TW201514675A - Method and system for controlling power supply of data center - Google Patents
Method and system for controlling power supply of data center Download PDFInfo
- Publication number
- TW201514675A TW201514675A TW102131918A TW102131918A TW201514675A TW 201514675 A TW201514675 A TW 201514675A TW 102131918 A TW102131918 A TW 102131918A TW 102131918 A TW102131918 A TW 102131918A TW 201514675 A TW201514675 A TW 201514675A
- Authority
- TW
- Taiwan
- Prior art keywords
- motherboard
- power supply
- data center
- bmc
- power consumption
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 19
- 238000010586 diagram Methods 0.000 description 4
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Landscapes
- Power Sources (AREA)
Abstract
Description
本發明涉及一種控制方法及系統,尤其是涉及一種資料中心電源調配方法及系統。The invention relates to a control method and system, in particular to a data center power supply method and system.
資料中心(Data Center)是由很多伺服器和記憶體主機組成的。所有這些數量龐大的伺服器和記憶體主機,使供電模組一直處於重負荷工作狀態。如果供電模組中一個電源裝置出現故障,有可能會導致供電不足,使整個資料中心不能正常穩定工作,所帶來的損失是無法估量的。業界為了避免這種情況的發生,通常的做法是使用冗餘電源裝置,當有電源裝置出現故障時,冗餘電源裝置接替故障電源裝置的工作,以保證整個資料中心正常穩定工作。但是,這種方式下很難確定配備多少數目的冗餘電源裝置才是最合理的。如果配備的冗餘電源裝置數目比較多,雖然可以保證資料中心正常穩定工作,但是事實上很多時候冗餘電源裝置都沒有發生作用,這會造成資源的浪費,增加成本。另外,即使已經配備冗餘電源裝置,有時仍無法避免整個資料中心無法正常工作的可能事件發生。例如只配備了兩個冗餘電源裝置,但是卻有三個電源裝置發生故障,則有一個故障電源裝置的工作無法接替。The Data Center is made up of many servers and memory hosts. All of these large number of servers and memory hosts keep the power supply module under heavy load. If a power supply unit in the power supply module fails, the power supply may be insufficient, and the entire data center may not work normally stably. The loss caused is incalculable. In order to avoid this situation, the industry usually uses redundant power supply devices. When a power supply device fails, the redundant power supply device takes over the work of the faulty power supply device to ensure normal and stable operation of the entire data center. However, in this way it is difficult to determine how many redundant power supply units are equipped to be the most reasonable. If the number of redundant power supply units is relatively large, although the data center can be guaranteed to work normally, in fact, many times redundant power supply devices do not work, which causes waste of resources and increases costs. In addition, even with redundant power supplies, it is sometimes impossible to avoid possible events where the entire data center is not working properly. For example, only two redundant power supply units are equipped, but three power supply units fail, and one failed power supply unit cannot be replaced.
鑒於以上內容,有必要提供一種資料中心電源調配方法及系統,可以智慧管理整個資料中心的供電狀況。In view of the above, it is necessary to provide a data center power distribution method and system, which can intelligently manage the power supply status of the entire data center.
所述資料中心電源調配方法包括:確定步驟:從資料中心的所有基板管理控制器BMC中確定一個主BMC;讀取步驟:控制主BMC透過電源管理匯流排從資料中心的供電模組中定期讀取整個資料中心的總功耗;選取步驟:當所讀取的總功耗已達到供電模組的供電上限時,根據預定原則選取資料中心中需要關閉的主機板;及控制步驟:發送控制命令至所選取的主機板,控制所選取的主機板暫時關閉。The data center power supply method includes: determining step: determining a master BMC from all the baseboard management controllers BMC in the data center; and reading step: controlling the master BMC to periodically read from the power supply module of the data center through the power management bus bar Take the total power consumption of the entire data center; select the step: when the total power consumption read has reached the power supply upper limit of the power supply module, select the motherboard to be closed in the data center according to the predetermined principle; and control step: send a control command To the selected motherboard, the selected motherboard is temporarily closed.
