SolidNode

A diagram of the design of the CLUMEQ Colossus supercomputer, from a recent presentation by Marc Parizeau of CLUMEQ.

Here’s one of the most unusual data center designs we’ve seen. The CLUMEQ supercomputing center in Quebec has worked with Sun Microsystems to transform a huge silo into a data center. The cylindrical silo, which is 65 feet high and 36 feet wide with two-foot thick concrete walls, previously housed a Van de Graaf particle accelerator. When the accelerator was decommissioned, CLUMEQ decided to convert the facility into a high-performance computing (HPC) cluster known as Colossus.

We first noted the development of the CLUMEQ site earlier this year when Marc Hamilton of Sun discussed its unique design, but offered scant details. Additional information about the design of the facility and its cooling system were discussed at the Sun HPC Consortium last month in Portland, Oregon.

The CLUMEQ Colossus cylinder features an interior “hot core” (as opposed to a hot aisle) in the center of the building and uses the outside ring of the facility as the cold air plenum. The cabinets are arranged in a ring on each floor, facing the outside of the silo. The floors supporting each ring of cabinets are comprised of grates rather than solid flooring to facilitate airflow through the facility.

The cooling coils and air handlers are located in the basement. Chilled air flows upward through the outside cold aisle and through the racks of servers. The waste heat exits the rear of the racks into the hot core, and is returned to the basement via the cold aisle.

The air flow pattern is maintained through differential air pressure – maintaining a higher air pressure in the cold aisle than the hot aisle. This keeps the air moving through the facility, which has a blowing capacity of 180,000 CFM and can cool up to 1.5 megawatts of electrical load. Up to 300 kilowatts of cooling capacity can be supplied by free cooling using fresh air from outside the facility.

“CLUMEQ silo totally blows up the paradigm of data center design,” says Nicolas Dube of Sun, who began work on the project as a graduate student at Universite Laval in Quebec. “The silo, by itself, is the CRAC (computer room air conditioner). The whole facility cools itself.”

As for computing horsepower, Colossus will have a peak of 86 teraflops of compute power. It’s equipped with a Sun Constellation HPC systems featuring 10 fully loaded Sun Blade 6048 chassis, 1 petabyte of Lustre storage and Sun J4400 storage arrays.

The data center racks are spread over three floors, with the switches on the second floor to keep the cable runs as short as possible.

For a full description of the CLUMEQ design, check out this video from Sun, which runs about 6 minutes.

Additional details are available in PDFs of presentations by Marc Parizeau of CLUMEQ and Nicolas Dube of Sun.

OTHER STORIES YOU MAY LIKE:

• Who Has The Most Web Servers?
• The Gallery of Exploding Servers
• Inside the ‘James Bond Villain’ Data Center
• Google’s Chiller-Less Data Center
• Inside a Cloud Computing Data Center

Original post:
Wild New Design: Data Center in A Silo

Popularity: 12% [?]

Anthony Wanger is the President and Founder of i/o Data Centers. He directs the company’s strategic affairs, handles acquisition activities, and oversees the company’s marketing, HR and legal functions.

The transition from physical, “offline” processes to digital, online processes is referred to as digitization or dematerialization. Processes that are digitized produce less carbon emissions than their analog counterparts. Data centers provide the infrastructure that enables this digitization to occur, serving as the foundation for the energy efficient enterprise.

ANTHONY WANGER
i/o Data Centers

Over the past 20 years there have been hundreds, if not thousands, of offline processes that have been digitized – everything from software distribution to financial transactions to medical record keeping. The combined effect of all of these digital processes is a macro-scale reduction in the overall use of materials and a more efficient distribution of the materials.

With the advent of the commercial Internet in the 1990s, companies have improved how they interact with their customers, partners and employees. Prior to the World Wide Web and email, businesses and government transacted in mostly inefficient and unconnected ways. For example, in order to pay a bill, a buyer would send a paper-based check by mail, which would be delivered to the recipient by way of a network of carbon-emitting postal vehicles.

Today, bills can be paid online in a matter of minutes. By digitizing this offline process, the need for material (i.e. paper) to be created and transported has been eliminated. This in turn has resulted in a significant reduction in CO2 emissions.

The impact of dematerialization has been quantified by a number of prominent corporations and research institutions including Microsoft, Intel, Lawrence Berkley Labs and Stanford University.

