ITU-T SERIES L SUPP 7-2014 ITU-T L 1300 C Supplement on rationale for minimum data set for evaluating energy efficiency and for controlling data centre equipment in view of power s.pdf

1、 I n t e r n a t i o n a l T e l e c o m m u n i c a t i o n U n i o n ITU-T Series L TELECOMMUNICATION STANDARDIZATION SECTOR OF ITU Supplement 7 (12/2014) SERIES L: CONSTRUCTION, INSTALLATION AND PROTECTION OF CABLES AND OTHER ELEMENTS OF OUTSIDE PLANT ITU-T L.1300 Supplement on rationale for mini

2、mum data set for evaluating energy efficiency and for controlling data centre equipment in view of power saving ITU-T L-series Recommendations Supplement 7 L series Supplement 7 (12/2014) i Supplement 7 to ITU-T L-series Recommendations ITU-T L.1300 Supplement on rationale for minimum data set for e

3、valuating energy efficiency and for controlling data centre equipment in view of power saving Summary Supplement 7 to ITU-T L-series Recommendations describes the rationale for a minimum data set for evaluating energy efficiency and for controlling data centre equipment with a view to power saving b

4、ased on Recommendation ITU-T L.1300. More precisely, this Supplement reports the data set necessary for: evaluation of energy efficiency and coordinated control to save power in data centres. Finally, it provides a summary of the minimum data set and gap analysis with other standards. History Editio

5、n Recommendation Approval Study Group Unique ID* 1.0 ITU-T L Suppl. 7 2014-12-19 5 11.1002/1000/12435 Keywords Best practice, data centre, energy efficient, information and communication technology and climate change (ICT the data set necessary for evaluation of energy efficiency in data centres; th

6、e data set necessary for coordinated control to save power in data centres; and a summary of the minimum data set and gap analysis with other standards. 2 Definitions This Supplement uses the following terms: 2.1 dynamic data: Data that should be obtained from data centre equipment periodically. 2.2

7、 static data: Data that should be treated statically. 3 Abbreviations and acronyms This Supplement uses the following abbreviations and acronyms: COP Coefficient Of Performance CRAC Computer Room Air Conditioner CRAH Computer Room Air Handler DCiE Data Centre infrastructure Efficiency DPPE Data cent

8、re Performance Per Energy ECMA European Computer Manufactures Association GEC Green Energy Coefficient HVAC Heating, Ventilation and Air Conditioning ICT Information and Communication Technology ITEE IT equipment Energy Efficiency ITEU IT Equipment Utilization KVM Keyboard, Video monitor, and Mouse

9、MIB Management Information Base PDU Power Distribution Unit PUE Power Usage Effectiveness RDU Rack Distribution Unit UPS Uninterruptible Power Supply 2 L series Supplement 7 (12/2014) VAV Variable Air Volume 4 Introduction Energy consumption in data centres is increasing year by year with the growth

10、 in their numbers. In order to mitigate global warming, reduction of power consumption in data centres is urgent. Generally, information and communication technology (ICT) equipment (e.g., servers, routers, switches, storage units) and facility equipment (e.g., power delivery components, heating, ve

11、ntilation and air conditioning (HVAC) system components) account for a large percentage of energy consumption in data centres. Thus, power-saving measures for these components are necessary to reduce overall power consumption. Evaluation of energy efficiency in data centres is necessary to know how

12、green the data centre is and to investigate the effect of power-saving measures. To evaluate whether the data centre is power efficient, it is necessary to continuously trace metrics that represent the energy efficiency of the data centre. An example of such a metric is power usage effectiveness (PU

13、E) or its reciprocal, data centre infrastructure efficiency (DCiE), both of which express the energy efficiency of facility equipment. In addition to power-saving measures for each component, energy consumption in data centres can be further reduced by implementing “coordinated control“ of ICT equip

14、ment and facility equipment. In general, ICT equipment and facility equipment are separately controlled. If facility equipment is controlled in a coordinated way according to the arrangement of the workload of ICT equipment, further power reduction can be achieved. To evaluate energy efficiency in d

15、ata centres and to control their equipment to achieve power saving, it is necessary to collect data from the data centre equipment. Currently, sensors can be installed to collect the data necessary, but the installation cost can be very high. It is also possible to obtain monitoring information dire

