Published on 30 Mar 2011
Validated on 19 Nov 2012
"Until we began using IBM Software to monitor and manage energy usage and thermal output, we didn’t even know that problems existed." - —Hilon Potter, Chief IT Architect, IBM Poughkeepsie Green Data Center
Autonomic Computing, Asset Management, Dynamic Infrastructure, Green/Sustainability, Cloud & Service Management, Service Management, Smarter Planet, Storage Consolidation
The IBM Poughkeepsie Green Data Center has experienced enormous growth over the years and has upgraded its existing systems with new higher performance, energy efficient models to keep pace.
As the IBM Poughkeepsie Green Data Center continued to increase processing capacity and deliver more services, staff knew that re-assessing hardware placement and cooling capabilities would be critical to data center operations
Through an exciting energy management program, IT staff can monitor, measure and visualize data center conditions and energy consumption to see issues in real time and implement preventive measures before failures occur. Visibility into hot-and cold spots helped IT staff uncover issues that had previously gone unnoticed. Greater control helped double processing power and enabled a data center infrastructure efficiency (DCiE) level of 84 percent, the highest category in this measurement.1
Reduced the amount of power required for cooling racks by 50 percent; Doubled the amount of processing power available without having to increase square footage; Achieved a data center infrastructure efficiency (DCiE) level of 84 percent; Avoided equipment failures to deliver continuous service to its users
The ability to visualize and automatically adjust data center energy usage and environmental conditions to improve energy efficiency and avoid equipment failures.
For companies that need to deliver more processing power without expanding their data centers, the IBM Poughkeepsie Green Data Center offers a model for smarter IT. Through an exciting energy management program, IT staff can monitor, measure and visualize data center conditions and energy consumption to see issues in real time and implement preventive measures before failures occur. Visibility into hot-and cold spots helped IT staff uncover issues that had previously gone unnoticed. Greater control helped double processing power and enabled a data center infrastructure efficiency (DCiE) level of 84 percent, the highest category in this measurement.1
Smarter IT: Insight into environment improves operational efficiency
Instrumented: Captures thermal readings, barometric pressure, humidity, power consumption and utilization information from third-party sensors, power distribution units and servers
Interconnected: Integrates disparate data and displays it in dashboards to provide a holistic view of data center operations
Intelligent: Real-time view of environmental conditions and energy usage enables IT staff to uncover and resolve issues that would have otherwise gone unnoticed
Like many data centers worldwide, the IBM Poughkeepsie Green Data Center has experienced enormous growth over the years and has upgraded its existing systems with new higher performance, energy efficient models to keep pace. But as it continued to increase processing capacity and deliver more services, staff knew that re-assessing hardware placement and cooling capabilities would be critical to data center operations.
“We got to the point where our air conditioning was overloaded,” says Allan Hoeft, executive project manager for the Poughkeepsie Green Data Center. “We had a business need to increase capacity and realized that we had absolutely hit a wall.”
The Poughkeepsie Data Center consists of two raised-floor rooms: the front room measuring 1,100 square feet and supporting around 110 kilowatts of IT processing; the back room providing an additional 1,600 square feet of space. Front room operations support a number of customer-centric service centers, including IBM’s System z® Benchmark Center, Executive Briefing Center, Design Center and New Technology Center. The systems are used for running benchmarks, proof of concepts, and product demos on IBM mixed hardware and software.
The data center hosts IBM® System x®, IBM POWER5™, IBM POWER6®, IBM POWER7®, IBM BladeCenter®, IBM System z and IBM System Storage® technologies. The diverse environment enables companies worldwide to test workloads—both large and small—on all types of IBM systems. For example: “We have two IBM System z10s that are at 90-95 percent utilization and run 24 hours a day,” says Hoeft. “We’ve had one client use the data center to simulate 300 million users across multiple z10s—this was the largest benchmark in the world for this—and we have clients that just need to run small benchmarks with Linux® on System z or z/OS®.”
Even while the new systems provided greater energy efficiency, the overall increase in processing power created a demand for additional cooling capacity within the data center. “We hit a thermal wall and the only way we knew was when the alarms went off on the computer room air conditioners to indicate they were exceeding the temperature limit,” says Hilon Potter, chief IT architect, IBM Poughkeepsie Green Data Center. “We had no idea how close the equipment was to failing. A rack of five BladeCenter chassis was actually running with the inlet temperatures higher than the specification allows.”