• A recent paper by Dr. Jonathan Koomey, senior researcher for Lawrence Berkley Labs, studied the impact on CO2 emissions resulting from the purchase of music online as compared to the purchase of a compact disc at a music store. Their research shows that the process of purchasing music online can reduce CO2 emissions, on a conservative basis, between 40%-80%.
• A comparative carbon footprint study of Microsoft’s Office 2007 product suite found that the digital delivery of their product to customers reduced carbon emissions by 88%.
• According to NPG Group, Apple leads the U.S. with 25% of all music sold, surpassing both Wal-Mart and Amazon.com. Apple’s iTunes music service has materially changed the way music is purchased and in so doing has eliminated a substantial portion of the carbon footprint related to the offline distribution of music.

iTunes, software distribution, online bill payment and many other digital services are delivered by a complex array of IT systems including servers and telecommunications networks. These servers and networks are located in data centers. Data centers provide the infrastructure (i.e. power, cooling, network access) required by these digital services to function.

Most corporate data centers are built to accommodate the IT needs of a single business unit or department. Large, commercial-grade data centers leverage the economies of scale to reduce energy consumption. Instead of operating ten smaller data centers, an organization could consolidate their IT infrastructure into one or two large data centers and reduce the costs and energy associated with operating separate cooling, UPS, backup power and network access systems.

Modern data centers use the latest technologies and engineering best practices including variable frequency drives, light-emitting diodes (LED) fixtures, thermal energy storage, photovoltaic (PV) solar arrays, ultrasonic humidification and sealed cabinets. Collectively, these systems contribute to a significant reduction in energy consumption – especially during peak load periods.

Conclusion
By combining digitized processes with the economies of scale recognized at large, modern, commercial-grade data centers, today’s enterprise can materially reduce the energy it consumes and greatly improve its efficiency. As consumers, businesses and government look for more efficient ways to communicate and transact, dematerialization and the data center will provide the foundation for a more energy efficient enterprise.

Industry Perspectives is a new content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating in Industry Perspectives.

Original post:
How Digitization Drives Energy Efficiency

Popularity: 13% [?]

IBM today announced plans to build a new data center in Auckland, New Zealand, investing $58 million (NZD $80 million) in a partnership with local developer Highbrook Development Ltd. The 56,000 square foot facility, which will feature a first phase of 16,000 square feet of data center space, will be located in the Highbrook Business Park in East Tamaki and be fully operational by late 2010. Additional data center pods will be developed as demand rises.

The New Zealand project is one of three new IBM computing facilities in Asia that were announced today. IBM is also building a new data center in Seoul, South Korea and a cloud computing lab in Hong Kong.

The Auckland data center will include a free cooling system which will leverage outside air during colder months, reducing the need for chillers.

“This is a highly significant investment in New Zealand’s future technological infrastructure,” said Jennifer Moxon, Managing Director of IBM New Zealand. “It is the result of long term strategic planning and signals IBM’s commitment to enable New Zealand to become a world class technology center and advances the growth of the digital economy.”

“IBM’s new facility is core infrastructure for the 21st century digital economy that the Government and ICT industry are mutually striving to develop,” said Brett O’Riley, Chief Executive Officer, NZICT Group. “It heralds the explosion of software as a service and cloud enabled computing which will drive a step change in productivity and innovation, while its design establishes the benchmark for green ICT in this country.”

Original post:
IBM Announces Data Centers in Pacific Rim

Popularity: 6% [?]

Quality Technology Services is moving quickly to expand its data center footprint, leveraging a recent $150 million investment from private equity firm General Atlantic. QualityTech said Tuesday that it will build out an additional 130,000 square feet of raised floor data center space in its massive Metro data center in downtown Atlanta.

Two large raised-floor pods will be added to the facility’s 200,000 square feet of existing data center space. The expansion will bring the 990,000 square foot building to nearly  65 percent of the planned raised-floor capacity.

“Our partnership with General Atlantic has allowed QualityTech to start the next step in our phased build-out of the Metro Atlanta Data Center facility,” said Mark Waddington, President of QualityTech. “The addition of over 130,000 square feet of raised floor space and supporting infrastructure positions this facility as one of the world’s largest and most technologically-advanced data center facilities.”

QualityTech said the expansion will focus on environmentally sustainable design practices. The company is planning to obtain certification through the LEED (Leadership in Energy and Environmental Design) program for energy-efficient buildings created by the US Green Building Council.

The Metro facility will include a Facilities Lab to test new energy conservation initiatives. The lab’s work will focus on air and water economization technologies as well as new developments in power distribution and consumption technology. QualityTech will also deploy advanced cooling and delivery methods to the floor that will allow customers to reach even higher levels of power density.

