Navigant Research Blog

It Takes a Lot of Energy to Catch ‘Em All

— July 29, 2016

Cloud ComputingPokémon GO has taken over the world. For those who have not yet played the game, it’s an augmented reality smartphone app where players walk around collecting Pokémon, battling in gyms, and generally having a good time. It’s also on the forefront of technological innovation, combining mapping data from Google with a narrative from the longstanding franchise. Niantic Labs, the developers of the game, have risen to the forefront of the technology world. Nintendo, one owner of the Pokémon franchise, became the most traded company by value of shares swapped on the Tokyo stock market this century. However, shortly after this rise, the stocks plummeted. Nintendo is not, after all, directly responsible for the development of the popular game and only owns a 32% stake in The Pokémon Company.

However, there is, as they say, a Butterfree in the ointment. The immense popularity of Pokémon GO has caused overrun servers and overheating data centers, making the free app crash every few hours. In addition, players are expressing frustration with the app’s  intense battery draining ability. A typical smartphone battery can drain in as few as 40 minutes of gameplay. The game is based entirely around GPS capabilities, which are notorious battery hogs. While GPS is running, a mobile device cannot enter a sleep state. In addition, communications channels with GPS satellites are very slow, and mapping software is processor-intensive, further compounding the energy intensity of such applications.

The intense data and energy use of the game has caused Werner Vogels, CTO of Amazon, to offer Niantic assistance in operating its servers. This intense usage of GPS capabilities, smartphone data, and server capacity promises to bring Pokémon GO to the top spot in smartphone application energy usage. According to SimilarWeb, in its first 4 days of use, the number of Pokemon GO users nearly surpassed Twitter users in the United States.

 Daily Active Users: Pokémon GO vs. Twitter

PokemonBlog

 (Source: SimilarWeb)

In terms of average time users spend using the app, Pokémon GO has surpassed social media sites WhatsApp, Instagram, Snapchat, and Facebook Messenger. The average player uses the app for 43 minutes a day. What’s more, Niantic plans to launch the app in over 200 countries as soon as servers are bolstered. With the current bulk of Pokémon trainers in the United States, a global phenomenon could have a large carbon footprint.

Pikachu-Powered Data Centers?

There’s little information available on the data centers that Niantic is using for the app, but the company is presumably using Google cloud data centers or something similar. Niantic was a part of Google until April 2015, when the two split. Google has always been known for its environmental stewardship in big data. The company’s data centers are reported to use 50% less energy than most in the industry, and it uses renewable energy to power over 35% of its operations. So while no data is available on Niantic’s end, it can be assumed that the company is using industry best practices in its data centers.

Niantic has not released any sort of impact statement on the app’s actual energy use, though it is almost certainly astronomical. Niantic is already hard at work developing improvements to the game, such as limiting the amount of personal data the app could access. The energy use could be measured to assess the app for potential energy improvements. A new tool called EnergyBox, developed by Ekhiotz Jon Vergara from Swedish Linkoping University, measures the energy consumption of mobile devices due to data communication. This tool finds that the way apps are designed helps to curb the energy used to send and receive large amounts of data. Niantic should take note of its app’s energy consumption before rolling it out globally, lest we be trapped in a Diglett-infested desert due to GO-related global warming.

 

Data Centers Drive Market for DC Distribution Networks

— July 15, 2015

The market for direct current (DC) distribution networks is not a single, cohesive market. Rather, it encompasses several disparate opportunities—telecommunications towers, data centers, grid-tied commercial buildings, and off-grid military networks—that revolve around different market assumptions, dynamics, and drivers.

Given the expense of current existing redundant alternating current (AC) uninterruptible power supply (UPS) systems, DC data centers would appear to be a no-brainer from an engineering point of view. Despite this, energy remains a small portion of the overall operations budget of data centers. As a result, the value proposition to conservative operations managers may still be a hard sell in the near term. However, DC microgrids can actually offer higher reliability than status quo AC solutions, so validating early adopter DC microgrids is a critical step forward for this market opportunity. The ABB 1 MW DC data center located in Zurich, Switzerland, is just one example of how this application is gaining momentum.

