JP2010165265A - Device, method and program for distributing server processing - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To suppress power consumption of a server system having a plurality of server apparatuses. <P>SOLUTION: A load distribution control part 2 transfers a received request from a client to one of the server apparatuses 1-1 to 1-N. A service processing position control part 4 obtains temperatures, operation states, performance information, and spatial position information of the server apparatuses 1-1 to 1-N, and a load distribution weighting parameter of the load distribution control part 2 from a service processing position information holding part 5, calculates new operation states of the server apparatuses 1-1 to 1-N and a new load distribution weighting parameter of the load distribution control part 2, and controls the server apparatuses 1-1 to 1-N. When the server apparatus having an overheat state higher than a proper temperature is present, the server processing distribution device 100A adjusts load distribution, and next adds the server apparatus to distribute processing. When the server apparatus having a low temperature state less than the proper temperature is present, the server processing distribution device 100A suspends the server apparatus, and next adjust the load distribution to integrate the processing. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a server processing distribution device, a server processing distribution method, and a server processing distribution program, and more particularly, a plurality of servers that receive a request from a client and execute a process (service), load distribution processing based on the spatial position of each server The present invention relates to a server processing distribution device, a server processing distribution method, and a server processing distribution program.

  A server processing distribution apparatus in which a plurality of servers perform load distribution processing is required to suppress unnecessary power consumption. Therefore, for example, in the server processing distribution device shown in Patent Document 1, as shown in the block diagram shown in FIG. 22, a plurality of servers each having a server power supply control unit operating with a power supply different from the server power supply. 914-1 to 914-4 and the load distribution device 913 that relays the client 911 have a predetermined configuration.

  The load balancer 913 receives a data request from the client 911 instead of the servers 914-1 to 914-4 and receives the received data request from any of the servers 914-1 to 914-4. From the data request transfer unit 923 to be transferred, the data supply amount measurement unit 926 that measures the amount of data that the servers 914-1 to 914-4 supply to the client 911, and the received data request and the measurement result, A server information measuring unit 922 that determines the number of operations and issues a command to the data request transfer unit 923 as to which of the servers 914-1 to 914-4 should be transferred, and a server stop or start command according to the determination result A power supply stop unit 924 and a power supply start unit 925 are provided.

  Further, Patent Document 2 discloses a management system as shown in the block diagram of FIG. 23 that can save energy for cooling a plurality of computers. The management system shown in FIG. 23 uses an arrangement destination selection data storage unit 933 including operation data representing the execution status of software by a computer and temperature data representing a temperature distribution, and arrangement destination selection data to other computers. Extract the overheated computer that is expected to have more heat dissipation than the other and the depopulated computer that is assumed to have less heat dissipation than the others, and move at least a portion of the software assigned to the overheated computer to the depopulated computer An instruction generation unit 932 that generates an instruction and an instruction unit 948 that outputs an instruction are provided.

  Further, Patent Document 3 does not require a device that manages the entire plurality of servers with the configuration shown in FIG. A load distribution processing system that saves power by independently adjusting a load and shifting load balancing / balancing between a plurality of servers to a sleep state or a normal state is shown. The load distribution processing system shown in FIG. 24 has a common group IP address in a server group including a plurality of servers {a, b, c}.

  Each server measures the processing load of its own server, holds the operating state, load information, sleep conditions, etc. for each server in the group, and notifies each other according to load fluctuations in the group. Each server accepts a process according to a load state in response to a process request using a group IP address from the PC. Each server appropriately switches to the sleep state in which only the auxiliary power supply is turned on by contacting the group with sleep according to the determination based on the load state and the sleep condition, turning off its main power supply.

  Patent Document 4 discloses a server processing distribution device that distributes and processes processing requests from clients to a plurality of servers for processing. On the other hand, Patent Literature 5 discloses a technology in which a thermometer for measuring exhaust temperature is provided at an exhaust port for heat radiation of a plurality of computers, and the load is distributed in descending order of load margin based on a temperature detection value and a load table. Yes.

  Therefore, the technology described in Patent Document 5 is applied to the server processing apparatus described in Patent Document 4, temperature and performance information is collected from each server, and load distribution is performed in consideration of temperature rise due to the effect of air conditioning depending on the installation position of the server. It is conceivable to perform control.

  Furthermore, based on the data for predicting a change in the external environment described in Patent Document 6 and past energy consumption patterns, a method for predicting a temporal variation pattern of energy consumption and controlling the operation pattern of the device is patented. By combining with the inventions described in Documents 4 and 5, it is assumed that stable load distribution control can be performed with predicted temperature and performance after load distribution change.

JP 2003-281008 A JP 2007-179437 A JP 2006-301749 A JP 2008-009800 A JP 2006-285317 A JP 2001-065959 A

  However, each of the server processing distribution devices disclosed in Patent Documents 1 to 3 has the following problems.

  The first problem is that it is impossible to accurately determine which process is executed in which server and which can be optimized in terms of energy. The reason is that the physical location of the computer is not considered. By concentrating the computers as much as possible and arranging the computers to apply the cool air from the air conditioner, the cooling efficiency is improved and the energy efficiency is improved. On the other hand, if the computers are concentrated too much, the temperature rises above the limit of the cooling capacity of the air conditioner, and heat accumulation occurs. Thus, it is difficult to perform proper server management unless the physical arrangement of computers is taken into consideration.

  The second problem is that it is impossible to accurately control which process and what load should be executed on which server. The reason for this is not clear which computer should be turned on / off (sleep state), and what ratio should be used when a part of the processing (load) is shared.

  The third problem is that it is impossible to stabilize the control of which process is executed on which server. The reason is that it is not presumed how much temperature the computer after the control will be and how much performance it will run. For example, immediately after turning off some of the computers and consolidating the processing, there is a possibility of performing control to turn on the computers again and distribute the processing due to overheating.

  Furthermore, in the server processing distribution apparatus that combines the inventions described in Patent Documents 4 and 5 described above, and the server processing distribution processing apparatus that combines the inventions described in Patent Documents 4 to 6, all of Patent Documents 4 to 6 are server physics. Since the spatial position information is not managed, there is a problem that a more accurate power consumption reduction effect cannot be obtained.

  SUMMARY An advantage of some aspects of the invention is that it provides a server processing distribution device, a server processing distribution method, and a server processing distribution program that reduce power consumption of a server system having a plurality of server devices.