16、ctly from data centre equipment by using, for example, a management information base (MIB) for ICT equipment. However, it is difficult to collect the data necessary in an integrated fashion because a data centre is generally a multivendor environment, and each vendor defines its own measurement poin

17、t and allocation address. Consequently, this Supplement describes the minimum data set necessary for evaluating energy efficiency and for controlling data centre equipment to save power in data centres. The data set includes two types of data: 1. data that should be obtained from data centre equipme

18、nt periodically (dynamic data); 2. data that should be treated statically (static data). The definition of such a minimum data set is intended to facilitate the evaluation and control of equipment under a multivendor environment. The means of collecting the data from data centre equipment is outside

19、 the scope of this Supplement and should be addressed as a future task. This Supplement considers: Dynamic data: Data that should be obtained from data centre equipment periodically. Static data: Data that should be treated statically. 5 Data set necessary for evaluation of energy efficiency in data

20、 centres This clause describes the minimum data set necessary for evaluating energy efficiency in data centres. The data set necessary is described, based on the analysis of the following metrics for evaluating energy efficiency of data centres. Table 1 lists examples of metrics, which represent the

21、 energy efficiency of a data centre. Table 2 shows the relation between the metrics and the parameters. L series Supplement 7 (12/2014) 3 Table 1 Examples of metrics representing energy efficiency of data centres Metrics Definitional identity Notes PUE: power usage effectiveness Total energy consump

22、tion of data centre/energy consumption of ICT equipment. PUE is proposed by the Green Grid b-TGG WP14, and determines the energy efficiency of a data centre infrastructure. The reciprocal of PUE is DCiE: data centre infrastructure efficiency. ITEU: IT equipment utilization Energy consumption of ICT

23、equipment/total rated power of ICT equipment. IT equipment utilization (ITEU) is proposed by the Green IT Promotion Council b-GITPC-DPPE, and represents the degree of energy saving by virtual and operational techniques using the potential ICT equipment capacity reducing waste capacity. A reduction i

24、n equipment to be installed is promoted by using only the number of devices needed to meet the required ICT capacity reducing waste capacity. ITEE: IT equipment energy efficiency (Total server capacity + Total storage capacity + Total NW equipment capacity)/total rated power of ICT equipment. IT equ

25、ipment energy efficiency (ITEE) is proposed by the Green IT Promotion Council b-GITPC-DPPE, and aims to promote energy saving by encouraging the installation of equipment with high processing capacity per unit of electric power. GEC: Green energy coefficient Green (natural energy) energy/total energ

26、y consumption of data centre. Green energy coefficient (GEC) is proposed by the Green IT Promotion Council b-GITPC-DPPE, and is a value obtained by dividing green energy produced and used in a data centre by the total power consumption. It has been introduced to promote the use of green energy. DPPE

27、: Data centre performance per energy Function of the four metrics PUE, ITEU, ITEE, GEC. Data centre performance per energy (DPPE) is proposed by the Green IT Promotion Council , and indicates the energy efficiency of data centres as a whole. It is expressed as the product of the other four metrics.

28、Table 2 Relation between metrics and parameters Parameters used in the metrics Type of information PUE ITEU ITEE GEC DPPE Total energy consumption of data centre Dynamic Energy consumption of ICT equipment Dynamic Green energy produced and used in data centre Dynamic Rated power of ICT equipment Sta

29、tic Server capacity Static NW equipment capacity Static Storage capacity Static In Table 2, energy consumption of ICT equipment includes all hardware, i.e., computing, storage, and network equipment, along with supplemental equipment, i.e., keyboard, video monitor, and mouse (KVM) switches, monitors

30、, and workstations or laptops used to monitor or otherwise control the data centre. On the other hand, total data centre energy consumption includes all of the ICT equipment power plus everything that supports it, such as power delivery components, HVAC system components, and other miscellaneous com

31、ponent loads, like physical security and building management systems. Table 3 gives examples of components quoted from b-TGG WP14. 4 L series Supplement 7 (12/2014) Table 3 Examples of components of facility and ICT equipment Facility Power Transfer switch Uninterruptible power supply (UPS) DC batte

32、ries or rectifiers (non-UPS telco nodes) Generator Transformer (step down) Power distribution unit (PDU) Rack distribution unit (RDU) Breaker panels Distribution wiring Lighting Heating, ventilation and air conditioning (HVAC) Cooling tower Condenser water pumps Chillers Chilled water pumps Computer

