Category Archives: Google Green Blog

Updates from Google’s green team on energy efficiency, renewable energy and corporate sustainability

Six Google data centers are diverting 100% of waste from landfill

Sustainability doesn’t end with a really low PUE for our data centers. Sustainability is an important business practice we strive to incorporate into all areas of our operations. A key part of this is how resources are managed. Here we define resources as the “things” that make up our data centers—both the buildings themselves, as well as all the stuff inside. This includes the waste that is generated at a data center—it’s a resource too. The more material we can reduce and use sustainably, the more effective and efficient our operations will be.

Over the past few years we’ve started focusing downstream on what resources we’re generating via waste. We’ve been working towards zero waste to landfill at our facilities, as well as reducing the amount of waste we’re generating. Today, we are announcing a new commitment to achieve Zero Waste to Landfill for our global data center operations.

At Google, Zero Waste to Landfill means that when waste leaves our data centers, none of it goes to a landfill—100 percent is diverted to a more sustainable pathway, with no more than 10% of it going to a waste-to-energy facility, unless waste-to-energy can be proved more valuable than alternative diversion paths. Our approach is based off thestandard created by UL Environment who we partnered with to ensure the guidelines we created for our facilities were aligned and compliant with how UL defines and monitors the process.

Six of our 14 sites are achieving 100 percent diversion rates. Globally across our data center operations we are diverting at least 86 percent of waste away from landfills. At our operating data centers in Europe and APAC we have reached 100 percent diversion from landfill which currently includes a contribution from waste to energy of greater than 10 percent. These data centers include: Dublin, Ireland; Hamina, Finland; St Ghislain, Belgium; Changhua County, Taiwan and Singapore. As we continue to implement new diversion strategies and ways to design waste out altogether that percentage will decrease.

Our data center in Mayes County, Oklahoma is our first Google data center to reach Zero Waste to Landfill.

So, how did we get here, where have we had big successes? There have been a couple of themes for success. Find projects that do double duty—those that not only reduce or divert waste, but also have an added benefit, like energy savings or improved process efficiency. For example, our Mayes County data center has deployed compactors to help manage waste. Not only does it help divert waste more effectively, it also gives us accurate weight data for tracking, reduces the number of pick-ups our vendor has to make (saving us and them time and money) and is cleaner overall for the site (reducing how much janitorial work is needed).

Second, sometimes you don’t have to eliminate a waste stream or find a new diversion pathway to reduce the amount of waste, instead you can also look at extending it’s life—then you’re buying less and disposing of less. The same concepts we apply to server management, we apply to our maintenance operations to keep the data centers up and running.

Third, expect the unexpected, waste streams do not stay the same, they change and evolve over time depending on your operations. Be prepared for random new waste products and be flexible. Frequently the last 10 to 20 percent of waste diversion can be the hardest to solve, but understanding these processes is critical to success.

We’ve learned a lot along this journey and will continue to learn more—the effort certainly has not been wasteful. Zero waste to landfill requires a careful attention to the types of materials you’re generating and a deep understanding of your resource pathways. All these learnings allow us to keep pushing towards zero waste to landfill, but also to start looking upstream to add circular economy practices into our operations. Zero waste to landfill is just the first step in a long process to sustainably manage our resources throughout the entire lifecycle of our data centers.

New Renewable Energy in Georgia Reduces Cost for All Customers


Last week, the Georgia Public Services Commission approved Georgia Power’s Integrated Resource Plan, a long-term planning tool that helps to guide the company’s development strategy. We’re pleased that as a result of efforts by Google and others, the plan calls for 1,500 megawatts (MW) of new renewable development for the state as well creation of an additional 200MW program for commercial and industrial customers who wish to buy renewables more directly.

