s up to networks and cloud providers to push energy efficiency forward—and developers can seize the opportunity to build efficient mobile apps, too. Every megabyte has a cost, even if it’s not readily visible, but it’s easy to forget that. From the solo developer to the cloud hoster, awareness and ownership of energy efficiency is key.