You are here

Covered Product Category: Data Center Storage

The Federal Energy Management Program (FEMP) provides acquisition guidance for data center storage, a product category covered by the ENERGY STAR program. Federal laws and requirements mandate that agencies buy ENERGY STAR-qualified products in all product categories covered by this program and any acquisitions that are not specifically exempted by law.

MEETING EFFICIENCY REQUIREMENTS FOR FEDERAL PURCHASES

The U.S. Environmental Protection Agency (EPA) sets the efficiency criteria for this product category in its ENERGY STAR program requirements. Manufacturers meeting these requirements are allowed to display the ENERGY STAR label on complying models. Visit the ENERGY STAR website for the most up-to-date data center storage efficiency levels and product specification information, and a list of qualified products.

DETERMINING WHEN ENERGY STAR IS COST-EFFECTIVE

The current ENERGY STAR data center storage specification (Version 1.0) contains only one efficiency requirement, which applies to power supply efficiency. Other qualification criteria in the specification pertain to product features and reporting requirements (e.g., on active and idle state efficiency), but not power efficiency.1 The cost difference for power supply units in ENERGY STAR-qualified versus non-ENERGY STAR data center storage products is considered minimal. Therefore, federal purchasers can assume that there is no production cost difference between ENERGY STAR-qualified and non-ENERGY STAR data center storage products.

COMPLYING WITH CONTRACTING REQUIREMENTS

These requirements apply to all forms of procurement, including construction guide specifications and project specifications; renovation, repair, maintenance, and energy service contracts; lease agreements; acquisitions made using purchase cards; and solicitations for offers. Energy efficiency requirements should be included in both the evaluation criteria of solicitations and the evaluations of solicitation responses.

Federal Acquisition Regulation (FAR) Part 23.206 requires federal agencies to insert the clause at FAR section 52.223-15 in solicitations and contracts that deliver, acquire, furnish, or specify energy-consuming products. FEMP recommends that agencies incorporate efficiency requirements into both the technical specification and evaluation sections of solicitations. Agencies may claim an exception to these requirements through a written finding that no ENERGY STAR-qualified or FEMP-designated product is available to meet the functional requirements, or that no such product is life cycle cost-effective for the specific application. Additional information on federal requirements is available.

BUYER TIPS: CHOOSING EFFICIENT PRODUCTS

Agencies should consult the Center of Expertise for Energy Efficiency in Data Centers website for information on energy-efficient design strategies and opportunities in data centers. DOE has partnered with key public and private stakeholders to provide technical information, tools, best practices, and analysis that assist government agencies with decreasing energy use in data centers. Purchasing efficient products for use in data centers can be an important component to meeting data center energy reduction goals.

Other factors to consider when purchasing data center storage include the following.

STORAGE PRODUCT SELECTION

The current ENERGY STAR data center storage specification (Version 1.0) includes Online 2, 3, and 4 storage products (among other storage product types such as near-online storage and removable media library). Buyers should investigate their individual storage needs before selecting the storage product types that maximize performance and energy cost savings for the specific applications.

SOLID STATE DRIVES

Solid state drives (SSDs) feature faster read and access speeds, but are more expensive than conventional hard disks. If SSDs are too costly for the application, buyers should consider purchasing systems that use only the smaller 2.5-inch form factor hard disk drives, which provide the same performance as the traditional 3.5-inch form factor drives but consume less energy.

CAPACITY OPTIMIZING METHODS

Capacity optimizing methods (COMs) are technologies used to improve data storage by shrinking stored data, thereby reducing energy consumption. COMs may include thin provisioning, data duplication, compression, and delta snapshots, as defined in the ENERGY STAR data center storage specification. Buyers can consult with product vendors or IT professionals for the type(s) of COMs suited for the application and workload.

USER TIPS: USING PRODUCTS MORE EFFICIENTLY

Buyers can track the power used by data storage products and implement power-saving technologies to optimize equipment efficiency.

POWER DATA MEASUREMENTS

Per the current ENERGY STAR data center storage specification (Version 1.0), Online 3 and 4 storage products are required to measure and report input power in watts through the full range of operation. Data are available to users in formats readable by third-party, non-proprietary management systems, over a standard network connection, and via embedded components or add-in devices that are packaged with the storage product. Users can make use of collected data to facilitate equipment selection and optimize operation strategies.

MASSIVE ARRAY OF IDLE DISKS

Many hard disks run on a continuous basis—this not only wears out the mechanical parts, but also consumes energy as long as the disks are running. For data that are accessed very infrequently and do not demand a fast response time, Massive Array of Idle Disks (MAID) is an alternative in which only those disk drives in active use are spinning at any given time, and hard disks not accessed for a longer period are switched off. If used properly, the MAID technology reduces power consumption and prolongs disk lifespan.

Lawrence Berkeley National Laboratory provided supporting analysis for this acquisition guidance.

Updated January 2015

1 EPA intends to set active and idle state efficiency criteria in Version 2.0, using data generated as a result of the Version 1.0 specification.