QualityTech offers both wholesale data center space and managed services. The company has built a network of 11 data center facilities spanning more than 2.5 million square feet of space in New York, New Jersey, Florida, Georgia and Silicon Valley.  Nearly 1.3 million square feet of Quality Tech’s data center space is in the Atlanta market, where the company operates a 375,000 square foot data center in Suwanee in addition to the Metro Atlanta property.

The company says it currently has space available for deployment in sizes up to 20,000 square feet in Atlanta and up to 15,000 square feet in Santa Clara.

Original post:
Quality Tech Expands Atlanta Metro Site

Popularity: unranked [?]

Which energy efficiency technologies work together to produce the best combined savings? IBM and Syracuse University will gather data on this question in an innovative new facility that combines on-site power distribution using microturbines, absorption chillers, the reuse of waste heat in nearby buildings, DC power distribution, and rear-door liquid cooling for each cabinet. Check out our a photo tour providing a detailed look at the technologies and equipment powering the Syracuse facility.

Original post:
Inside IBM’s Green Data Center Testbed

Popularity: unranked [?]

At the Gartner Data Center Conference we spoke with Pete Van Deventer, the president and CEO of SynapSense Corp., which makes wireless monitoring products that help data center operators measure the environment and optimize their facilities for energy efficiency. Van Deventer discusses the market for wireless monitoring, how it may be impacted by new carbon regulations, and SynapSense’s development of technology to use monitoring tools to automate the management of data center cooling. This video runs about 7 minutes, 45 seconds.

For more coverage of information about energy efficiency, check out our Green Data Centers Channel. For additional video, check out our DCK video archive and the Data Center Videos channel on YouTube.

Original post:
SynapSense CEO on Wireless Monitoring

Popularity: 6% [?]

We’re a little tardy with our review of last month’s most popular story, as we were busy at the Gartner Data Center Conference last week. But there continues to be huge interest in data center power and cooling innovations at Facebook and Google, which accounted for four of our top five stories for the month. Here’s a look at the 10 most popular stories of November 2009, ranked by total page views:

• Facebook Follows Google to Data Center Savings (Nov. 27)
• A Closer Look at Google’s New Cooling Design (Nov. 30)
• WoW’s Back End: 10 Data centers, 75,000 Cores (Nov. 25)
• Google Patent Reveals Data Center Innovations (Nov. 30)
• Inside A Cloud Computing Data Center (Nov. 4)
• Should Servers Come With Batteries? (Nov. 27)
• Power Outage Effects Rackspace Cloud (Nov. 3)
• Facebook Goes Green With New Data Centers (Nov. 5)
• Rackspace: We Have Work to Do (Nov. 4)
• Microsoft’s Windows Azure Cloud Container (Nov. 18)

You can stay current on the latest data center news by subscribing to our RSS feed and daily e-mail updates.

Original post:
Best of Data Center Knowledge, November 2009

Popularity: 12% [?]

Elisabeth Stahl is the Chief Technical Strategist, Performance Marketing for IBM Systems and Technology Group.

Anyone who took Psychology 101 in school remembers Maslow’s hierarchy. In this groundbreaking 1943 paper, Abraham Maslow outlined a pyramid to demonstrate the five distinct levels of human “need.” His pyramid showed basic, physiological needs at the bottom, more sophisticated needs at the top, all leading to the ultimate, final phase characterized by a profound level of “self-actualization” of identity and purpose.

However, this theory assumes that only personal growth can be achieved by moving through the five levels. In an interesting twist, we can actually apply Maslow’s hierarchy to data center energy efficiency, allowing us to ultimately realize the maximum potential for our IT organizations.

ELISABETH STAHL
IBM

The world has grown passionately interested in energy efficiency within the data center in the last several years. As energy prices climb and organizations outgrow their power and cooling limits, it becomes imperative for data center managers to address IT energy efficiency through green initiatives. Many of us have become intimately familiar with recommendations for improving the efficiency of our data centers. Create those hot and cold aisles, use those pillows and baffles, update those cables and that lighting, consolidate and virtualize, and maybe try some innovative water cooling technology.