  In order to achieve the above object, a server processing distribution apparatus according to the present invention is a computer that operates under program control, and a plurality of server devices that respond to requests from clients and a distribution unit that distributes requests from clients And collecting means for collecting temperature and performance information of each of the plurality of server devices, and holding operation information, physical space position information and load distribution settings for each of the plurality of server devices together with the collected temperature and performance information. New operation status and load distribution settings are calculated from service processing position information holding means, operation information held by the service processing position information holding means, physical space position information and load distribution settings, and temperature and performance information. Control a plurality of server devices based on the calculation means to perform, the new operating state and the load distribution setting And having a control means.

  In order to achieve the above object, a server processing distribution method according to the present invention is a computer that operates under program control, and requests from a client to a plurality of server devices that respond to requests from the client. The distribution step, the collection step for collecting the temperature and performance information of each of the plurality of server devices, the operation information, the physical space position information, and the load distribution setting for each of the plurality of server devices. A new operation state and load distribution setting are calculated from the holding step held in the holding unit together with the information, the operation information held by the holding unit, the physical space position information and the load distribution setting, and the temperature and performance information. A control system that controls multiple server devices based on the calculation step and the new operating state and load distribution settings. A step, characterized in that it comprises a.

  Furthermore, in order to achieve the above object, the server processing distributed program of the present invention is a computer that operates under program control, and requests from a client to a plurality of server devices that respond to requests from the client. The distribution step, the collection step for collecting the temperature and performance information of each of the plurality of server devices, the operation information, the physical space position information, and the load distribution setting for each of the plurality of server devices. A new operation state and load distribution setting are calculated from the holding step held in the holding unit together with the information, the operation information held by the holding unit, the physical space position information and the load distribution setting, and the temperature and performance information. Control multiple server devices based on calculation steps and new operating conditions and load distribution settings And that control step, characterized in that is executed by a different computer from the server device.

  According to the present invention, a server device that performs power control is determined in consideration of a temperature state, power consumption, performance index, and air conditioning efficiency, including the physical space position of the server device, and the process is moved. Thus, it is possible to accurately determine which process is executed in which server device and which can be optimized in terms of energy, thereby reducing power consumption of a server system having a plurality of server devices.

It is a block diagram which shows the structure of the 1st Embodiment of this invention. It is a flowchart for operation | movement description of the 1st Embodiment of this invention. FIG. 3 is a flowchart (part 1) for explaining detailed operations of some steps in FIG. 2; FIG. FIG. 3 is a flowchart (part 2) for explaining detailed operations of some steps in FIG. 2; FIG. It is a flowchart for detailed operation | movement description of the one part step in FIG. It is a flowchart for detailed operation | movement description of the one part step in FIG. It is a block diagram which shows the structure of the 2nd Embodiment of this invention. It is a flowchart (the 1) for operation | movement description of the 2nd Embodiment of this invention. It is a flowchart (the 2) for operation | movement description of the 2nd Embodiment of this invention. It is a block diagram which shows the structure of the 3rd Embodiment of this invention. It is a block diagram which shows the structure of the 4th Embodiment of this invention. It is a block diagram which shows the structure of the 5th Embodiment of this invention. It is a figure which shows the rack arrange | positioned in the server room in which the server apparatus which performs the load distribution process of one Example of this invention was accommodated, and those spatial position coordinates. FIG. 6 is a diagram illustrating a correspondence between a temperature of a server device that performs load distribution processing and a business name to be executed according to an embodiment of the present invention. It is a figure which shows the temperature distribution in the server room shown in FIG. It is a figure which shows an example of the temperature distribution in the server room shown in FIG. 13 when a part of business increases rapidly. FIG. 17 is a diagram showing the correspondence between the temperature of a server device and the name of a business to be executed when load distribution is performed for each server device according to an embodiment of the present invention in the temperature distribution state shown in FIG. 16. It is a figure which shows an example of the temperature distribution in a server room when the load distribution shown in FIG. 17 is performed. It is a figure which shows an example of the temperature distribution in a server room when some of several server apparatus in a server room will be in a low temperature state. FIG. 20 is a diagram illustrating the correspondence between the temperature of a server device and the business name to be executed or stopped when load distribution is performed on each server device according to an embodiment of the present invention in the temperature distribution state illustrated in FIG. 19. It is a figure which shows an example of the temperature distribution in a server room when the load distribution shown in FIG. 20 is performed. 10 is a block diagram of an example of a server processing distribution device shown in Patent Literature 1. FIG. 10 is a block diagram of an example of a server management apparatus shown in Patent Literature 2. FIG. It is a figure which shows an example of the load distribution processing system shown in patent document 3. FIG.

  Next, each embodiment of the present invention will be described with reference to the drawings.

(First embodiment)
FIG. 1 shows a block diagram of a first embodiment of a server processing distribution apparatus according to the present invention. The server processing distribution device 100A of this embodiment is a computer that operates under program control, and N (N is a natural number of 2 or more) server devices 1-1 that respond to a request from a client (not shown). ˜1-N, load distribution control unit 2 that distributes requests from clients, temperature / performance information collection unit 3 that collects temperature and performance information from server devices 1-1 to 1-N, and server device 1-1. ˜1-N calculates the new server operating state and load distribution setting from the operation information and load distribution setting, and controls each of the service processing position control unit 4 that performs control and the server devices 1-1 to 1-N. The service processing position information holding unit 5 holds operation information, physical space position information, and load distribution settings.

  Next, an outline of the operation of the server processing distribution apparatus 100A of this embodiment will be described.

  The server devices 1-1 to 1 -N receive the request from the client once by the load distribution control unit 2. The load distribution control unit 2 transfers the received request from the client to any of the server devices 1-1 to 1-N. Which server device is transferred is determined by a load distribution weighting parameter (ω) set from the outside. For example, there are three server devices (when N = 3), server device 1-1 (referred to as A), server device 1-2 (referred to as B), and server device 1-3 (referred to as this). C)), the load distribution weighting parameters are ωA = 0.2, ωB = 0.3, and ωC = 0.5 in that order, the request from the client is A: B: C = 2: 3: 5 so that {C, B, C, A, C, B, C, A, B, C} are sorted in order (the details of the algorithm are omitted because they are not the essence of the invention) To do.)