This is a big deal for a region that is still in the early stages of scaling up opportunities for renewable energy. Our utility provider Georgia Power, responding to customer demand for wind, solar, and biomass, now has almost 2,000MW of renewables online in the state, and approval for an additional 1,500MW as a result of this IRP. This is a win for clean energy advocates and all Georgia Power customers, as the renewables coming online will only be authorized if they are cheaper than Georgia Power’s existing grid power, meaning that each MW of renewables coming online will reduce the cost of energy for all customers.

The 200MW C&I purchasing program is the result of urging by Google and a consortium of national and international businesses.  We participated in the regulatory process to encourage Georgia Power to adopt more, cost effective renewables, and enable commercial and industrial customers to directly procure renewable power in the state. While the details of this program will need to be fleshed out and approved by the Georgia PSC, we are hopeful that this program will give companies like Google a scalable and sustainable way to source clean energy in Georgia.  We look forward to continue working with Georgia Power, the PSC, and other stakeholders in the development of this program and share updates on our progress.


Google Data Center in Douglas County, Georgia

New renewable energy in Georgia reduces cost for all customers

Last week, the Georgia Public Services Commission approved Georgia Power’s Integrated Resource Plan, a long-term planning tool that helps to guide the company’s development strategy. We’re pleased that as a result of efforts by Google and others, the plan calls for 1,500 megawatts (MW) of new renewable development for the state as well creation of an additional 200MW program for commercial and industrial customers who wish to buy renewables more directly.

Douglas County data center
Google Data Center in Douglas County, Georgia
This is a big deal for a region that is still in the early stages of scaling up opportunities for renewable energy. Our utility provider Georgia Power, responding to customer demand for wind, solar, and biomass, now has almost 2,000MW of renewables online in the state, and approval for an additional 1,500MW as a result of this IRP. This is a win for clean energy advocates and all Georgia Power customers, as the renewables coming online will only be authorized if they are cheaper than Georgia Power’s existing grid power, meaning that each MW of renewables coming online will reduce the cost of energy for all customers.

The 200MW C&I purchasing program is the result of urging by Google and a consortium of national and international businesses.  We participated in the regulatory process to encourage Georgia Power to adopt more, cost effective renewables, and enable commercial and industrial customers to directly procure renewable power in the state. While the details of this program will need to be fleshed out and approved by the Georgia PSC, we are hopeful that this program will give companies like Google a scalable and sustainable way to source clean energy in Georgia.  We look forward to continue working with Georgia Power, the PSC, and other stakeholders in the development of this program and share updates on our progress.

First solar-powered plane completes maiden round-the-world tour, setting 19 world records

At 4:05am local time today, an atypical plane landed on a tarmac in Abu Dhabi: Si2, a futuristic aircraft entirely powered by solar energy. It was imagined and built by the two Swiss explorers Bertrand Piccard and André Borschberg, who founded Solar Impulse to promote the use of clean energies. They set the goal of circumnavigating the world by air, powered by the sun, with no fuel or polluting emissions. Starting in 2004, it took the team more than a decade to design and proof test this unique aircraft. Si2 took off in March 2015 for a 17-leg journey, spanning over 26,000 miles and using 11,000 kWh worth of solar energy. After 510 flying hours, Si2 has set 19 world records, according to the Fédération Aéronautique Internationale (FAI), on this historic expedition. 

 
Google helped build and host Solar Impulse’s digital presence, and on the first day of their round-the-world journey, we jointly launched the #FutureIsClean initiative, a platform to encourage the world to support the adoption of necessary clean technologies.

Solar Powered Plane
We’re deeply committed to powering the world with clean energy. Our goal is 100% renewable power, and so far we've committed to purchase nearly 2.5 gigawatts of renewable energy—equivalent to taking more than 1 million cars off the road—making us the largest non-utility purchaser of renewable energy in the world. 

But commitment also comes through advocacy. That’s why in 2013, Google became the internet and technology partner of Solar Impulse: to raise awareness for what's possible with clean technology and renewable energy. Everybody could use the plane’s technologies on the ground to reduce our world’s energy consumption, save natural resources and improve our quality of life. 