So what is the hierarchy for IT energy efficiency and how can we realize the full potential? Let’s outline it step by step:

• Level One: The lowest level of this hierarchy is when energy is still seen as a basic commodity; it has no influence on data center choices. The infrastructure is not monitored and many hot spots exist on the data center floor;
• Level Two: The next level assumes that some thermal monitoring is performed. Some simple decisions have been made to help the infrastructure such as straightforward consolidations and elementary server virtualization;
• Level Three: In this stage, organizations have implemented tools to start actively monitoring IT and non-IT assets such as air conditioners.
• Level Four: Next, we focus on more significant optimization including monitoring metrics such as the power usage effectiveness (PUE), more sophisticated virtualizations using storage and the network, and even potentially employing free cooling which can save thousands of dollars in cooling costs while simultaneously cutting back on greenhouse gas emissions.
• Level Five: Finally, the highest level, the “self-actualization” of data center energy efficiency, can be reached. Here we have the ultimate cooling and powering infrastructure. We see exploitation of the virtualized environment with significant workload management capabilities. This level is most notable for introducing the data center finally as an actual Energy Producer.

How cool is that? We can effectively take the heat we don’t want in the data center and basically reuse it where we need it while saving energy and lowering emissions.

Companies around the world are already implementing this innovative heat reuse approach, sometimes referred to as “district heating,” in a concentrated effort to reach their full potential as an energy producer. For example, by working with IBM, the Swiss Federal Institute of Technology Zurich is building a first-of-a-kind water-cooled supercomputer dubbed Aquasar that will directly repurpose excess heat for the university buildings. The system, which also features a water-cooling system, is expected to cut energy consumption by 40% and carbon-dioxide emissions by up to 85%, resulting in dual IT savings.

Through adopting this final level of the IT energy efficiency hierarchy, we can build a scalable, flexible, and green data center that is dynamic in its infrastructure. Through this “self-actualization” we can potentially save on energy costs; as a producer we might also even be able to make money as well.

Industry Perspectives is a new content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating in Industry Perspectives.

Original post:
Reaching Self-Actualization in IT Energy Efficiency

Popularity: 12% [?]

Fred Stack is vice president of marketing for Liebert Precision Cooling at Emerson Network Power. Stack is responsible for new product development roadmaps that reflect evolving market demands and incorporate new technology.

The recent expansion of data center temperature limits has raised various questions and concerns throughout the industry. Some data center and IT managers have voiced reservations about elevating the overall data center temperature and reducing their outage ride through time while others seek all possible efficiency gains.

FRED STACK
Emerson
Network Power

As a refresher, the purpose of the ASHRAE-recommended envelope is to give guidance on maintaining high reliability while efficiently operating data centers. The newest ASHRAE recommended environmental range for class one and two data centers is 18 to 27° C (64.4 to 80.6° F) (dry bulb temperature). The allowable envelope which broadens the range to 32° C (90° F) is where IT equipment manufacturers test product in order to verify that the unit will function with no damage for periods of times generally assumed to be counted in hours or a few days for emergency operation.

These higher temperature ranges are being promoted to emphasize to the industry that these are server input temperatures, not cooling unit return temperatures. The temperatures in the cold aisles can be significantly increased from the averages across the industry for a significant increase in the cooling capacity and efficiency of the cooling equipment.

The higher temperature ranges are also being used to increase the number of hours an air or water economizer can be used in a data center. Water economizers have minimal operational concerns, just some maintenance procedures, while air economizers require data center managers recognize the potential impact of energy to humidify dry air along with the potential impact of particulate and gaseous contamination.

One additional caution has been explained in the ASHRAE publications. That is the move from 25° C to 27° C (77° F to 80.6° F) can have a negative impact on the IT equipment’s power dissipation. Most IT manufacturers start to increase air moving device speed around 23° C (73° F) to improve component cooling and offset the increased ambient air temperature. Increases in the server intake temperature from 23° to 27°C will double the server fan energy consumption. A typical server fan consumes 10 percent of the total server energy at low temperatures.

This small change in input temperature to 27°C (80.6°F) will increase the total server energy by 10 percent. A move to 29°C (84.2°F) increases the total server energy by 20 percent. These increases in server energy, while an improvement to your PUE metric, actually negate all the energy savings from the effort. For this reason it is usually recommended to only move above 24°C (75.2°F) when an economizer provides access to “free cooling.”

Ultimately, raising temperatures in data centers should be carried out in a systematic manner to best determine where hot spots may occur. Care should be taken to ensure that data centers are operating efficiently without risking infrastructure shutdown because of high temperatures or contamination.

Industry Perspectives is a new content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating in Industry Perspectives.

Original post:
Managing the Expanding Temperature Envelope

Popularity: unranked [?]

A diagram of an “air wand” indicating the location of cooling vents in the wand, a key feature of a patent application by Google data center engineers.

Google has revealed some of the secret technology inside its mighty data centers, but its engineers are busy cooking up new secrets.

