Product energy efficiency

Product energy efficiency was formalized as one of the company's corporate objectives when IBM established its product stewardship program in 1991. Through the collaboration of IBM Research and our product development teams, we have combined hardware and software technologies to improve the energy efficiency of IT equipment and in turn, data centers.

One of IBM's product energy efficiency goals is to improve the computing power delivered for each kilowatt-hour (kWh) of electricity used by each new generation of server. In 2017, IBM released its POWER9™-based Power Systems™ Accelerated Compute (AC922) server for high-performance computing analytics and artificial intelligence. When compared to comparable IBM POWER8® products, IBM POWER9-based servers have Server Efficiency Rating Tool (SERT) weighted geomean active efficiency scores up to three times higher -- which represents three times the performance or work delivered without any increase in power use. The SERT was created by the Standard Performance Evaluation Corporation. IBM also introduced its next-generation mainframe, the IBM z14™ server, in 2017. On average, the IBM z14 server delivers 23 percent or more work per kilowatt depending on the choice of components and cooling method.

IBM also has a goal to qualify its new server and storage products to the U.S. Environmental Protection Agency's ENERGY STAR® program criteria where practical, and where criteria have been developed for the specific server or storage product type. Visit our ENERGY STAR certified products page for more information.

In addition to its ongoing program and objectives regarding energy efficient products, IBM has initiated and invested in innovations and integrated solutions involving its hardware, software and services business to address the energy efficiency of IT equipment and the data center. Visit our Data center energy efficiency page for more information.

IBM continues to utilize innovations in semiconductor, hard drive/storage system and networking technologies to improve server and storage system performance for each unit of power consumed by the equipment. For more information, see:


IBM Power Systems

IBM announced its first server product featuring the POWER9 processor family, the IBM Power Systems Accelerated Compute Server AC922, in December 2017. The AC922 has two POWER9 processors with up to two terabytes of memory and support for up to six graphics processing units. The POWER9 servers are designed for high-performance computing, analytics, and artificial intelligence (AI) workloads. Initial SERT test data from servers with POWER9 processors indicate that the servers' performance-to-power efficiency increased by a factor of up to three over comparable POWER8-based products. The increased performance-to-power capabilities of the POWER9 servers enable them to deliver more workload than previous-generation servers, offering clients an opportunity to consolidate and reduce the number of servers required to perform a given workload and to decrease the energy demand and consumption of their data center operations.

IBM Power Systems client efficiency examples:

IBM Power Systems servers can make material improvements in the efficiency of data center operations. Since the POWER9 servers were not released until the end of 2017, the examples below detail improved efficiencies of client operations achieved using POWER8 products:

  • One client deployed two S824 servers and one E870 server to replace its legacy IT infrastructure and Enterprise Resource Planning system. The Power Systems servers reduced the number of physical servers required compared to competitors' offerings, resulting in less space, maintenance, power and cooling requirements and more energy efficient operation. The system upgrade also delivered 73 to 83 percent increased performance thereby improving customer service and internal administrative processing times.
  • Another client upgraded their server systems to a POWER8 platform, reducing the number of servers needed to support their workload by 50 percent with a commensurate reduction in energy consumption.

IBM Z servers

The IBM z14™ server, announced in 2017, is designed for the cloud and is the only server that can deliver 100 percent encryption of application, cloud service and database data without changes to applications. It is capable of managing blockchain, AI and high-volume transactional workloads securely and efficiently. On average, the z14 delivers 23 percent or more work per kilowatt (kW) than the IBM z13® through increased processor capacity and performance, a 3.2x increase in memory capacity and the ability to store twice as much data on the same storage disk space. The z14 server also offers several energy- saving options.

  • Capability to operate using high-voltage direct current power, which reduces conversion losses and increases energy efficiency
  • A water-cooled model, which delivers 6 percent more computing capacity per kW compared to the air- cooled model
  • Capability to operate at temperatures of up to 40 degrees Celsius for defined periods of time, enabling use of direct air cooling to reduce cooling costs

With its high utilization rates, the z14 offers one of the most efficient computing platforms when measured by the workload delivered per unit of energy consumed.


Storage systems

IBM continues to improve its range of storage products -- including the IBM FlashSystem® 900, the IBM XIV®, the Storwize® family, the DS8880 enterprise storage family, and tape systems -- offering clients solutions across all their data storage needs. IBM storage products are supported by software-defined storage and capacity optimization methods (COMs) that maximize the utilization of available storage capacity and assign data to the storage tier commensurate with the importance of the data. COM functions include software-based data management capabilities such as Easy Tier®, thin provisioning, data compression and de-duplication, and storage virtualization. These capabilities can reduce the storage hardware and energy footprint as well as the capacity required to accomplish a given storage task.

IBM has also continued to expand its software-defined Spectrum Scale™ storage offerings, which enable storage automation and virtualization in both on-premise and cloud environments. Spectrum Scale storage enables the reduction of storage energy consumption and costs through data consolidation and the use of data placement technologies to optimize the use of available storage devices, including tape storage. The desired outcome is to maximize the amount of data stored on a minimum number of storage products, in turn minimizing energy use.

Storage systems client solution examples:

  • One client, a provider of marketing as a service solutions, implemented IBM Spectrum Accelerate™ on IBM Cloud to deliver a high-availability, easily expandable, enterprise storage solution for its clients. Application performance was increased by 15 to 20 percent while hardware requirements and operating expenses, including energy cost, were reduced by up to 60 percent.
  • A client in the medical services field upgraded a portion of their legacy data storage equipment to IBM FlashSystem 900 products, reducing the hardware footprint and energy consumption by over 80 percent. The FlashSystem storage was integrated with IBM XIV and Storwize products using IBM Spectrum Virtualize™ to enable significant data compression and optimization of data storage across the hardware products, significantly increasing the quantity of data stored on the system. In addition, the upgrades enabled the client team to achieve sub-millisecond response times, improving the productivity of their medical staff.

These are just two examples of the improvements in storage utilization and reduction in energy use that can be achieved through the deployment of new storage technologies utilizing COMs capabilities and software-defined storage capabilities such as the IBM Spectrum Accelerate and Virtualize software. Similar productivity and energy efficiency improvements are being achieved across the thousands of IBM storage systems that IBM sells each year.