所述資料中心電源調配系統包括:確定模組,用於從資料中心的所有基板管理控制器BMC中確定一個主BMC;讀取模組,用於控制主BMC透過電源管理匯流排從資料中心的供電模組中定期讀取整個資料中心的總功耗;選取模組,用於當所讀取的總功耗已達到供電模組的供電上限時,根據預定原則選取資料中心中需要關閉的主機板;及控制模組,用於發送控制命令至所選取的主機板,控制所選取的主機板暫時關閉。The data center power distribution system includes: a determining module for determining a main BMC from all the baseboard management controllers BMC of the data center; and a reading module for controlling the main BMC to pass through the power management bus from the data center. The power module periodically reads the total power consumption of the entire data center; the module is selected to select the host to be shut down in the data center according to a predetermined principle when the total power consumption read has reached the power supply upper limit of the power supply module. And a control module, configured to send a control command to the selected motherboard, and control the selected motherboard to be temporarily closed.
相較於習知技術,本發明所述之資料中心電源調配方法及系統,可以提供一種節能、智慧的解決方案,當資料中心的總功耗達到供電模組所能提供的電能上限時,透過選擇性地讓某些主機板暫時關閉來降低功耗,以避免整個資料中心不能正常工作。Compared with the prior art, the data center power supply method and system of the present invention can provide an energy-saving and intelligent solution. When the total power consumption of the data center reaches the upper limit of the power that the power supply module can provide, Selectively turn off some motherboards temporarily to reduce power consumption to prevent the entire data center from working properly.
圖1係為本發明資料中心電源調配系統較佳實施方式之應用環境圖。1 is an application environment diagram of a preferred embodiment of a data center power distribution system of the present invention.
圖2係為本發明資料中心電源調配系統較佳實施方式之功能模組圖。2 is a functional block diagram of a preferred embodiment of a data center power distribution system of the present invention.
圖3係為本發明資料中心電源調配方法較佳實施方式之流程圖。3 is a flow chart of a preferred embodiment of a data center power distribution method according to the present invention.
圖4係為本發明中各個BMC計算主機板的平均電源功耗和CPU平均負載之流程圖。FIG. 4 is a flow chart of average power consumption and average CPU load of each BMC computing motherboard in the present invention.
參閱圖1所示,係為本發明資料中心電源調配系統較佳實施方式之應用環境圖。所述資料中心電源調配系統10運行於控制電腦1中,所述控制電腦1透過網路與資料中心2連接。所述控制電腦1中還包括透過資料匯流排相連的儲存器11和處理器12。所述資料中心2中包括多個伺服器或記憶體主機(圖中未示出),該多個伺服器或記憶體主機中包括多個主機板20(圖中以兩個為例)及供電模組30,每個主機板20中包括BMC(Baseboard Management Controller,基板管理控制器)21及電源管理晶片(例如INA219AIDR)22。Referring to FIG. 1 , it is an application environment diagram of a preferred embodiment of a data center power distribution system of the present invention. The data center power distribution system 10 operates in a control computer 1, and the control computer 1 is connected to the data center 2 via a network. The control computer 1 further includes a storage 11 and a processor 12 connected through a data bus. The data center 2 includes a plurality of servers or memory hosts (not shown), and the plurality of server or memory hosts include a plurality of motherboards 20 (two in the figure) and power supply The module 30 includes a BMC (Baseboard Management Controller) 21 and a power management chip (for example, INA219AIDR) 22 in each motherboard 20.
所述供電模組30由多個電源裝置組成,每個電源裝置可以為資料中心2中的一個或多個主機板20進行供電。所述各個主機板20中的BMC 21與供電模組30之間以電源管理匯流排(PMBus)連接,所述BMC 21可以透過該電源管理匯流排對供電模組30進行讀寫訪問。所述供電模組30的供電線路經過電源管理晶片22再連接到各個主機板20,同時將電源管理晶片22掛載在各個BMC 21的系統管理匯流排(SMBus)上。所述電源管理晶片22中包括特定的寄存器,則所述BMC 21透過系統管理匯流排讀取該寄存器可以獲知經過該電源管理晶片22的電流和電壓的大小,從而計算出各個主機板20的電源功耗。The power supply module 30 is composed of a plurality of power supply devices, and each of the power supply devices can supply power to one or more of the motherboards 20 in the data center 2. The BMC 21 of each of the motherboards 20 and the power supply module 30 are connected by a power management bus (PMBus), and the BMC 21 can read and write the power supply module 30 through the power management bus. The power supply lines of the power supply module 30 are connected to the respective motherboards 20 via the power management chip 22, and the power management chips 22 are mounted on the system management bus (SMBus) of each BMC 21. The power management chip 22 includes a specific register, and the BMC 21 reads the register through the system management bus bar to know the current and voltage of the power management chip 22, thereby calculating the power of each motherboard 20. Power consumption.