  Each of the server devices 1-1 to 1-N has a function of transmitting the status of the server device called a service processor to the operation management server. The service processor is independent of a main processor that executes a program that performs service processing for a client, transmits the status of the server device regardless of the power status of the main processor, and controls the power status of the main processor of the server device. .

  The temperature / performance information collection unit 3 communicates with the server devices 1-1 to 1-N (specifically, the service processors of the server devices 1-1 to 1-N), respectively, and the server devices 1-1 to 1 are connected. -Collect N temperature, operating status, and performance information. “Operating state” refers to power on / off, sleep state, and main processor operating mode. “Performance information” includes a CPU (processor) usage rate, a memory usage amount, a network communication amount, a disk input / output amount (I / O amount), and the like. These explanations are described in detail in known documents (for example, NEC Technical Bulletin, Vol.58 No.1, pp.48-52 “Server Management of Express5800”), and the operation thereof is described in the present invention. Since it is not a gist, explanation is omitted.

  The temperature / performance information collection unit 3 passes the information collected from the server devices 1-1 to 1-N to the service processing position information holding unit 5 and stores it. The service processing position information holding unit 5 also holds the physical space position information of each of the server devices 1-1 to 1-N and the load distribution weighting parameter of the load distribution control unit 2.

  The service processing position control unit 4 receives the temperature, operating state, performance information, space position information, and load distribution weighting parameters of the load distribution control unit 2 from the service processing position information holding unit 5. It acquires, calculates the new operating state of the server equipment 1-1 to 1-N and the load distribution weighting parameter of the load distribution control unit 2, and controls it.

  The temperature / performance information collection unit 3 polls the server devices 1-1 to 1-N at a certain frequency in order to collect the temperature, operating state, and performance information of the server devices 1-1 to 1-N. It may be configured to perform polling after waiting for an instruction from the service processing position control unit 4, or may wait for communication from the server devices 1-1 to 1-N.

  Next, the overall operation of the server processing distribution device 100A of this embodiment will be described in detail with reference to the flowcharts of FIG. 1 and FIGS.

  First, processing in the service processing position control unit 4 in FIG. 1 will be described. The service processing position control unit 4 first waits if a predetermined time has not elapsed since the previous control execution (step S1 in FIG. 2). The waiting time may be changed according to a control method described later. Next, the service processing position control unit 4 receives from the service processing position information holding unit 5 the temperature, operating state, performance information, space position information, and load distribution of the load distribution control unit 2 of the server devices 1-1 to 1-N. A weighting parameter is acquired (step S2 in FIG. 2). Subsequently, the service processing position control unit 4 checks whether there is a server device having a temperature higher than the appropriate temperature based on the various information acquired in step S2 (step S3 in FIG. 2). If there is any server device, the process proceeds to step S4 in FIG. 2, and if not, the process proceeds to step S7 in FIG.

  The service processing position control unit 4 determines whether or not the load distribution weighting parameter has been changed for an overheated server device having a temperature higher than the appropriate temperature (step S4 in FIG. 2). If not, the load distribution weighting parameter of the overheated server device is subtracted (step S6 in FIG. 2).

  For example, there are three server devices (when N = 3), server device 1-1 (referred to as A), server device 1-2 (referred to as B), and server device 1-3 (referred to as this). C)) in order, ωA = 0.2, ωB = 0.3, ωC = 0.5, and the server device C is in an overheated state, the subtraction process in step S6 Thus, it can be considered that the load distribution weighting parameters are ωA = 0.24, ωB = 0.36, and ωC = 0.4.

  Here, the load distribution weighting parameter of the server device C is multiplied by 0.8, and 0.1 of the reduction is proportionally distributed and added to the load distribution weighting parameters of the other servers, but the reduction ratio is 0.8. It is not necessary to be 8, and a constant subtraction may be used, or proportional distribution may be used instead of proportional distribution. The reason why the reduction is proportionally distributed and added to the load distribution weighting parameters of other server devices is that it is easier to calculate the update of the load distribution weighting parameters when the sum of the load distribution weighting parameters is constant. is there.

  On the other hand, if the service processing position control unit 4 determines that the load distribution weighting parameter is changed in step S4, the service processing position control unit 4 determines that the load distribution weighting parameter is changed in step S6 but cannot be improved. Then, another server device is activated, and part or all of the processing in the overheated server device is moved to the activated another server device (step S5 in FIG. 2). The process in step S5 will be described in detail separately.

  If the service processing position control unit 4 determines in step S3 that there is no server device having a temperature higher than the appropriate temperature, or if the subtraction process in step S6 is performed, the service processing position control unit 4 subsequently becomes lower than the appropriate temperature. It is checked whether or not a server device exists (step S7 in FIG. 2). If there is a server device having a temperature lower than the appropriate temperature, the process returns to step S8 in FIG. 2, and if not, the process returns to step S1 in FIG.

  When there is a server device whose temperature is lower than the appropriate temperature, the service processing position control unit 4 first determines whether the server device has been stopped or operated (step S8 in FIG. 2), and has not been attempted. In this case, the corresponding server device is stopped, and the process at the corresponding server device is moved to another server device (step S9 in FIG. 2).

  If it is necessary to start the server device immediately after attempting to stop / operate the server device, the server device has failed to stop / operate (reduction), and after that, the server device has not been reduced. The server device load distribution weighting parameters are added (step S10 in FIG. 2).

  For example, there are three server devices (when N = 3), server device 1-1 (referred to as A), server device 1-2 (referred to as B), and server device 1-3 (referred to as this). C)) in order, ωA = 0.2, ωB = 0.3, ωC = 0.5, and the server device A is in a low temperature state, the addition processing in step S10 Thus, it can be considered that the load distribution weighting parameters are ωA = 0.25, ωB = 0.28, and ωC = 0.47.

  Here, the load distribution weighting parameter of the server device A is multiplied by 1.25, and the increment 0.05 is proportionally distributed to the load distribution weighting parameters of the other server devices B and C and subtracted. The ratio need not be 1.25, may be a constant addition, or may be equally divided, not proportionally distributed.