An example: Google is seeking to patent an advanced data center cooling system that provides precision cooling inside racks of servers, automatically adjusting to temperature changes while dramatically reducing the energy required to run chillers.

The cooling design, which could help Google slash the power bill for its servers, reinforces Google’s focus on its data centers as a competitive advantage in its battle with Microsoft and other rivals for leadership in cloud computing. The company has customized much of the operation of its data centers, which serve as the engines powering its massive Internet business. Google builds its own servers and networking switches, and now appears to be customizing the racks that hold them.

Precision Cooling via ‘Air Wands’
The innovative rack cooling design features an adjustable piping system, including “air wands” that provide small amounts of cold air to components within a server tray. The chilled air enters the top of a rack through two vertical standpipes, which branch off into air wands – long, thin pipes lined with vents that release cold air.

The air wands can pivot to target cold air on specific components, or be swung to one side to allow equipment to be removed from the rack. Dampers on each standpipe can open and close to regulate the volume of air flowing into the pipe and air wands, while the vents on each individual air wand can be adjusted to point up or down, allowing for a highly configurable system. (See A Closer Look at Google’s New Cooling Design for a diagram).

Exaflop and Its History
It’s not clear whether Google is already using the cooling system. But the patent application was submitted by Exaflop LLC, whose 2008 patent for a UPS system integrating batteries with server power supplies helped Google achieve 99.9 percent UPS efficiency and record low Power Usage Effectiveness (PUE) scores. The address for Exaflop is 1600 Amphitheatre Parkway in Mountain View, Calif., which is Google’s headquarters. The inventors listed on the patent are Google employees Jimmy Clidaras and Winnie Leung.

The system designed by Clidaras and Leung addresses many of the most vexing challenges in data center energy efficiency. It allows Google to apply small amounts of cold air precisely where it is needed, rather than cooling an entire server room and seeking to steer the airflow into each rack and across the hot server components.

Going Beyond Containers
Google has used data center containers to isolate hot and cold air and gain greater control over airflow to its servers. The new design takes this concept to a more granular level of management. The air wands can apply cool air directly to the “hot spots” inside a server tray, meaning less air is wasted or misdirected in the server room or container. This could allow Google to use a smaller chiller plant in its data centers, saving energy in the process.

Chillers, which are used to refrigerate water for use in data center cooling systems, require a large amount of electricity to operate. With the growing focus on power costs, many data centers are trying to reduce their reliance on chillers.

This has boosted adoption of “free cooling,” the use of fresh air from outside the data center to support the cooling systems. This approach allows data centers to use outside air when the temperature is cool, while falling back on chillers on warmer days. The new design could be used as supplemental cooling in a data center using free cooling, or in facilities located in areas where fresh air cooling isn’t feasible.

Limitations of Free Cooling
Google is operating a chiller-less data center in Belgium, where the climate allows nearly year-round use of free cooling. But this strategy will only work in cooler regions, and Google’s global ambitions may eventually require data centers in hotter climates unsuitable for free cooling.

Google can gain additional control over its cooling system through automated monitoring and management, as the system is designed to respond to changes within the rack as temperatures fluctuate. “The temperature sensor output can be fed to a computer program that triggers air distribution in the event of the board temperature crossing a threshold,” the patent reads. “Each temperature sensor may be connected to a PID control loop with a damper, so the corresponding damper is opened … with an increase in temperature sensed for a particular area.”

Some Secrets Revealed, While Others Incubate
Google’s data center designs were kept secret for many years, consistent with the company’s belief that its data center innovations gave it a competitive advantage. In April Google discussed its data center operations for the first time, joining a growing industry conversation about best practices for energy efficiency.

The company revealed its data center containers, custom server design and on-board UPS, among other innovations. But some industry observers concluded that there was more in the pipeline that Google wasn’t discussing.

“Both the board and the data center designs shown in detail where not Google’s very newest but all were excellent and well worth seeing,” James Hamilton noted at the time. “I like the approach of showing the previous generation technology to the industry while pushing ahead with newer work. This technique allows a company to reap the potential competitive advantages of its R&D investment while at the same time being more open with the previous generation.”

RELATED STORIES:

• Google’s Custom Web Server Revealed
• Google Claims ‘World’s Most Efficient Data Centers
• Efficient UPS Aids Google’s Extreme PUE
• Google Unveils its Data Center Containers
• Google’s Chiller-Less Data Center

Original post:
Google Patent Reveals Data Center Innovations

Popularity: unranked [?]