所述儲存器11用於儲存所述資料中心電源調配系統10的程式碼等資料。所述處理器12用於執行所述資料中心電源調配系統10的各功能模組,以完成本發明。可以理解,所述控制電腦1中還應該包括其他必要的硬體系統與軟體系統,如主機板、作業系統等,由於這些設備都是本領域技術人員的習知常識,本實施方式中不再一一描述。The storage device 11 is configured to store data such as code of the data center power distribution system 10. The processor 12 is configured to execute various functional modules of the data center power distribution system 10 to complete the present invention. It can be understood that the control computer 1 should also include other necessary hardware systems and software systems, such as a motherboard, an operating system, etc., since these devices are common knowledge of those skilled in the art, in this embodiment, One by one description.
參閱圖2所示,係為本發明資料中心電源調配系統較佳實施方式之功能模組圖。Referring to FIG. 2, it is a functional module diagram of a preferred embodiment of the data center power distribution system of the present invention.
所述資料中心電源調配系統10包括確定模組100、讀取模組200、判斷模組300、選取模組400及控制模組500。The data center power distribution system 10 includes a determination module 100, a reading module 200, a determination module 300, a selection module 400, and a control module 500.
所述確定模組100用於從資料中心2的所有BMC 21中確定一個主BMC。在本實施方式中,按照BMC 21的ID號最小的原則確定主BMC,剩餘的BMC 21為從BMC,主BMC負責管理各個從BMC。The determining module 100 is configured to determine one master BMC from all BMCs 21 of the data center 2. In the present embodiment, the primary BMC is determined according to the principle that the ID number of the BMC 21 is the smallest, and the remaining BMC 21 is the secondary BMC, and the primary BMC is responsible for managing each secondary BMC.
所述讀取模組200用於控制主BMC透過電源管理匯流排從供電模組30中定期讀取整個資料中心2的總功耗。The reading module 200 is configured to control the main BMC to periodically read the total power consumption of the entire data center 2 from the power supply module 30 through the power management bus.
所述判斷模組300用於判斷所讀取的總功耗是否已達到供電模組30的供電上限(即供電模組30能提供的最大功耗)。The determining module 300 is configured to determine whether the total power consumption read has reached the power supply upper limit of the power supply module 30 (ie, the maximum power consumption that the power supply module 30 can provide).
所述選取模組400用於當所讀取的總功耗已達到供電模組30的供電上限時,根據預定原則選取可以關閉的主機板20。在本實施方式中,所述預定原則包括主機板20的平均電源功耗和CPU(圖中未示出)平均負載等。The selection module 400 is configured to select the motherboard 20 that can be turned off according to a predetermined principle when the total power consumption read has reached the power supply upper limit of the power supply module 30. In the present embodiment, the predetermined principle includes the average power consumption of the motherboard 20 and the average load of the CPU (not shown) and the like.
當以主機板20的平均電源功耗為原則選取可以關閉的主機板20時,所述選取模組400從各個從BMC中讀取各個主機板20的平均電源功耗,然後選取平均電源功耗最低的主機板20。當以CPU平均負載為原則選取可以關閉的主機板20時,所述選取模組400從各個從BMC中讀取各個主機板20的CPU平均負載,然後選取CPU平均負載最低的主機板20。When the motherboard 20 that can be turned off is selected based on the average power consumption of the motherboard 20, the selection module 400 reads the average power consumption of each motherboard 20 from each slave BMC, and then selects the average power consumption. The lowest motherboard 20. When the motherboard 20 that can be turned off is selected based on the average CPU load, the selection module 400 reads the average CPU load of each motherboard 20 from each slave BMC, and then selects the motherboard 20 with the lowest average CPU load.
所述平均電源功耗和CPU平均負載均為一定時間段內的平均值,各個BMC 21週期性地獲取對應主機板20的電源功耗和CPU的負載,並更新主機板20的平均電源功耗和CPU平均負載。當某一主機板20的平均電源功耗和CPU平均負載較低時,說明該主機板20在這一時間段內處於相對空閒的狀態。但是,平均電源功耗和CPU平均負載較低並不能說明該主機板20當前的電源功耗和CPU負載也較低。因此為了兼顧即時性的考慮,各個BMC 21每隔預定時間(例如30分鐘)將對應主機板20的平均電源功耗和CPU平均負載清零並重新計算。具體計算過程參閱圖4所示。The average power consumption and the average CPU load are average values in a certain period of time, and each BMC 21 periodically acquires the power consumption of the corresponding motherboard 20 and the load of the CPU, and updates the average power consumption of the motherboard 20 . And the average CPU load. When the average power consumption and the average CPU load of a certain motherboard 20 are low, it indicates that the motherboard 20 is in a relatively idle state during this period of time. However, the average power consumption and the average CPU load are not sufficient to indicate that the current power consumption and CPU load of the motherboard 20 are also low. Therefore, in order to take into account the immediacy consideration, each BMC 21 clears and recalculates the average power consumption and CPU average load of the corresponding motherboard 20 every predetermined time (for example, 30 minutes). The specific calculation process is shown in Figure 4.