Distributing DC enables replacement of AC-DC converters within individual devices with a smaller number of larger, more efficient converters. LED lighting installations that run on 24V DC lines, for example, require up to 15% less energy than the same lights running on fixture-level rectifiers. Nevertheless, losses in the linings limit 24V DC distributions to just 10 meters, so manufacturers are developing 380V DC wiring to extend comparable benefits to entire data centers and other commercial buildings. Asia Pacific is expected to lead this market in both the near and long-term, with China alone having already deployed hundreds of DC data centers.

DC Data Center Network Implementation Revenue by Region,
Base Scenario, World Markets: 2015-2024

DC Data Network Implementation Revenue

(Source: Navigant Research)

The core challenge facing DC distribution networks lies with the need for standards and open grid architectures that can help integrate the increasing diversity of resources being plugged into retail power grids. Even DC advocates maintain that distribution networks operating at the municipal level may always remain AC systems. The efficiency gains accrued by sticking with DC instead of converting to AC (and then back to DC) are not as great at this higher voltage level. This may remain the sweet spot for AC technology, serving the vital role of interconnecting large wholesale transfers from high-voltage DC (HVDC). In fact, DC microgrids and nanogrids could, ironically enough, extend the life of the incumbent AC distribution system by taking loads off that system in an intelligent and dynamic way.

The focus of the industry, working through the efforts of the EMerge Alliance, is currently medium-voltage DC distribution networks. These systems are mostly concentrated on the data center market segment, but can also apply to commercial buildings—especially those of considerable scale, such as big box retailers (Costco, Walmart, etc.). At present, the majority of progress in developing DC-based technologies has occurred at either the high-voltage (more than 1,000V) or low-voltage (less than 100V) level of electricity service. Since microgrids and building-scale nanogrids typically operate at medium-voltage (roughly 380V to 400V), much work needs to be done to bridge this voltage innovation gap, and this goal is the focus of companies such as ABB, Bosch, Emerson Network Power, and others.

As noted in a previous blog, Bosch is encountering a few regulatory issues when it comes to deploying DC microgrids, primarily an artifact of assumptions that distributed renewables and energy storage are interconnected to the alternating current utility grid. But surprisingly, DC fits in well with AC infrastructure, and is especially accommodating for integrating cutting-edge distributed energy resources.

 

Google Expands Data Center Fleet at Retiring TVA Coal-Fired Generation Site

— July 7, 2015

In late June, Google announced plans to site its 14th massive data center at the Widows Creek Tennessee Valley Authority (TVA)  coal-fired generation site in Alabama.  The $600 million facility will also re-purpose the 60-year-old coal-fired site, which will soon be retired, leveraging existing electric transmission and distribution infrastructure. This infrastructure might have otherwise become another stranded utility asset, ultimately abandoned.  While the data center is planned to be powered by 100 percent renewable energy, access to existing electric transmission infrastructure provides access to renewable power not generated at the data center site, as well as additional backup capabilities sometimes necessary to operate 24/7.

The size of and the electrical power requirements for Google data centers are huge. Google has a history of building creative large-scale data centers, which are some of the largest electric power consumers on the transmission grid.  The company’s data center designs are typically state of the art, utilizing the latest cooling technologies to keep aisle after aisle of servers running at optimal temperatures on a 24/7 basis.  Google’s designs are also known for their creative use of renewable energy where possible.

TVA’s Widows Creek Facility

TVA

(Source: Tennessee Valley Authority)

Model  Behavior

Gary Demasi, Google’s director of data center energy and location strategy, reported that, “The idea of re-purposing a former coal generating site and powering our new facility with renewable energy — especially reliable, affordable energy that we can count on 24/7 with the existing infrastructure in place — was attractive.”