  Next, the processing in step S5 in FIG. 2 will be described in detail with reference to FIG. The service processing position control unit 4 first sets the server device in the overheated state that exceeds the appropriate temperature as the server H (step S101 in FIG. 3A). Subsequently, among the server devices that are not operating and within the proper temperature range, the server device that is the closest to the server H and that is not executing the processing shared by the server H is referred to as the server K (FIG. 3). (A) Step S102). If the server K is not turned on, the server K is started by turning on the power (step S103 in FIG. 3A).

  Next, in order to distribute the load so that the service shared by the server H in the overheated state is further processed by the server K, the service processing position control unit 4 counts the number of server devices that share the processing. Is set to n, the load distribution weighting parameter (hereinafter also referred to as load distribution weighting coefficient) ωK of the server K is set as an average load distribution weighting parameter of n servers (step S104 in FIG. 3A). That is, if the sum of the load distribution weighting coefficients is “1”, the ωK is set to 1 / n.

  Next, 1 / n of the load distribution weighting parameter ωK of the server K is subtracted from the load distribution weighting parameters of the remaining (n−1) server devices. If the subtraction is performed equally, the load distribution weighting parameter ωj for each of the remaining (n−1) server devices is ωj−1 / (n × (n−1)) (FIG. 3 ( A) Step S105).

  As another load distribution method different from step S105, 1 / n of the load distribution weighting parameter ωK of the server K is subtracted by a value proportionally distributed by the remaining (n−1) server devices. You may do it. In this case, the load distribution weighting parameter ωj of each of the remaining (n−1) server devices is ωj {1- (1 / n)} (step S130 in FIG. 3B). If there are a plurality of processes shared by the server H, the same load distribution weighting parameter is updated for all the processes.

  Then, the service processing position control unit 4 applies the new setting to the server device and the load distribution control unit 2 (step S106 in FIG. 3A). If the load distribution weighting parameter is zero or very small in a certain server device, it may be determined that the server device is unnecessary and the server device may be suspended.

  Next, the process in step S9 in FIG. 2 will be described in detail with reference to FIG. The service processing position control unit 4 first sets the server device in the low temperature state to be reduced that is below the appropriate temperature as the server L (step S151 in FIG. 4A). If a plurality of server devices are targeted, the servers L are sequentially set from the server far from the center of gravity of the space of the active server. Further, when the server device in a low temperature state to be reduced performs only one service on the system, if it is reduced, there is no server device that provides that service.

  Subsequently, the service processing position control unit 4 updates the load distribution weighting parameter so that the remaining server devices n share the load distribution weighting parameter ωL of the server L to be reduced. Thereby, the load distribution weighting parameter ωj of each of the remaining n server devices becomes ωj + (ωL / n) (step S152 in FIG. 4A).

  As another load distribution method different from step S152, the load distribution weighting parameter ωL of the server L to be reduced may be added as a value proportionally distributed by the remaining n server devices. In that case, the load distribution weighting parameter ωj of each of the remaining n server devices is ωj / (1−ωL) (step S160 in FIG. 4B).

  If there are a plurality of processes shared by the server L, the same load distribution weighting parameter is updated for all the processes. Then, the service processing position control unit 4 applies a new setting to the server device and the load distribution control unit 2 (step S153 in FIG. 4A).

  Next, the server selection process for newly sharing the process shown in step S102 of FIG. 3 will be described in more detail using the flowchart of FIG.

  The service processing position control unit 4 includes, for each of the non-operating server devices, a physical space position distance D, power consumption W, a processing performance index P, and an air conditioning efficiency index from the server device of interest. A is sequentially obtained (steps S120 to S123 in FIG. 5). Subsequently, the service processing position control unit 4 calculates the power-on appropriateness index Sj of each non-operating server device by the following equation (step S124 in FIG. 5).

Sj = (σd / Dj) + (σW / Wj) + σp × Pj + σa × Aj (1)
In equation (1), σd, σW, σp, and σa are weight parameters for calculating an appropriate index for power-on by integrating distance, power consumption, performance index, and air conditioning efficiency.

  Finally, the service processing position control unit 4 selects a server device whose power-on appropriate index Sj obtained in step S124 has the maximum value as a server device to be turned on (step S125 in FIG. 5). Note that step S102 shown in FIG. 3 is described when σd = 1, σW = σp = σa = 0. Thus, when the weight parameter σ is zero, calculation of each distance, power consumption, performance index, and air conditioning efficiency may be omitted.

  Next, with reference to the flowchart of FIG. 6, the server selection process for stopping the operation shown in step S <b> 151 of FIG. 4A will be described in more detail. The service processing position control unit 4 sequentially obtains the temperature T, the power consumption W, the processing performance index P, and the air conditioning efficiency index A for each of the active server devices (step S170 in FIG. 6). ~ S173). Subsequently, the service processing position control unit 4 calculates the power-off appropriate index Sj of each active server device by the following equation (step S174 in FIG. 6).

Sj = (λt / Tj) + λW × Wj + (λp / Pj) + λa / Aj (2)
However, in the equation (2), λt, λW, λp, and λa are weight parameters for calculating an appropriate index when the power is turned off by integrating temperature, power consumption, performance index, and air conditioning efficiency.

  Finally, the service processing position control unit 4 selects the server device whose power-off appropriate index Sj obtained in step S174 has the maximum value as the server device to be turned off (step S175 in FIG. 6). Note that step S151 shown in FIG. 4 has been described when λt = 1 and λW = λp = λa = 0. Thus, when the weight parameter λ is zero, calculation of each temperature, power consumption, performance index, and air conditioning efficiency may be omitted.

  Next, the effect of the server processing distribution apparatus 100A of this embodiment will be described.

  The server processing distribution apparatus 100A according to the present embodiment mainly selects the server device that is the target of the operation state (power supply) control and updates the load distribution setting by using the physical space position and temperature state of the server device. . That is, when there is an overheated server device that is higher than the appropriate temperature, the server processing distribution device 100A first adjusts load distribution and then adds the server device to distribute the processing. On the other hand, when there is a low-temperature server device that is below the appropriate temperature, the server processing distribution device 100A first stops the server device, and then adjusts load distribution to consolidate processing. Therefore, according to the present embodiment, it is possible to accurately and accurately control the operating state (number of operating units), processing position, and load distribution of server devices.