所述控制模組500用於發送控制命令至所選取的主機板20,控制所選取的主機板20暫時關閉,並發送提醒資訊給管理員。The control module 500 is configured to send a control command to the selected motherboard 20, control the selected motherboard 20 to be temporarily turned off, and send reminder information to the administrator.
所述判斷模組300還用於當所述讀取模組200所讀取的總功耗未達到供電模組30的供電上限時,判斷是否存在被動關機的主機板20。所述被動關機指主機板20接收到所述控制模組500發送的控制命令而暫時關閉。當存在被動關機的主機板20時,所述判斷模組300根據主BMC重新讀取的資料中心2當前的總功耗預估所述被動關機的主機板20能否重新開啟。The determining module 300 is further configured to determine whether there is a passively shut down motherboard 20 when the total power consumption read by the reading module 200 does not reach the power supply upper limit of the power supply module 30. The passive shutdown means that the motherboard 20 receives the control command sent by the control module 500 and is temporarily turned off. When there is a passively shut down motherboard 20, the determining module 300 predicts whether the passively powered motherboard 20 can be re-opened according to the current total power consumption of the data center 2 re-read by the primary BMC.
在本實施方式中,當存在被動關機的主機板20時,每次主BMC讀取資料中心2的總功耗後,所述判斷模組300都會計算讓所述被動關機的主機板20啟動後,供電模組30能否保證資料中心2中所有主機板20正常工作。主要是透過判斷運算式P_C+m*P_A<P_S是否成立來進行預估,其中P_C表示資料中心2當前的總功耗,P_S表示供電模組30能提供的最大功耗,P_A表示所述被動關機的主機板20的最大功耗。為了保證各個主機板20均能穩定工作,m的值要比1略大(例如1.1),具體數值需要經過多次試驗總結。In this embodiment, when there is a passively shut down motherboard 20, each time the main BMC reads the total power consumption of the data center 2, the determining module 300 calculates the motherboard 20 that is passively turned off. Whether the power supply module 30 can ensure that all the motherboards 20 in the data center 2 work normally. The estimation is mainly performed by judging whether the arithmetic expression P_C+m*P_A<P_S is established, wherein P_C represents the current total power consumption of the data center 2, P_S represents the maximum power consumption that the power supply module 30 can provide, and P_A represents the passive power. The maximum power consumption of the motherboard 20 that is turned off. In order to ensure that each motherboard 20 can work stably, the value of m is slightly larger than 1 (for example, 1.1), and the specific values need to be summarized by many experiments.
所述控制模組500還用於當預估出所述被動關機的主機板20可以重新開啟時,發送控制命令至所述被動關機的主機板20,控制所述被動關機的主機板20重新開啟。The control module 500 is further configured to: when it is estimated that the passively shut down motherboard 20 can be re-enabled, send a control command to the passively shut down motherboard 20, and control the passively shut down the motherboard 20 to be turned back on. .
參閱圖3所示,係為本發明資料中心電源調配方法較佳實施方式之流程圖。Referring to FIG. 3, it is a flowchart of a preferred embodiment of a power distribution method for a data center of the present invention.
步驟S10,所述確定模組100從資料中心2的所有BMC 21中確定一個主BMC。In step S10, the determining module 100 determines one master BMC from all the BMCs 21 of the data center 2.
步驟S12,所述讀取模組200控制主BMC透過電源管理匯流排從供電模組30中定期讀取整個資料中心2的總功耗。In step S12, the reading module 200 controls the main BMC to periodically read the total power consumption of the entire data center 2 from the power supply module 30 through the power management bus.
步驟S14,所述判斷模組300判斷所讀取的總功耗是否已達到供電模組30的供電上限(即供電模組30能提供的最大功耗)。若所讀取的總功耗已達到供電模組30的供電上限,則執行步驟S16。若所讀取的總功耗未達到供電模組30的供電上限,則執行步驟S20。In step S14, the determining module 300 determines whether the total power consumption read has reached the power supply upper limit of the power supply module 30 (ie, the maximum power consumption that the power supply module 30 can provide). If the total power consumption read has reached the upper power supply limit of the power supply module 30, step S16 is performed. If the total power consumption read does not reach the power supply upper limit of the power supply module 30, step S20 is performed.