Patrick Gammons, Google’s senior manager for data center energy and location strategy added, “Thanks to an arrangement with Tennessee Valley Authority, our electric utility, we’ll be able to scout new renewable energy projects and work with TVA to bring the power onto their electrical grid. Ultimately, this contributes to our goal of being powered by 100% renewable energy.”

With the seemingly insatiable need for power, infrastructure, and real estate that large data centers have, this 100% renewable data center plan provides a model for other utilities nationwide to use when determining how to redevelop coal plant sites. With the Energy Information Administration’s (EIA’s) recent announcement that the coal plant retirement timetable is accelerating, in part due to the U.S. Environmental Protection Agency’s Clean Power Plan mandate, Google’s clean energy leadership is certainly an inspiration.

 

Data Centers Morphing Into Virtual Power Plants

— February 12, 2013

What is a “virtual power plant?” The term means different things to different people in different parts of the world.  Pike Research has come up with its own definition: A system that relies upon software to remotely and automatically dispatch and optimize generation, demand-side, or storage resources (including PEVs and bi-directional inverters) in a single, secure web-connected platform.

At their core, VPPs tap existing grid networks to tailor electricity supply and demand services for a customer, utility, or grid operator.  Without any large-scale fundamental infrastructure upgrades, VPPs can stretch supplies from existing generators and utility demand reduction programs.

The latest VPP model to emerge is based not on geographic proximity – typically the top consideration – but rather on enterprise ownership of global operations.  Ironically enough, the farther away each facility linked in the VPP, the better!  Companies such as PowerAssure are investigating ways for companies that use large global data center operations, such as Apple and Google, to create enterprise VPPs that span the globe, whereby data centers shut down operations and shift load from the regions of the world in daylight to the nighttime half of the globe, where power is cheaper.  The technology to carry out this level of global energy arbitrage – known as “following the moon” – is nearly here (though some engineers may disagree). “Data centers can modulate their IT loads based on external events, such as the price of power, and in the process, save money and get paid for providing demand response (DR) services,” Peter Maltbaek, vice president of worldwide sales for PowerAssure, told me.

Changing Models and Mindsets

The U.S. Environmental Protection Agency (EPA) recently revised rules governing limits imposed upon use of diesel generators that should help increase the availability of DR throughout the United States.  The chief challenge for global enterprise VPPs comes on similar regulatory restraints as well as the accounting end of such transactions.  Of course, if large numbers of large energy users employed this strategy, it could wreak havoc with local grid stability instead of enhancing reliability.  How national and regional regulators would respond to such a business model, based largely on financial flows instead of engineering smarts, is unclear.

Another challenge is changing the mindset of data center owners.  “They need 100% availability and are leery of anyone fooling around with their power supply, especially since it is only typically 3% of total costs,” added Maltbaek.

Lawrence Berkeley National Laboratories (LBNL) released a study last year that looked at data centers and their potential for DR.  ABB, which has invested in PowerAssure and has its own Decathlon DCIM VPP offering for data centers, has already installed a 1 megawatt (MW) DC microgrid at a data center in Zurich, Switzerland providing DR through use of its emergency generators; this system is currently being expanded to 10 MW, will later go to 30 MW, and will then be aggregated with three other data centers in the region.

In Germany, meanwhile, Siemens claims that recent regulatory reforms will allow it to boost its supply-side VPP capacity to 3,000 MW by 2018.  Last year, the company announced that it would increase the capacity of its VPP from less than 10 MW to 200 MW by 2015.  The company says that Germany has enough spare capacity on its transmission lines to create VPPs that span the entire country.

 

Blog Articles

Most Recent

By Date

Tags

Clean Transportation, Digital Utility Strategies, Electric Vehicles, Energy Technologies, Policy & Regulation, Renewable Energy, Smart Energy Practice, Smart Energy Program, Transportation Efficiencies, Utility Transformations

By Author


{"userID":"","pageName":"Data Centers","path":"\/tag\/data-centers","date":"12\/15\/2017"}