  Thus, according to the present embodiment, the server device that performs power control is determined in consideration of the temperature state, power consumption, performance index, and air conditioning efficiency, including the physical space position of the server device. In addition, since the processing is moved in consideration of the physical spatial position of the server device, it is possible to accurately determine which processing can be performed in which server device can be optimized in terms of energy.

  Further, according to the present embodiment, determination of a server device for turning on / off the power (sleep state) and setting of load distribution are performed using position information and load distribution information of the physical space position of the server device. In order to make a change, it is possible to accurately control which process and what load should be executed in which server device. As described above, according to the present embodiment, power consumption of a server system having a plurality of server devices can be suppressed.

(Second Embodiment)
Next, a second embodiment of the present invention will be described in detail with reference to the drawings. FIG. 7 shows a block diagram of a second embodiment of the server processing distribution apparatus according to the present invention. In the figure, the same components as those in FIG. The server processing distribution device 100B of this embodiment has a configuration in which a temperature / performance prediction processing unit 6 is added to the server processing distribution device 100A of the first embodiment. The service processing position control unit 4B performs slightly different operations from the service processing position control unit 4 shown in FIG. The temperature / performance prediction processing unit 6 predicts and calculates the temperature / performance of an operating server device when the operating state and load distribution setting of each of the server devices 1-1 to 1-N are updated.

  Next, the overall operation of the server processing distribution device 100B of this embodiment will be described in detail with reference to FIG. 7 and the flowcharts of FIGS.

  First, processing in the service processing position control unit 4B in FIG. 7 will be described. The service processing position control unit 4B basically performs the operation according to the flowchart of FIG. 2 in the same operation as the service processing position control unit 4 shown in FIG. However, the service processing position control unit 4B performs an operation according to the flowchart of FIG. 8 instead of the processing of step S5 shown in FIG. 2 and FIG. The operation according to the flowchart of FIG. 9 is performed instead of the process of step S9. In FIGS. 8 and 9, the same components as those in FIGS. 3A and 4A are denoted by the same reference numerals.

  First, a process of moving part or all of the processing in the overheated server device to another activated server device by the service processing position control unit 4B will be described with reference to the flowchart of FIG.

  First, the service processing position control unit 4B sets the server device in the overheated state that exceeds the appropriate temperature as the server H (step S101 in FIG. 8). Subsequently, the service processing position control unit 4B lists the non-operating servers K that can newly share the processing of the server H among the non-operating server devices (step S201 in FIG. 8).

  Subsequently, the service processing position control unit 4B performs an operation state and load distribution setting when load distribution is performed so that the server H further processes the process shared by the server H with respect to a certain server K. (Steps S104 and S105 in FIG. 8).

  Next, the service processing position control unit 4B provides the temperature / performance prediction processing unit 6 with the temperature / performance of each server device when the server K is operated in the operating state and load distribution setting obtained in step S104. Prediction calculation (step S202 in FIG. 8). The service processing position control unit 4B acquires the temperatures obtained by the prediction calculation by the temperature / performance prediction processing unit 6 for all of the server devices K in the appropriate temperature range where there is a possibility of starting and not operating. Then, it is determined whether or not there is a non-operating server device K having a predicted temperature within the appropriate temperature range (step S203 in FIG. 8).

  When the service processing position control unit 4B determines in step S203 that there is no non-operating server device having a predicted temperature within the appropriate temperature range, there is no control solution. Return to step S1. On the other hand, if the service processing position control unit 4B determines in step S203 that there are non-operating server devices with predicted temperatures within the appropriate temperature range, among the non-operating server devices with predicted temperatures within the appropriate temperature range, A server device (referred to as server K) arranged at the spatial position closest to the server H is selected (step S204 in FIG. 8). Thereafter, the service processing position control unit 4B starts by turning on the power of the server K selected in step S204 (step S103 in FIG. 8), and the server device and the load distribution control according to the operating state and load distribution setting of the selected server K. The unit 2 is controlled (step S106 in FIG. 8).

  Next, a process corresponding to step S9 in which the service processing position control unit 4B stops the server device in the low temperature state and moves the processing in the server device to another server device will be described with reference to the flowchart of FIG.

  First, the service processing position control unit 4B sets the server device having a temperature lower than the appropriate temperature as the server L (step S151 in FIG. 9). Subsequently, the service processing position control unit 4B divides the load distribution weight coefficient ωL of the server L by the remaining number n of server devices and adds the value added to the load distribution weight coefficient ωj of each server device to a new load distribution weight coefficient It calculates | requires as (omega) j (step S152 of FIG. 9).

  Next, the service processing position control unit 4B calculates the temperature / performance of each server device when the remaining n server devices are operated with the new load distribution weighting coefficient ωj obtained in step S152. The unit 6 performs prediction calculation (step S251 in FIG. 9). Subsequently, the service processing position control unit 4B checks whether all of them operate properly based on the temperature / performance prediction results of the remaining n server devices (step S252 in FIG. 9). As a result, if all of the remaining n server devices do not operate properly, there is no control solution, and the process returns to step S1 in FIG. On the other hand, when one or more of the remaining n server devices operate properly, the server L is stopped, the server device that is farthest from the proper temperature is selected, and its operating state and load distribution setting are set. Therefore, control is performed on the server device and the load distribution control unit 2 (step S153 in FIG. 9).

  Next, the effect of this embodiment will be described.

  In the present embodiment, before executing control on the server device and the load distribution control unit 2, the operation state to be updated and the temperature / performance state of the server when the load distribution setting is applied are predicted, and the setting is performed. Therefore, it is possible to perform stable control as compared with the first embodiment.

(Third embodiment)
Next, a third embodiment of the present invention will be described in detail with reference to the drawings. FIG. 10 shows a block diagram of a third embodiment of the server processing distribution apparatus according to the present invention. In the figure, the same components as those in FIG. The server processing distribution device 100C of this embodiment is configured to include a virtual machine control unit 8 instead of the load distribution control unit 2 of the server processing distribution device 100A of the first embodiment. Further, the service processing position control unit 4C performs an operation slightly different from that of the service processing position control unit 4 shown in FIG.

  When the virtual machine is used, the virtual machine operates in each of the server devices 1-1 to 1-N, and the server program operates on the virtual machine. In the load distribution control unit 2 illustrated in FIG. 1, the load of the server devices 1-1 to 1-N is adjusted by updating the load distribution weighting parameters and controlling the distribution of requests from the clients. . On the other hand, the virtual machine control unit 8 in the present embodiment adjusts the load of each server device by causing the server devices 1-1 to 1-N to simultaneously execute the virtual machines.