步驟S16,所述選取模組400根據預定原則選取可以關閉的主機板20。In step S16, the selection module 400 selects the motherboard 20 that can be closed according to a predetermined principle.
步驟S18,所述控制模組500發送控制命令至所選取的主機板20,控制所選取的主機板20暫時關閉,並發送提醒資訊給管理員。In step S18, the control module 500 sends a control command to the selected motherboard 20, controls the selected motherboard 20 to be temporarily turned off, and sends reminder information to the administrator.
步驟S20,所述判斷模組300判斷是否存在被動關機的主機板20。若存在被動關機的主機板20,則執行步驟S22。若不存在被動關機的主機板20,則返回步驟S12。In step S20, the determining module 300 determines whether there is a passively turned off motherboard 20. If there is a passively shut down motherboard 20, step S22 is performed. If there is no passively shut down motherboard 20, then return to step S12.
步驟S22,所述判斷模組300根據主BMC重新讀取的資料中心2當前的總功耗預估所述被動關機的主機板20能否重新開啟。若預估出所述被動關機的主機板20可以重新開啟,則執行步驟S24。若預估出所述被動關機的主機板20不能重新開啟,則返回步驟S12。In step S22, the determining module 300 predicts whether the passively powered motherboard 20 can be re-opened according to the current total power consumption of the data center 2 re-read by the primary BMC. If it is estimated that the passively shut down motherboard 20 can be turned back on, step S24 is performed. If it is estimated that the passively shut down motherboard 20 cannot be turned back on, it returns to step S12.
步驟S24,所述控制模組500發送控制命令至所述被動關機的主機板20,控制所述被動關機的主機板20重新開啟。In step S24, the control module 500 sends a control command to the passively shut down motherboard 20 to control the passively shutdown motherboard 20 to be turned back on.
參閱圖4所示,係為本發明中各個BMC 21計算主機板的平均電源功耗和CPU平均負載之流程圖。在本實施方式中,所述BMC 21主要是指從BMC。Referring to FIG. 4, it is a flowchart of calculating the average power consumption and the average CPU load of the motherboard of each BMC 21 in the present invention. In the present embodiment, the BMC 21 mainly refers to a slave BMC.
步驟S600,所述BMC 21啟動總時間為預定時間T的計時。In step S600, the BMC 21 starts the time when the total time is the predetermined time T.
步驟S602,所述BMC 21讀取對應CPU的當前負載,並計算從計時開始到當前時間這一時間段內的CPU平均負載。Step S602, the BMC 21 reads the current load of the corresponding CPU, and calculates the average CPU load during the period from the start of the counting to the current time.
步驟S604,所述BMC 21透過系統管理匯流排從所述電源管理晶片22中讀取電壓和電流資料。在本實施方式中,所述BMC 21每隔預設間隔時間(例如一分鐘)讀取一次對應CPU的當前負載和電源管理晶片22中的電壓和電流資料。In step S604, the BMC 21 reads voltage and current data from the power management chip 22 through the system management bus. In the present embodiment, the BMC 21 reads the current load of the corresponding CPU and the voltage and current data in the power management chip 22 every predetermined interval time (for example, one minute).
步驟S606,所述BMC 21根據所讀取的電壓和電流資料計算對應主機板20的電源功耗,並計算從計時開始到當前時間這一時間段內的平均電源功耗。Step S606, the BMC 21 calculates the power consumption of the corresponding motherboard 20 based on the read voltage and current data, and calculates the average power consumption during the period from the start of the timer to the current time.
步驟S608,所述BMC 21判斷是否接收到主BMC發送的控制命令。若接收到控制命令,則執行步驟S610。若未接收到控制命令,則執行步驟S612。In step S608, the BMC 21 determines whether a control command sent by the primary BMC is received. If a control command is received, step S610 is performed. If the control command is not received, step S612 is performed.
步驟S610,所述BMC 21根據接收到的控制命令開啟或關閉對應的主機板20,然後執行步驟S612。In step S610, the BMC 21 turns on or off the corresponding motherboard 20 according to the received control command, and then performs step S612.
步驟S612,所述BMC 21判斷當前計時是否達到預定時間T。若達到則執行步驟S614,若未達到則返回步驟S602。In step S612, the BMC 21 determines whether the current timing has reached the predetermined time T. If yes, step S614 is performed, and if not, step S602 is returned.