A global community formed to join the #FutureIsClean movement, following the progression of the Si2 during its travel around the world on www.solarimpulse.com, and tuning in for the pilot’s conversations with the Mission Control Center in Monaco (MCC). A virtual cockpit, built with the help of Google engineers and platforms, provided the telemetrics of Si2 (altitude, speed, battery level, equipment on board, etc.) and immersed children and supporters in the technical and human challenges that Solar Impulse embarked upon. 

Today, we join the rest of the world in congratulating the Solar Impulse team for this outstanding accomplishment. Solar Impulse's pioneering spirit enabled them to push human boundaries and demonstrate that clean technologies can achieve goals we once thought were impossible. 

First solar-powered plane completes maiden round-the-world tour, setting 19 world records


Posted by Amy Atlas, Communications Lead for Google Green

At 4:05am local time today, an atypical plane landed on a tarmac in Abu Dhabi: Si2, a futuristic aircraft entirely powered by solar energy. It was imagined and built by the two Swiss explorers Bertrand Piccard and André Borschberg, who founded Solar Impulse to promote the use of clean energies. They set the goal of circumnavigating the world by air, powered by the sun, with no fuel or polluting emissions. Starting in 2004, it took the team more than a decade to design and proof test this unique aircraft. Si2 took off in March 2015 for a 17-leg journey, spanning over 26,000 miles and using 11,000 kWh worth of solar energy. After 510 flying hours, Si2 has set 19 world records, according to the Fédération Aéronautique Internationale (FAI), on this historic expedition.

Google helped build and host Solar Impulse’s digital presence, and on the first day of their round-the-world journey, we jointly launched the #FutureIsClean initiative, a platform to encourage the world to support the adoption of necessary clean technologies.

We’re deeply committed to powering the world with clean energy. Our goal is 100% renewable power, and so far we've committed to purchase nearly 2.5 gigawatts of renewable energy—equivalent to taking more than 1 million cars off the road—making us the largest non-utility purchaser of renewable energy in the world.

But commitment also comes through advocacy. That’s why in 2013, Google became the internet and technology partner of Solar Impulse: to raise awareness for what's possible with clean technology and renewable energy. Everybody could use the plane’s technologies on the ground to reduce our world’s energy consumption, save natural resources and improve our quality of life.

A global community formed to join the #FutureIsClean movement, following the progression of the Si2 during its travel around the world on www.solarimpulse.com, and tuning in for the pilot’s conversations with the Mission Control Center in Monaco (MCC). A virtual cockpit, built with the help of Google engineers and platforms, provided the telemetrics of Si2 (altitude, speed, battery level, equipment on board, etc.) and immersed children and supporters in the technical and human challenges that Solar Impulse embarked upon.

Today, we join the rest of the world in congratulating the Solar Impulse team for this outstanding accomplishment. Solar Impulse's pioneering spirit enabled them to push human boundaries and demonstrate that clean technologies can achieve goals we once thought were impossible.


DeepMind AI reduces energy used for cooling Google data centers by 40%

Posted by Rich Evans, Research Engineer, DeepMind and Jim Gao, Data Center Engineer, Google

From smartphone assistants to image recognition and translation, machine learning already helps us in our everyday lives. But it can also help us to tackle some of the world’s most challenging physical problems -- such as energy consumption.  Large-scale commercial and industrial systems like data centers consume a lot of energy, and while much has been done to stem the growth of energy use, there remains a lot more to do given the world’s increasing need for computing power.


Reducing energy usage has been a major focus for us over the past  10 years: we have built our own super-efficient servers at Google, invented more efficient ways to cool our data centers and invested heavily in green energy sources, with the goal of being powered 100 percent by renewable energy. Compared to five years ago, we now get around 3.5 times the computing power out of the same amount of energy, and we continue to make many improvements each year.