  Next, the operation of the server processing distribution device 100C of this embodiment that is different from the first embodiment will be described. First, since the virtual machine control unit 8 manages the power supply of the server devices 1-1 to 1-N, the operating state of the virtual machine, and the like, the service processing position control unit 4C directly controls the server device 1 -1 to 1-N are not controlled, and only the information on the operating state of the server to be set and the load distribution setting is passed to the virtual machine control unit 8.

  In addition, the load distribution setting is an arrangement of the number of virtual machines operated on each of the server devices 1-1 to 1-N, and the number of virtual machines operated on the server devices 1-1 to 1-N depends on the service. This is a value obtained by rounding up the product of the total number of virtual machines to be executed and the load distribution weighting parameter to a natural number.

  In the distribution of service requests from clients by the load distribution control unit 2 as in the first embodiment, it is assumed that the programs executed on the server devices are the same. The load distribution control can finely distribute the same service program, but is not suitable for operating different service programs. On the other hand, when the virtual machine control unit 8 is used as in the present embodiment, it is possible to adjust the load distribution by freely arranging the virtual machines without depending on the service program.

  Next, the effect of this embodiment will be described.

  In the present embodiment, a virtual machine control unit 8 is included instead of the load distribution control unit 2, and the arrangement of virtual machines that provide services to the clients is changed instead of distributing service requests from the clients. As a result, according to the present embodiment, as in the first embodiment, it is possible to accurately and appropriately control the operating state (operating number), processing position, and load distribution of the server devices 1-1 to 1-N. Is possible.

(Fourth embodiment)
Next, a fourth embodiment of the present invention will be described in detail with reference to the drawings. FIG. 11 shows a block diagram of a fourth embodiment of a server processing distribution apparatus according to the present invention. In the figure, the same components as those in FIG. The server processing distribution device 100D according to the present exemplary embodiment includes a situation extraction unit 10 and a situation / service instead of the service processing position control unit 4 and the service processing position information holding unit 5 in the server processing distribution device 100A according to the first exemplary embodiment. An arrangement correspondence analysis unit 11, a server management unit 12, a service arrangement example data holding unit 13, and a situation / operation correspondence learning unit 14 are further configured.

  The status extraction unit 10 includes service requests from clients on the server devices 1-1 to 1-N, environmental conditions such as load and temperature of each client, date and time, and operating states of the server devices 1-1 to 1-N. Then, a setting situation including the load distribution weighting parameters is extracted. The service arrangement example data holding unit 13 holds the environmental status and the setting status as a pair. A flag indicating whether or not the system has been operated satisfactorily is given to the environmental status / setting status.

  The situation / service arrangement correspondence analysis unit 11 receives the current environmental situation and the setting situation from the situation extraction unit 10, and is the same or similar to the current environmental situation / setting situation for the service arrangement example data holding unit 13. Search for past environmental and setting cases. If the search is a hit and the search result does not indicate that the system is operating satisfactorily, the situation / service arrangement correspondence analysis unit 11 receives the service request status from the same or similar past client (part of the environmental status). The setting status in which the system operates satisfactorily is extracted, and the server management unit 12 is requested to update the setting.

  If the server management unit 12 receives a setting update request from the situation / service arrangement correspondence analysis unit 11, the server management unit 12 operates the server devices 1-1 to 1-N and the load distribution control unit 2 according to the setting status. Update load distribution weighting parameters.

  The situation / operation correspondence learning unit 14 performs a good operation if there is no abnormality in temperature, load, and power consumption within a certain time range with respect to the environmental situation / setting situation held in the service arrangement example data holding unit 13. A flag is assigned, and if there is an abnormality, a flag indicating a defective operation is assigned. Further, the situation / operation correspondence learning unit 14 performs machine learning on the correspondence between the environmental situation / setting situation and the flag. As a result of the machine learning, a past environmental situation / setting situation that is the same as or similar to the current environmental situation / setting situation in the situation / service arrangement correspondence analysis unit 11 is searched, and the current environmental situation / setting situation is searched. Is used to determine whether the operation is good or bad.

  Next, the effect of this embodiment will be described.

  In the server processing distribution apparatus 100D of the present embodiment, the service processing position control unit 4 and the service processing position information holding unit 5 in the server processing distribution apparatus 100A determine the service processing position by an algorithm, whereas past cases are used. If the situation is the same as the case where the accumulated operation has caused a defective operation in the past, the setting is updated so as to be favorable. As a result, according to the server processing distribution device 100D of the present embodiment, it is possible to accurately and appropriately control the operating state (number of operating units), processing positions, and load distribution of server devices, as in the first embodiment. It is.

(Fifth embodiment)
Next, a fifth embodiment of the present invention will be described in detail with reference to the drawings. FIG. 12 shows a block diagram of a fifth embodiment of a server processing distribution apparatus according to the present invention. In the figure, the same components as those in FIG. In the description of each of the embodiments described above, it is assumed that the server devices 1-1 to 1-N have a function of transmitting the operation state of the server device called a service processor to the operation management server.

  On the other hand, the server processing distribution device 100E of the present embodiment has sensors 7-1 provided for the server devices 1′-1 to 1′-N with respect to the state (particularly temperature) corresponding to each server device individually. Measurement is performed independently by ˜7-N, and the measured temperature is transmitted to the temperature / performance information collecting unit 3.

  According to 100E of the present embodiment, basically the same operation as in the first embodiment is performed except that the detection state of the server device is different. Therefore, as in the first embodiment, the operation state of the server device ( The number of operating units), processing position, and load distribution can be appropriately controlled with high accuracy.

  Next, the operation of the server processing distribution apparatus of the present invention will be described using a specific embodiment.

  As shown in FIG. 13A, the server processing distribution apparatus according to the present embodiment is installed in server equipment housed in each of four racks A, B, C, and D arranged in the server room 200. On the other hand, load distribution processing based on those spatial positions is performed. FIG. 13B shows the rack names and the position coordinates of the racks A to D in the server room 200.