步驟S614,所述BMC 21將對應主機板20的平均電源功耗和CPU平均負載資料清零。In step S614, the BMC 21 clears the average power consumption and CPU average load data of the corresponding motherboard 20.
使用本發明資料中心電源調配方法及系統,可以在供電模組30發生供電不足的時候,按預先約定好的策略,根據各個主機板20的平均電源功耗或CPU平均負載選擇一塊主機板20讓其關機,以保證資料中心2中盡可能多的主機板20能正常穩定工作。By using the data center power distribution method and system of the present invention, when the power supply module 30 is insufficiently powered, a motherboard board 20 can be selected according to the average power consumption of the main board 20 or the average CPU load according to a pre-agreed strategy. It is turned off to ensure that as many motherboards 20 as possible in the data center 2 can work normally and stably.
綜上所述,本發明符合發明專利要件,爰依法提出專利申請。惟,以上所述者僅爲本發明之較佳實施方式,本發明之範圍並不以上述實施方式爲限,舉凡熟悉本案技藝之人士爰依本發明之精神所作之等效修飾或變化,皆應涵蓋於以下申請專利範圍內。In summary, the present invention complies with the requirements of the invention patent and submits a patent application according to law. However, the above description is only a preferred embodiment of the present invention, and the scope of the present invention is not limited to the above-described embodiments, and equivalent modifications or variations made by those skilled in the art in accordance with the spirit of the present invention are It should be covered by the following patent application.
1‧‧‧控制電腦1‧‧‧Control computer
2‧‧‧資料中心2‧‧‧Data Center
10‧‧‧資料中心電源調配系統10‧‧‧Data Center Power Supply System
11‧‧‧儲存器11‧‧‧Storage
12‧‧‧處理器12‧‧‧ Processor
20‧‧‧主機板20‧‧‧ motherboard
21‧‧‧BMC21‧‧‧BMC
22‧‧‧電源管理晶片22‧‧‧Power Management Wafer
30‧‧‧供電模組30‧‧‧Power supply module
100‧‧‧確定模組100‧‧‧Determining modules
200‧‧‧讀取模組200‧‧‧Reading module
300‧‧‧判斷模組300‧‧‧Judgement module
400‧‧‧選取模組400‧‧‧Selection module
500‧‧‧控制模組500‧‧‧Control Module
無no
10‧‧‧資料中心電源調配系統 10‧‧‧Data Center Power Supply System
100‧‧‧確定模組 100‧‧‧Determining modules
200‧‧‧讀取模組 200‧‧‧Reading module
300‧‧‧判斷模組 300‧‧‧Judgement module
400‧‧‧選取模組 400‧‧‧Selection module
500‧‧‧控制模組 500‧‧‧Control Module
Claims (12)
確定步驟:從資料中心的所有基板管理控制器BMC中確定一個主BMC;
讀取步驟:控制主BMC透過電源管理匯流排從資料中心的供電模組中定期讀取整個資料中心的總功耗;
選取步驟:當所讀取的總功耗已達到供電模組的供電上限時,根據預定原則選取資料中心中需要關閉的主機板;及
控制步驟:發送控制命令至所選取的主機板,控制所選取的主機板暫時關閉。A data center power supply method, the method comprising:
Determining step: determining a master BMC from all baseboard management controllers BMC in the data center;
The reading step: controlling the main BMC to periodically read the total power consumption of the entire data center from the power supply module of the data center through the power management bus;
Selection step: when the total power consumption read has reached the power supply upper limit of the power supply module, the motherboard to be closed in the data center is selected according to a predetermined principle; and the control step is: sending a control command to the selected motherboard, the control station The selected motherboard is temporarily turned off.
所述預定原則包括主機板的平均電源功耗和CPU平均負載;
當以主機板的平均電源功耗為原則選取需要關閉的主機板時,從各個從BMC中讀取各個主機板的平均電源功耗,然後選取平均電源功耗最低的主機板作為需要關閉的主機板;
當以CPU平均負載為原則選取需要關閉的主機板時,從各個從BMC中讀取各個主機板的CPU平均負載,然後選取CPU平均負載最低的主機板作為需要關閉的主機板。For example, the data center power supply method described in claim 1 of the patent scope, wherein:
The predetermined principle includes an average power consumption of the motherboard and an average CPU load;
When selecting the motherboard to be shut down based on the average power consumption of the motherboard, read the average power consumption of each motherboard from each BMC, and then select the motherboard with the lowest average power consumption as the host to be shut down. board;
When the motherboard to be shut down is selected based on the average CPU load, the average CPU load of each motherboard is read from each BMC, and then the motherboard with the lowest average CPU load is selected as the motherboard to be shut down.