Understanding the data center’s thermal footprint
Data center staff worked with IBM Systems and Technology Group (STG) Lab Services to understand and model thermal conditions in its existing configuration. IBM mobile measurement technology (MMT)2 was used to provide a detailed map of temperature, humidity and airflow measurements at key locations throughout the data center room. Using computational fluid dynamic models, the team then assessed what changes were needed so that critical systems would provide continuous service.
The thermal profile analysis showed several technologies, such as the 95 percent utilized System z servers, were already very efficient, but other technologies had room for improvement. Initial data center improvements included the use of IBM Cool Blue® rear door heat exchangers, which remove heat from racks using water-cooling methods; the implementation of new cooling distribution units; and the application of IBM and industry best practices for a raised floor data center. This work provided immediate and significant energy savings including a 50 percent reduction in the amount of power required for cooling racks.
This was only the first step, according to Hoeft. “Data centers are dynamic,” says Hoeft. “Invariably the week after you conduct an MMT analysis, you're ready to change something so we needed the ability to understand the impact changes had on the environment.”
An integrated approach to energy management
After targeting specific technologies, the data center staff next implemented an Integrated Service Management solution for the data center. This solution provided a holistic view on the visibility, control and automation over energy usage and environmental conditions. IBM Tivoli® Monitoring for Energy Management with IBM Systems Director Active Energy Manager enables the organization to integrate thermal, barometric pressure, humidity, power consumption and utilization information from third-party sensors, power distribution units and the systems themselves. The information, which is stored in a common repository using IBM Tivoli Data Warehouse, is then displayed in dashboards through IBM Tivoli Business Service Manager software to help staff see in near-real time how the facility is operating.
IBM Maximo® Asset Management for Energy Optimization software also uses the information collected to create real-time 3-D temperature and humidity maps that help staff identify “hot” and “cold” areas in the data center. The sensors provide information on all relevant sections of the data center, from the floor to just above the tops of the racks. If a hot spot appears, staff can quickly drill down to find out which assets are at that location and may be affected.
IBM Tivoli Netcool®/OMNIbus software aggregates any events or alerts from the servers and software, and automates fault deduplication, isolation and resolution. For example, data center staff is implementing an automated baffle in the raised floor. Once installed, if the facility temperature exceeds acceptable conditions, IBM Tivoli Netcool/OMNIbus will receive the alert and automatically instruct the baffle to open, increasing the air flow between the front and back rooms.
Gaining new insight for improved efficiency
In addition to targeting specific technologies and gaining a holistic insight to the data center’s energy management, the team went one step further to identify potentially hidden issues. This integrated energy management approach has been critical in providing new insight into data center operations. For example, staff saw that temperature readings around its IBM System z10® servers were higher halfway up the system than at the base. After investigating, the team found that the cause was airflow contamination from the back of the frame because a rear plate wasn’t properly installed.
“Once we installed the required backplate, it immediately improved the cooling,” says Potter. “We also found that when we did a shutdown and turned a rack off in the room, the barometric pressure under the floor didn’t drop to zero, and it should have. We discovered that the contractors had cut a hole underneath the floor going into the back room. Until we began using IBM Software to monitor and manage energy usage and thermal output, we didn’t know that problems existed.”
While confirming that temperatures don’t exceed specifications has been a significant focus, so is preventing over cooling. “In many cases, data centers are over cooled,” says Hoeft. “If you understand what the temperature dynamics are, you can actually shut off air conditioners and this can save a tremendous amount of energy.”
Doubling equipment power without adding space
How does this new insight translate into savings? First, it has helped the data center double the amount of processing power available without having to increase its square footage. “With the insight we gained from Tivoli Monitoring for Energy Management, we were able to move 660 square feet of equipment from our back room into the front room—almost doubling the amount of equipment in the front room without any problem.”