  In addition, as shown in FIG. 14 (A), a total of eight server devices stored in each of the racks A to D are assigned server names, and two are stored in each rack. The temperature and business name of each server device are shown. Here, the relationship between the business name and the load is shown in FIG. In other words, in the server processing distribution apparatus of this embodiment, each of the eight server devices is arranged in which rack, temperature, business running on this server (server program, server type (role It is assumed that something is managed as shown in FIG.

  Here, the business name “Web” is a web server program such as Apache or Microsoft IIS, for example, the business name “DB” is a database server program such as Oracle or PostgreSQL, and the business name “AP” is For example, an application server program such as WebLogic.

  It is assumed that the load of each business is updated as needed by service monitoring. Here, the appropriate temperature range is 30 ° C. to 40 ° C.

  FIG. 15 shows an image of the temperature distribution in the server room shown in FIG. The higher the temperature, the darker the color, and the higher the temperature around rack A and rack B in which two server devices are operating.

  Here, it is assumed that the load of a part of business increases rapidly. Then, it is assumed that the temperature of the server devices COM1 and COM2 in the rack A executing the business Web has increased to 50 ° C. FIG. 16 shows the temperature distribution in the server room in this case, and the temperature around rack A and rack B is higher than in the case of FIG.

  In this case, the server processing distribution apparatus according to this embodiment operates as follows. Since the overheating state cannot be solved by changing the load distribution of the server devices COM1 and COM2, a server for executing the business Web is added (step S5 in FIG. 2). The server devices that are not executing the business Web are server devices COM3 to COM8 as shown in FIG. Further, the barycentric coordinates of the spatial position of the rack A of the operating server devices COM1 and COM2 are (1.67, 1.5), and the server devices COM3 and COM4 of the rack B are close to the physical distance.

  Therefore, the server processing distribution apparatus of this embodiment shares the business Web processing with the server device COM3. As a result, the loads on the server devices COM1 and COM2 that are in an overheated state are reduced, and the temperature of these server devices approaches the appropriate range. On the other hand, in the server device COM3, since the DB business is also executed, the load distribution is made lower than that of the server devices COM1 and COM2. If the temperature of each of the server devices COM1, COM2, and COM3 still exceeds the appropriate range even after waiting for a certain time in step S1 of FIG. 2, the other server device COM4 in the rack B also performs business. Share Web processing.

  As described above, the server processing distribution apparatus according to the present embodiment sequentially increases the server devices that execute the business Web while adjusting the load distribution. FIGS. 17A and 17B show changes in temperature, business, and load management due to the load distribution adjustment described above. FIG. 18 shows the temperature distribution in the server room after the setting change of business or the like as shown in FIGS. 17 (A) and 17 (B). As shown in FIG. 18, the temperature distribution in the server room is substantially uniform in the server room.

  Next, it is assumed that the access to the business Web is reduced and the temperature of each server device COM1, COM2, COM5 to COM8 is lowered to 25 ° C. FIG. 19 shows the temperature distribution in the server room in this case.

  In this case, the server processing distribution apparatus of this embodiment operates as follows. The center-of-gravity coordinates of the rack positions of all the active servers are (2, 1.5), and any server device is equidistant in terms of distance, so any server device may be used, but first the server device COM1 is stopped. (Step S9 in FIG. 2). By stopping the server device COM1, the load distribution to other server devices is increased and adjusted. If the temperatures of the server devices COM2, COM5 to COM8 are still below the appropriate range, the server devices are stopped in the order of COM2, COM5, COM6, and COM8. When the server device COM7 is stopped, the server device COM7 is excluded from being stopped because it is the server device that provides the only business AP on the system.

  As described above, the server processing distribution apparatus according to the present embodiment sequentially increases the server devices in the low temperature state to be stopped while adjusting the load distribution. 20A and 20B show changes in the management of temperature, work, and load due to the load distribution adjustment described above. In FIG. 20A, “sleep” means stop. FIG. 21 shows the temperature distribution in the server room after the setting is changed as shown in FIGS.

  In addition, this invention is not limited to the above embodiment and Example, For example, you may make it combine 5th Embodiment with either of 1st-4th Embodiment. A configuration in which the load distribution control unit 2 in the server distributed processing device 100D of the fifth embodiment is replaced with the virtual machine control unit 8 shown in FIG. 10 is also possible. According to the server distributed processing apparatus having this configuration, the server management unit 12 derives the operating state and arrangement setting of a virtual machine that has operated properly in the past from the current temperature, performance information, and processing request of the server device. .

  Furthermore, the present invention also includes server distributed processing programs that cause the server processing distribution apparatuses 100A to 100E of the first to fifth embodiments and the combination of the above embodiments to be executed by a computer.

  According to the present invention, in a server system composed of a plurality of server devices, an operation management application for determining which process is executed in which server can be optimized in terms of energy and controlling the server system Applicable to. Instead of energy evaluation, evaluation including system performance and reliability may be used.

  The present invention can also be applied to production management in a production system using each server device as a manufacturing machine.

1-1 to 1-N, 1′-1 to 1′-N Server device 2 Load distribution control unit 3 Temperature / performance information collection unit 4, 4B, 4C, 4D Service processing position control unit 5 Service processing position information holding unit 7-1 to 7-N Sensor 6 Temperature / performance prediction processing unit 8 Virtual machine control unit 10 Situation extraction unit 11 Situation / service arrangement correspondence analysis unit 12 Server management unit 13 Service / arrangement example data holding unit 14 Situation / operation correspondence Relation learning unit 100A, 100B, 100C, 100D, 100E Server processing distribution device 200 Server room

Claims (33)

