1、 AMERICAN NATIONAL STANDARD FOR TELECOMMUNICATIONS ATIS-0600015.01.2014 Energy Efficiency for Telecommunication Equipment: Methodology for Measurement and Reporting Server Requirements As a leading technology and solutions development organization, ATIS brings together the top global ICT companies t
2、o advance the industrys most-pressing business priorities. Through ATIS committees and forums, nearly 200 companies address cloud services, device solutions, emergency services, M2M communications, cyber security, ehealth, network evolution, quality of service, billing support, operations, and more.
3、 These priorities follow a fast-track development lifecycle from design and innovation through solutions that include standards, specifications, requirements, business use cases, software toolkits, and interoperability testing. ATIS is accredited by the American National Standards Institute (ANSI).
4、ATIS is the North American Organizational Partner for the 3rd Generation Partnership Project (3GPP), a founding Partner of oneM2M, a member and major U.S. contributor to the International Telecommunication Union (ITU) Radio and Telecommunications sectors, and a member of the Inter-American Telecommu
5、nication Commission (CITEL). For more information, visit . AMERICAN NATIONAL STANDARD Approval of an American National Standard requires review by ANSI that the requirements for due process, consensus, and other criteria for approval have been met by the standards developer. Consensus is established
6、 when, in the judgment of the ANSI Board of Standards Review, substantial agreement has been reached by directly and materially affected interests. Substantial agreement means much more than a simple majority, but not necessarily unanimity. Consensus requires that all views and objections be conside
7、red, and that a concerted effort be made towards their resolution. The use of American National Standards is completely voluntary; their existence does not in any respect preclude anyone, whether he has approved the standards or not, from manufacturing, marketing, purchasing, or using products, proc
8、esses, or procedures not conforming to the standards. The American National Standards Institute does not develop standards and will in no circumstances give an interpretation of any American National Standard. Moreover, no person shall have the right or authority to issue an interpretation of an Ame
9、rican National Standard in the name of the American National Standards Institute. Requests for interpretations should be addressed to the secretariat or sponsor whose name appears on the title page of this standard. CAUTION NOTICE: This American National Standard may be revised or withdrawn at any t
10、ime. The procedures of the American National Standards Institute require that action be taken periodically to reaffirm, revise, or withdraw this standard. Purchasers of American National Standards may receive current information on all standards by calling or writing the American National Standards
11、Institute. Notice of Disclaimer they are commonly found in many networks. Examples of servers include but are not limited to the following: File server Mail server Database server Authentication server Web server Media server Game server Servers are often characterized by workloads that require the
12、server to process, convert, or otherwise service requests from network-connected equipment. 5 Metric Definition 5.1 Preamble SPEC has defined a power efficiency metric called SPECpower_ssj2008 to measure the performance of servers across a variety of different load points. This benchmark measures se
13、rver side Java performance. More information on this benchmark can be found at http:/www.spec.org. SPECpower supports several different models of power meters and automatically produces an official report for benchmark runs (see the SPEC web site for one example report). The SPECpower_ssj2008 benchm
14、ark runs on a system under test, with a separate server running as a controller to collect information from the system under test and from the power analyzer. Figure 1 below (adapted from the SPECpower_ssj2008 User Guide) shows a basic configuration: 3This benchmark is available from the Standard Pe
15、rformance Evaluation Corporation, 6585 Merchant Place, Suite 100, Warrenton, VA 20187, USA ATIS-0600015.01.2014 3 Figure 5. 1 - Typical SPECpower_ssj2008 Configuration SPECpower is currently defined to work with AC-powered systems. While early versions of the benchmark could work with DC-powered sys
16、tems with selected power meters, the benchmark has been updated to enforce the requirement to work with AC-powered systems only. For DC-powered systems, a modified configuration can be used. ATIS-0600015.01.2014 4 Figure 5. 2 - SPECpower_ssj2008 Configuration for DC-powered Systems When DC systems a
17、re tested with SPECpower, the results will include the power consumed by the rectifier used during the tests. The power consumed by the rectifier must be removed from the test results to get the TEER value. ATIS-0600015.04 defines a method for determining the rectifier efficiency; the rectifier must
18、 be tested to this method and the ATIS-0600015.04 results included with the test report for DC-powered servers. The duration of the test is controlled by the SPECpower_ssj2008 software in the controller. This may deviate from the duration specified in the general standard (ATIS-0600015.2013). 5.2 Se