Major breakthroughs, however, are few and far between -- which is why we are excited to share that by applying DeepMind’s machine learning to our own Google data centers, we’ve managed to reduce the amount of energy we use for cooling by up to 40 percent. In any large scale energy-consuming environment, this would be a huge improvement. Given how sophisticated Google’s data centers are already, it’s a phenomenal step forward.


The implications are significant for Google’s data centers, given its potential to greatly improve energy efficiency and reduce emissions overall. This will also help other companies who run on Google’s cloud to improve their own energy efficiency. While Google is only one of many data center operators in the world, many are not powered by renewable energy as we are. Every improvement in data center efficiency reduces total emissions into our environment and with technology like DeepMind’s, we can use machine learning to consume less energy and help address one of the biggest challenges of all -- climate change.


One of the primary sources of energy use in the data center environment is cooling. Just as your laptop generates a lot of heat, our data centers -- which contain servers powering Google Search, Gmail, YouTube, etc. -- also generate a lot of heat that must be removed to keep the servers running. This cooling is typically accomplished via large industrial equipment such as pumps, chillers and cooling towers. However, dynamic environments like data centers make it difficult to operate optimally for several reasons:

  1. The equipment, how we operate that equipment, and the environment interact with each other in complex, nonlinear ways. Traditional formula-based engineering and human intuition often do not capture these interactions.

  1. The system cannot adapt quickly to internal or external changes (like the weather). This is because we cannot come up with rules and heuristics for every operating scenario.

  1. Each data center has a unique architecture and environment. A custom-tuned model for one system may not be applicable to another. Therefore, a general intelligence framework is needed to understand the data center’s interactions.

To address this problem, we began applying machine learning two years ago to operate our data centers more efficiently. And over the past few months, DeepMind researchers began working with Google’s data center team to significantly improve the system’s utility. Using a system of neural networks trained on different operating scenarios and parameters within our data centers, we created a more efficient and adaptive framework to understand data center dynamics and optimize efficiency.


We accomplished this by taking the historical data that had already been collected by thousands of sensors within the data center -- data such as temperatures, power, pump speeds, setpoints, etc. -- and using it to train an ensemble of deep neural networks. Since our objective was to improve data center energy efficiency, we trained the neural networks on the average future PUE (Power Usage Effectiveness), which is defined as the ratio of the total building energy usage to the IT energy usage. We then trained two additional ensembles of deep neural networks to predict the future temperature and pressure of the data center over the next hour. The purpose of these predictions is to simulate the recommended actions from the PUE model, to ensure that we do not go beyond any operating constraints.


We tested our model by deploying on a live data center. The graph below shows a typical day of testing, including when we turned the machine learning recommendations on, and when we turned them off.

DC_PUE.png
Google DeepMind graph showing results of machine learning test on power usage effectiveness in Google data centers

Our machine learning system was able to consistently achieve a 40 percent reduction in the amount of energy used for cooling, which equates to a 15 percent reduction in overall PUE after accounting for electrical losses and other non-cooling inefficiencies. It also produced the lowest PUE the site had ever seen.


Because the algorithm is a general-purpose framework to understand complex dynamics, we plan to apply this to other challenges in the data center environment and beyond in the coming months. Possible applications of this technology include improving power plant conversion efficiency (getting more energy from the same unit of input), reducing semiconductor manufacturing energy and water usage, or helping manufacturing facilities increase throughput.


We are planning to roll out this system more broadly and will share how we did it in an upcoming publication, so that other data center and industrial system operators -- and ultimately the environment -- can benefit from this major step forward.

DeepMind AI reduces energy used for cooling Google data centers by 40%

From smartphone assistants to image recognition and translation, machine learning already helps us in our everyday lives. But it can also help us to tackle some of the world’s most challenging physical problems -- such as energy consumption.  Large-scale commercial and industrial systems like data centers consume a lot of energy, and while much has been done to stem the growth of energy use, there remains a lot more to do given the world’s increasing need for computing power.