Additionally, the solution has enabled the organization to achieve a data center infrastructure efficiency (DCiE) metric of 84 percent (and conversely a power usage effectiveness, PUE, level of 1.19), representing the highest levels of efficiency in the industry.1 Fundamental in accomplishing this has been the new insight, which has enabled data center staff to resolve energy “leaks” and reduce energy usage and, subsequently, energy costs.
“A 1.19 PUE is a leadership number in the industry,” says Hoeft. “There are very few companies that have achieved this level. We’ve achieved this through the data center changes along with ongoing monitoring and management of our energy footprint.”
Most importantly, the combination of IBM systems, software and technologies have helped the data center avoid equipment failure due to problematic environmental conditions. “Above all, IT organizations are measured on delivering continuous service,” says Hoeft. “That’s one of the reasons our business scenario resonates with companies that tour our facility. Every data center will continue to grow the capacity it delivers. Energy and thermal management enables that to happen without having to guess whether the equipment is within spec or the room is out of cooling capacity.”
- Reduced the amount of power required for cooling racks by 50 percent
- Doubled the amount of processing power available without having to increase square footage
- Achieved a data center infrastructure efficiency (DCiE) level of 84 percent
- Avoided equipment failures to deliver continuous service to its users
- IBM® Maximo® Asset Management for Energy Optimization
- IBM Tivoli® Business Service Manager
- IBM Tivoli Data Warehouse
- IBM Tivoli Monitoring for Energy Management
- IBM Systems Director Active Energy Manager
- IBM Tivoli Netcool®/OMNIbus
- IBM BladeCenter® HS22
- IBM Power® 570
- IBM System x®
- IBM System z10®
- IBM System Storage® DS8000®, DS4000®
- IBM System Storage N series
- IBM System Storage TS3100 Tape Library Express
- IBM XIV® Storage System
- IBM Systems and Technology Group Lab Services—Measurement and Management Technologies
For more information
To learn more about how IBM can help you transform your business, please contact your IBM sales representative or IBM Business Partner.
Visit us at: ibm.com/tivoli
Additionally, financing solutions from IBM Global Financing can enable effective cash management, protection from technology obsolescence, improved total cost of ownership and return on investment. Also, our Global Asset Recovery Services help address environmental concerns with new, more energy-efficient solutions. For more information on IBM Global Financing, visit: ibm.com/financing
IBM products and services that were used in this case study.
Tivoli Netcool/OMNIbus, Maximo Asset Management for Energy Optimization, Tivoli Business Service Manager, Tivoli Data Warehouse, Tivoli Monitoring for Energy Management, IBM Systems Director Active Energy Manager
STG Lab Services: Other
© Copyright IBM Corporation 2011 IBM Corporation Software Group Route 100 Somers, NY 10589 U.S.A. Produced in the United States of America March 2011 All Rights Reserved IBM, the IBM logo, ibm.com, Let’s build a smarter planet, smarter planet, the planet icons and Tivoli are trademarks of International Business Machines Corporation in the United States, other countries or both. If these and other IBM trademarked terms are marked on their first occurrence in this information with a trademark symbol (® or ™), these symbols indicate U.S. registered or common law trademarks owned by IBM at the time this information was published. Such trademarks may also be registered or common law trademarks in other countries. A current list of IBM trademarks is available on the web at “Copyright and trademark information” at ibm.com/legal/copytrade.shtml Other company, product and service names may be trademarks or service marks of others. References in this publication to IBM products or services do not imply that IBM intends to make them available in all countries in which IBM operates. 1 Data center infrastructure efficiency (DCiE) is calculated by dividing IT equipment power by total facility power. Power Usage Effectiveness (PUE) is the reciprocal of DCiE, calculating efficiency by dividing total facility power by IT equipment power. Levels of efficiency for these benchmarks are as follows: Very inefficient=PUE of 3.0 and DCiE of 33%; Inefficient=PUE of 2.5 and DCiE of 40%; Average=PUE of 2.0 and DCiE of 50%; Efficient= PUE of 1.5 and DCiE of 67%. Very Efficient=PUE of 1.2 and DCiE of 83%. 2 IBM mobile measurement technology was developed by IBM Watson Research Center and is available through IBM Systems and Technology Group Lab Services and IBM Global Business Services. TIC14173-USEN-00