所述資料中心中各個主機板中的BMC與供電模組之間以電源管理匯流排連接,所述BMC透過該電源管理匯流排對供電模組進行讀寫訪問;
所述供電模組的供電線路經過電源管理晶片再連接到各個主機板,並且所述電源管理晶片掛載在各個BMC的系統管理匯流排上,所述BMC透過系統管理匯流排從該電源管理晶片中獲取經過該電源管理晶片的電流和電壓的大小,從而計算出各個主機板的電源功耗。For example, the data center power supply method described in claim 2, wherein:
A power management bus is connected between the BMC and the power supply module in each of the data centers, and the BMC reads and writes the power supply module through the power management bus.
The power supply line of the power supply module is connected to each motherboard through a power management chip, and the power management chip is mounted on a system management bus of each BMC, and the BMC manages the bus from the power management chip through the system management bus. The current and voltage of the power management chip are obtained to calculate the power consumption of each motherboard.
當所讀取的總功耗未達到供電模組的供電上限時,判斷是否存在被動關機的主機板;
當存在被動關機的主機板時,根據主BMC重新讀取的資料中心當前的總功耗預估所述被動關機的主機板能否重新開啟;及
當預估出所述被動關機的主機板可以重新開啟時,發送控制命令至所述被動關機的主機板,控制所述被動關機的主機板重新開啟。The data center power supply method according to claim 1, wherein the method further comprises the steps of:
When the total power consumption read does not reach the power supply upper limit of the power supply module, it is determined whether there is a passively shut down motherboard;
When there is a passively shut down motherboard, whether the passively shutdown motherboard can be re-opened according to the current total power consumption of the data center re-read by the main BMC; and when the passive shutdown of the motherboard is estimated When re-opening, a control command is sent to the passively shut down motherboard, and the passively shut down motherboard is restarted.
確定模組,用於從資料中心的所有基板管理控制器BMC中確定一個主BMC;
讀取模組,用於控制主BMC透過電源管理匯流排從資料中心的供電模組中定期讀取整個資料中心的總功耗;
選取模組,用於當所讀取的總功耗已達到供電模組的供電上限時,根據預定原則選取資料中心中需要關閉的主機板;及
控制模組,用於發送控制命令至所選取的主機板,控制所選取的主機板暫時關閉。A data center power distribution system, the system comprising:
Determining a module for determining a primary BMC from all baseboard management controllers BMC of the data center;
The reading module is configured to control the main BMC to periodically read the total power consumption of the entire data center from the power supply module of the data center through the power management bus;
Selecting a module for selecting a motherboard to be closed in the data center according to a predetermined principle when the total power consumption read has reached the power supply upper limit of the power supply module; and a control module for transmitting the control command to the selected The motherboard is controlled to temporarily close the selected motherboard.
所述預定原則包括主機板的平均電源功耗和CPU平均負載;
當以主機板的平均電源功耗為原則選取需要關閉的主機板時,從各個從BMC中讀取各個主機板的平均電源功耗,然後選取平均電源功耗最低的主機板作為需要關閉的主機板;
當以CPU平均負載為原則選取需要關閉的主機板時,從各個從BMC中讀取各個主機板的CPU平均負載,然後選取CPU平均負載最低的主機板作為需要關閉的主機板。For example, the data center power distribution system described in claim 7 of the patent scope, wherein:
The predetermined principle includes an average power consumption of the motherboard and an average CPU load;
When selecting the motherboard to be shut down based on the average power consumption of the motherboard, read the average power consumption of each motherboard from each BMC, and then select the motherboard with the lowest average power consumption as the host to be shut down. board;
When the motherboard to be shut down is selected based on the average CPU load, the average CPU load of each motherboard is read from each BMC, and then the motherboard with the lowest average CPU load is selected as the motherboard to be shut down.
所述資料中心中各個主機板中的BMC與供電模組之間以電源管理匯流排連接,所述BMC透過該電源管理匯流排對供電模組進行讀寫訪問;
所述供電模組的供電線路經過電源管理晶片再連接到各個主機板,並且所述電源管理晶片掛載在各個BMC的系統管理匯流排上,所述BMC透過系統管理匯流排從該電源管理晶片中獲取經過該電源管理晶片的電流和電壓的大小,從而計算出各個主機板的電源功耗。For example, the data center power distribution system described in claim 8 of the patent scope, wherein:
A power management bus is connected between the BMC and the power supply module in each of the data centers, and the BMC reads and writes the power supply module through the power management bus.