33、 room air conditioners (CRACs) Computer room air handlers (CRAHs) Dry cooler Supply fans Return fans Air economizer Water-side economizer Humidifier In-row, in-rack and in-chassis cooling solutions Physical security Fire suppression Water detection Physical security servers or devices Building manag

34、ement system Servers or devices used to control management of data centre Probes or sensors IT equipment Computer devices Servers Network devices Switches Routers IT support systems Printers PCs/workstations Remote management (KVM/console/etc.) Miscellaneous devices Security encryption, storage encr

35、yption, appliances, etc. Storage Storage devices switches, storage array Backup devices media libraries, virtual media libraries Telecommunications All telco devices Source: b-TGG WP14 Table 2 indicates that the dynamic data necessary for calculating these metrics are the total energy consumption of

36、 the data centre and the energy consumption of ICT equipment, both of which are used in the calculation of PUE. Green energy produced and used in the data centre is also dynamic data, but does not need to be included because the scope of this Supplement includes a data set for L series Supplement 7

37、(12/2014) 5 evaluating power efficiency, not energy efficiency. Thus, consideration of the source of power is outside the scope of this Supplement. Other parameters in Table 2 are static data that can be obtained from equipment specifications. To determine the total energy consumption of the data ce

38、ntre and the energy consumption of ICT equipment, some approaches to measuring PUE are described in b-TGG WP14, as shown in Table 4. PUE calculation is more precise at higher levels, and locations to measure annual energy consumption of the data centre and of ICT equipment, as well as the minimum me

39、asurement interval, are defined for each level. Figure 1 shows an example of a power feeding system in a data centre with the measurement locations indicated. The higher the level, the shorter the measurement interval recommended, and thus, it would be easier to obtain energy consumption of ICT equi

40、pment if the input power of ICT equipment and the output power of power equipment (i.e., UPS, rectifier, PDU) could be collected as dynamic data from the equipment. On the other hand, for total power consumption, it is recommended that input power be measured in the utility meter when calculating PU

41、E. In addition, cooling equipment accounts for a large percentage of energy consumption in data centres, thus its input power should also be collected. Therefore, inclusion of the input power of ICT equipment and cooling equipment, and the output power of the UPS, rectifier and PDU to the data set,

42、is necessary to evaluate the power efficiency. Table 4 Measurement approaches of PUE Level 1 (basic) Level 2 (intermediate) Level 3 (advanced) IT equipment power Total facility power where: UPS Data centre input power PDU Data centre input power less shared HVAC Server Data centre input power less s

43、hared HVAC plus building lighting, security Minimum measurement interval 1 month/1 week Daily Continuous (XX min) Figure 1 Example of a power feeding system of a data centre 6 Data set necessary for coordinated control to save power in data centres This clause describes the minimum data set necessar

44、y for coordinated control to reduce the total energy consumption of data centres. First, the effect of coordinated control is explained in clause 6.1. 6 L series Supplement 7 (12/2014) Next, the description of the control architecture and the scope of the data set follow in clause 6.2. Then, clause

45、6.3 describes the details of the data set on ICT equipment, cooling equipment, power equipment, and the equipment configuration. 6.1 Power saving due to coordinated control The effects of controlling multiple pieces of cooling equipment in a coordinated way are described here. The efficiency of a co

46、mputer room air conditioner (CRAC) varies with the load factor and operation mode. When each CRAC is controlled to satisfy its own temperature setting, the overall efficiency of all CRACs is not necessarily high. As shown in Figure 2, when one of the CRACs in variable air volume (VAV) mode has a dis

47、proportionate load and the thermostats of adjacent CRACs remain switched off, the overall efficiency may be low. An experiment was conducted to see if the total energy consumption could be reduced by adjusting the load balance among the CRACs. The temperature setting of each CRAC was adjusted so the

48、y would operate at higher efficiency. Figure 2 Illustration of a server room (coefficient of performance (COP) The correlation between the outside temperature and the energy consumption of CRACs is shown in Figure 3, which illustrates that the energy consumption of CRACs can be reduced by adjusting

49、the load balance. For example, when the outside temperature is 15C, the reduction in energy consumption of CRACs is estimated to be 17.5%. Note that this estimated value is only valid for the experimental environment, but the result indicates that coordinated control among CRACs makes it possible to save a large amount of power. For power saving, it is important to control multiple CRACs by optimally responding to the heat generation due to the operation of ICT equipment. L series Supplement 7