Reducing energy usage has been a major focus for us over the past  10 years: we have built our own super-efficient servers at Google, invented more efficient ways to cool our data centers and invested heavily in green energy sources, with the goal of being powered 100 percent by renewable energy. Compared to five years ago, we now get around 3.5 times the computing power out of the same amount of energy, and we continue to make many improvements each year.

Major breakthroughs, however, are few and far between -- which is why we are excited to share that by applying DeepMind’s machine learning to our own Google data centers, we’ve managed to reduce the amount of energy we use for cooling by up to 40 percent. In any large scale energy-consuming environment, this would be a huge improvement. Given how sophisticated Google’s data centers are already, it’s a phenomenal step forward.

The implications are significant for Google’s data centers, given its potential to greatly improve energy efficiency and reduce emissions overall. This will also help other companies who run on Google’s cloud to improve their own energy efficiency. While Google is only one of many data center operators in the world, many are not powered by renewable energy as we are. Every improvement in data center efficiency reduces total emissions into our environment and with technology like DeepMind’s, we can use machine learning to consume less energy and help address one of the biggest challenges of all -- climate change.

One of the primary sources of energy use in the data center environment is cooling. Just as your laptop generates a lot of heat, our data centers -- which contain servers powering Google Search, Gmail, YouTube, etc. -- also generate a lot of heat that must be removed to keep the servers running. This cooling is typically accomplished via large industrial equipment such as pumps, chillers and cooling towers. However, dynamic environments like data centers make it difficult to operate optimally for several reasons: 

  1. The equipment, how we operate that equipment, and the environment interact with each other in complex, nonlinear ways. Traditional formula-based engineering and human intuition often do not capture these interactions.
  2. The system cannot adapt quickly to internal or external changes (like the weather). This is because we cannot come up with rules and heuristics for every operating scenario.
  3. Each data center has a unique architecture and environment. A custom-tuned model for one system may not be applicable to another. Therefore, a general intelligence framework is needed to understand the data center’s interactions.
To address this problem, we began applying machine learning two years ago to operate our data centers more efficiently. And over the past few months, DeepMind researchers began working with Google’s data center team to significantly improve the system’s utility. Using a system of neural networks trained on different operating scenarios and parameters within our data centers, we created a more efficient and adaptive framework to understand data center dynamics and optimize efficiency.

We accomplished this by taking the historical data that had already been collected by thousands of sensors within the data center -- data such as temperatures, power, pump speeds, setpoints, etc. -- and using it to train an ensemble of deep neural networks. Since our objective was to improve data center energy efficiency, we trained the neural networks on the average future PUE (Power Usage Effectiveness), which is defined as the ratio of the total building energy usage to the IT energy usage. We then trained two additional ensembles of deep neural networks to predict the future temperature and pressure of the data center over the next hour. The purpose of these predictions is to simulate the recommended actions from the PUE model, to ensure that we do not go beyond any operating constraints.

We tested our model by deploying on a live data center. The graph below shows a typical day of testing, including when we turned the machine learning recommendations on, and when we turned them off.

Green_-_07_20_16_-_Deepmind_Reduces_Energy.width-1600.png
Google DeepMind graph showing results of machine learning test on power usage effectiveness in Google data centers

Our machine learning system was able to consistently achieve a 40 percent reduction in the amount of energy used for cooling, which equates to a 15 percent reduction in overall PUE overhead after accounting for electrical losses and other non-cooling inefficiencies. It also produced the lowest PUE the site had ever seen. 

Because the algorithm is a general-purpose framework to understand complex dynamics, we plan to apply this to other challenges in the data center environment and beyond in the coming months. Possible applications of this technology include improving power plant conversion efficiency (getting more energy from the same unit of input), reducing semiconductor manufacturing energy and water usage, or helping manufacturing facilities increase throughput.

We are planning to roll out this system more broadly and will share how we did it in an upcoming publication, so that other data center and industrial system operators -- and ultimately the environment -- can benefit from this major step forward.