The power supply line of the power supply module is connected to each motherboard through a power management chip, and the power management chip is mounted on a system management bus of each BMC, and the BMC manages the bus from the power management chip through the system management bus. The current and voltage of the power management chip are obtained to calculate the power consumption of each motherboard.
判斷模組,用於當所讀取的總功耗未達到供電模組的供電上限時,判斷是否存在被動關機的主機板;
所述判斷模組還用於當存在被動關機的主機板時,根據主BMC重新讀取的資料中心當前的總功耗預估所述被動關機的主機板能否重新開啟;及
所述控制模組還用於當預估出所述被動關機的主機板可以重新開啟時,發送控制命令至所述被動關機的主機板,控制所述被動關機的主機板重新開啟。The data center power distribution system of claim 7, wherein the system further comprises:
The determining module is configured to determine whether there is a passively shut down motherboard when the total power consumption read does not reach the power supply upper limit of the power supply module;
The determining module is further configured to: when there is a passively shutdown motherboard, predict whether the passively powered motherboard can be re-opened according to the current total power consumption of the data center re-read by the primary BMC; and the control module The group is further configured to send a control command to the passively shut down motherboard when the motherboard that is passively shut down is estimated to be re-opened, and control the passively shut down the motherboard to be turned back on.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW102131918A TW201514675A (en) | 2013-09-05 | 2013-09-05 | Method and system for controlling power supply of data center |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW102131918A TW201514675A (en) | 2013-09-05 | 2013-09-05 | Method and system for controlling power supply of data center |
Publications (1)
Publication Number | Publication Date |
---|---|
TW201514675A true TW201514675A (en) | 2015-04-16 |
Family
ID=53437596
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW102131918A TW201514675A (en) | 2013-09-05 | 2013-09-05 | Method and system for controlling power supply of data center |
Country Status (1)
Country | Link |
---|---|
TW (1) | TW201514675A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107682223A (en) * | 2017-09-29 | 2018-02-09 | 郑州云海信息技术有限公司 | A kind of method of quick test power management bus functionality |
-
2013
- 2013-09-05 TW TW102131918A patent/TW201514675A/en unknown
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107682223A (en) * | 2017-09-29 | 2018-02-09 | 郑州云海信息技术有限公司 | A kind of method of quick test power management bus functionality |
CN107682223B (en) * | 2017-09-29 | 2021-09-28 | 郑州云海信息技术有限公司 | Method for rapidly testing power management bus function |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110221946B (en) | Method and apparatus for power analysis of a storage system | |
US9304562B2 (en) | Server rack system and power management method applicable thereto | |
US9195588B2 (en) | Solid-state disk (SSD) management | |
US7363517B2 (en) | Methods and apparatus to manage system power and performance | |
TWI410789B (en) | System on a chip and method for the same | |
TWI355580B (en) | Power control system of a high density server and | |
US10852795B2 (en) | Method for system power management and computing system thereof | |
US8694804B2 (en) | Computer system and power management method thereof | |
KR20110004328A (en) | Power management by selective approval of high power states of computer system hardware devices | |
US10788872B2 (en) | Server node shutdown | |
CN111142644A (en) | Hard disk operation control method and device and related components | |
US20160349812A1 (en) | Managing Power in a High Performance Computing System for Resiliency and Cooling | |
CN101634882B (en) | High-density server power supply control system and method thereof | |
US20170220354A1 (en) | Server node shutdown | |
CN112230755A (en) | Power management method, device, equipment and machine-readable storage medium | |
CN104423528A (en) | Power allocation method and power allocation system for data centre | |
CN107533348A (en) | For the power in elasticity and cooling management high performance computing system | |
US9639144B2 (en) | Power state adjustment | |
TW201514675A (en) | Method and system for controlling power supply of data center | |
TW201322697A (en) | Baseboard management controller electronic device and controlling method thereof | |
CN110096224A (en) | Power control in memory sub-system | |
WO2014196473A1 (en) | Server system and method and program for controlling same | |
CN103139107A (en) | Base plate management controller, electronic device and control method of electronic device | |
CN116483402A (en) | A power module upgrading method, server, electronic equipment and storage medium | |
TW201640275A (en) | Deep off power states |