19、rver TEER The SPECpower framework allows multiple different parameters to be adjusted for internal benchmarking runs, but all results used for compliance with this specification must meet the requirements called out in the SPECpower Run and Reporting Rules document, with the notable deviation that D
20、C-powered servers will be tested with rectifier power included. During each benchmark cycle, SPECpower performs several calibration runs before testing the system at different load levels (100%, 90%, 80%, 70%, 60%, 50%, 40%, 30%, 20%, 10%, and “Active Idle”). SPECpower computes the average power con
21、sumption at each of these load levels. Refer to Figure 3 below for an example. ATIS-0600015.01.2014 5 SPECpower_ssj2008 - TransactionsWoodcrest 3.0, 4x 1GB, 1x HDD, Pwr Mgmt OnSPECpower_ssj2008 Transactions xx Server, 4x1GB, 1x HDD, Pwr Mgmt On PeakThroughput40%20%100%ActiveIdlecalibration levels80%
22、Power(Watts)Figure 5. 3 -Example SPECpower Workload Iteration 5.2.1 Determining the TEER Value The SPECpower output document shall be included with the reported test data. The standard report provides the average power consumption and work performed at each load level shown in Figure 2 above. In add
23、ition, SPECpower_ssj2008 computes an overall SPECpower_ssj2008 value that is the sum of the work performed at each level divided by the sum of the average power consumption at each level. 5.2.1.1 TEER for AC-powered Systems For AC-powered systems, the TEER value shall be calculated as follows, round
24、ed to the nearest whole value: TEER = ssj_ops / power / 10 For example, a product with a SPECpower_ssj2008 rating of 1088 would have a TEER value of 109. 5.2.1.2 TEER for DC-powered Systems For DC-powered systems, the TEER value must account for the power consumed by the rectifier. Assume a SPECpowe
25、r run for a given DC server produced the data shown in Table 5.1. Table 5. 1- Example DC Server Data from SPECpower Tool Performance PowerPerformance to Power Ratio Target Load Actual Load ssj_ops Average Active Power (W)100% 99.8% 22,884,076 3,086 7,41590% 89.9% 20,617,898 2,722 7,57680% 80.0% 18,3
26、42,838 2,249 8,15670% 70.1% 16,059,905 2,041 7,86860% 60.0% 13,744,552 1,817 7,56450% 50.0% 11,456,289 1,608 7,12340% 40.0% 9,176,474 1,418 6,47030% 30.0% 6,879,431 1,238 5,55520% 20.0% 4,583,962 1,069 4,28710% 10.0% 2,296,016 908 2,528Active Idle 0 536 0ssj_ops / power = 6,743If this had been an AC
27、-powered system, its TEER value would have been 674. However, the DC-powered system for this test was used with a 5 kW rectifier that had the following efficiency at the load levels specified in ATIS-0600015.04: ATIS-0600015.01.2014 6 Table 5. 2 - Example 5 kW Rectifier Efficiency (tested per ATIS-0
28、600015.04) % of Full Load Efficiency (%) Watts 10% 93.1% 500 20% 95.5% 100030% 96.2% 1500 40% 96.4% 200050% 96.5% 2500 60% 96.4% 300070% 96.1% 3500 80% 96.0% 400090% 95.7% 4500 100% 95.4% 5000It can be seen that a value such as 1238 watts is 238W above the 1000 W value shown in Table 5.2. Since the
29、efficiency at 1000W is 95.5% and the efficiency at 1500W is 96.2%, the average change in efficiency per watt is 0.0014%/W. This means that the interpolated efficiency should be 95.5%+ (0.0014%/W * 238W) = 95.8%. Multiplying the average active AC power of 1238 watts (measured by the AC power meter) b
30、y 0.958 (the interpolated 95.8% efficiency factor) gives an average active DC power of 1186 watts. Using linear interpolation like the example above, a modified form of Table 5.1 can be produced, as shown in Table 5.3: Table 5.3 Example of Calculated DC Server Data Performance PowerTarget Load Actua
31、l Load ssj_ops Average Active AC Power (W) Rectifier Efficiency Average Active DC Power (W) 100% 99.8% 22,884,076 3,086 96.3% 2,972 90% 89.9% 20,617,898 2,722 96.5% 2,626 80% 80.0% 18,342,838 2,249 96.4% 2,168 70% 70.1% 16,059,905 2,041 96.4% 1,968 60% 60.0% 13,744,552 1,817 96.3% 1,750 50% 50.0% 11
32、,456,289 1,608 96.2% 1,547 40% 40.0% 9,176,474 1,418 96.1% 1,363 30% 30.0% 6,879,431 1,238 95.8% 1,186 20% 20.0% 4,583,962 1,069 95.6% 1,022 10% 10.0% 2,296,016 908 95.1% 864 Active Idle 0 536 93.3% 500 ssj_ops = 126,041,442 DC_power = 17,966 ssj_ops / DC_power = 7,015 For DC-powered systems, the TE
33、ER value shall be calculated as follows, rounded to the nearest whole value: TEER = (ssj_ops / DC_power) / 10 For example, a product with the data shown in Table 5.3 would have a TEER value of 702. ATIS-0600015.01.2014 7 6 Test Procedure 6.1 Equipment Configuration The test methodology and test cond
34、itions used to independently verify the power consumption shall meet the requirements called out in both the SPECpower Run and Reporting Rules document and ATIS-0600015.2013, as well as this document. Where conflicts occur, this document and the general requirements standard take precedence. 7 Repor
35、ting & Documentation 7.1 Specific Measurements Outside the Scope of the General Document All products claiming compliance with this specification shall provide the information as defined in ATIS-0600015.2013. In addition, the standard report produced by SPECpower_ssj2008 shall be appended to the doc
36、ument. DC-powered equipment shall also append the rectifier test data as defined in ATIS-0600015.04 and the corresponding data similar to Table 5.3 of this document. Note that the SPEC organization has defined requirements for public disclosure of SPECpower results. Any public disclosure of such results is subject to the terms of the SPECpower license.