A plurality of computer devices each operating under program control, and responding to requests from clients;
A distribution unit that distributes the request from the client;
Collecting means for collecting temperature and performance information of each of the plurality of server devices;
Service processing position information holding means for holding operation information, physical space position information and load distribution setting of each of the plurality of server devices together with the collected temperature and performance information;
Calculation means for calculating a new operation state and load distribution setting from the operation information, physical space position information and load distribution setting held by the service processing position information holding means, and the temperature and performance information;
Control means for controlling the plurality of server devices based on the new operating state and load distribution setting;
A server processing distribution apparatus comprising: A plurality of computer devices each operating under program control, and responding to requests from clients;
A distribution unit that distributes the request from the client;
Collecting means for collecting temperature and performance information of each of the plurality of server devices;
Service processing position information holding means for holding operation information, physical space position information and load distribution setting of each of the plurality of server devices together with the collected temperature and performance information;
A temperature / performance prediction means for predicting and calculating the temperature / performance of an operating server device when the operating state and load distribution setting of each of the plurality of server devices are updated;
Calculation means for calculating a new operation state and load distribution setting from the operation information, physical space position information and load distribution setting and temperature and performance information held by the service processing position information holding means;
A determination unit that determines whether or not the new operating state and load distribution setting calculated by the calculation unit are appropriate from the prediction result of the temperature / performance information predicted by the temperature / performance prediction unit;
Control means for controlling the plurality of server devices based on the new operating state and load distribution setting determined to be more appropriate than the determination means;
A server processing distribution apparatus comprising: A plurality of server devices each of which operates a virtual machine and a server program operates on the virtual machine and responds to a request from a client;
Collecting means for collecting temperature and performance information of each of the plurality of server devices;
Service processing position information holding means for holding operation information, physical space position information and load distribution setting of each of the plurality of server devices together with the collected temperature and performance information;
Calculation means for calculating a new operation state and load distribution setting from the operation information, physical space position information and load distribution setting held by the service processing position information holding means, and the temperature and performance information;
Virtual machine control means for controlling the number of virtual machines to be executed simultaneously by the plurality of server devices based on the new operating state and load distribution setting;
A server processing distribution apparatus comprising: A plurality of computer devices each operating under program control, and responding to requests from clients;
A distribution unit that distributes the request from the client;
Collecting means for collecting temperature and performance information of each of the plurality of server devices;
Extraction means for extracting the environmental status including the temperature, performance information and processing request from the client from the information collected by the collecting means, and the load distribution setting status of each of the server equipment;
Holding means for holding a correspondence relationship between the environmental situation extracted by the extraction means and the load distribution setting situation;
Whether the current environmental situation extracted by the extracting means and the current environmental situation that is the same or similar to the current load allocation setting situation and the load distribution setting situation that has been properly operated in the past are held by the holding means. A search means for searching;
Control means for controlling the plurality of server devices according to the environmental status held in the hitting holding means and the load distribution setting status that has been properly operated in the past, when the search by the search means hits; ,
A server processing distribution apparatus comprising: A plurality of server devices each of which operates a virtual machine and a server program operates on the virtual machine and responds to a request from a client;
Collecting means for collecting temperature and performance information of each of the plurality of server devices;
Extraction means for extracting the environmental status including the temperature, performance information and processing request from the client from the information collected by the collecting means, and the load distribution setting status of each of the server equipment;
Holding means for holding a correspondence relationship between the environmental situation extracted by the extraction means and the load distribution setting situation;
Whether the current environmental situation extracted by the extracting means and the current environmental situation that is the same or similar to the current load allocation setting situation and the load distribution setting situation that has been properly operated in the past are held by the holding means. A search means for searching;
When the search by the search means is hit, the server is simultaneously executed by the plurality of server devices based on the environmental status held in the hit holding means and the load distribution setting status that has been properly operated in the past. Control means for controlling the number of the virtual machines;
A server processing distribution apparatus comprising: It is determined whether or not there is an abnormality in the temperature, load and power consumption of the server device within a certain time range with respect to the environmental status and load distribution setting status held in the holding means, and the determination A learning unit that gives a flag indicating a result, and further machine-learns a correspondence relationship between the environmental status held in the holding unit and the load distribution setting status;
As a result of the machine learning by the learning means, the current environmental situation in the search means and the same or similar environmental situation as the current load allocation setting situation and the load distribution setting situation that has been properly operated in the past are obtained by the holding means. Used to determine whether the current environmental status and the current load distribution setting status are in good operation when searching for
When the machine learning result indicates that the load distribution setting state held by the holding unit is not a good operation, the control unit updates the load distribution setting state to a load distribution setting state that performs a good operation. 6. The server processing distribution device according to claim 4 or 5, wherein:   7. The apparatus according to claim 1, further comprising a measuring unit that individually measures the temperature of each of the plurality of server devices and outputs the measured temperature to the collecting unit. The server processing distribution device according to one item. The calculating means includes
When a server device having a temperature higher than a preset appropriate temperature is detected among the plurality of server devices, the physical space from the high-temperature server device among the non-operating server devices at or below the proper temperature 8. The load distribution weighting factor is calculated for sharing the processing of the high-temperature server device with a non-operating server device in the closest position. 8. Server processing distribution device. The calculating means includes
When a server device having a temperature higher than a preset appropriate temperature is detected among the plurality of server devices, the power consumption, processing performance, and air conditioning efficiency index of each of the non-operating server devices in the proper temperature range And calculating an appropriate power-on indicator using one or more of the four parameters consisting of the distance from the high-temperature server device,
The control means includes
8. The server device according to claim 1, wherein the non-operating server device is controlled as a server device to be operated that shares processing of the high-temperature server device based on the power-on appropriate index. The server processing distribution device described. The calculating means includes
Among the plurality of server devices, when a low-temperature server device that is in a state of lower temperature than a preset appropriate temperature and performs processing in a shared manner with other server devices is detected, the low-temperature server 10. The load distribution weighting coefficient for calculating the load distribution weighting factor for stopping the processing of the equipment and for sharing the processing of the low-temperature server equipment that has stopped the processing with the remaining operating server equipment. The server processing distribution device according to claim 1. The calculating means includes
Of the plurality of server devices, when a plurality of low-temperature server devices that are in a state of lower temperature than the preset appropriate temperature and that perform processing in a shared manner with other server devices are detected, Calculate an appropriate power-off indicator using one or more of the four parameters consisting of the temperature, power consumption, processing performance, and air conditioning efficiency index of each low-temperature server device,
The control means includes
Based on the appropriate power-off indicator, a server device to be stopped is selected from the plurality of low-temperature server devices, and the remaining server devices that are in operation are assigned to the processing of the low-temperature server device to be stopped. The server processing distribution device according to claim 1, wherein: A distribution step of distributing a request from the client to a plurality of server devices each responding to a request from the client, each of the computers operating by program control;
A collection step of collecting temperature and performance information of each of the plurality of server devices;
A holding step of holding operation information, physical space position information, and load distribution settings of each of the plurality of server devices in a holding unit together with the collected temperature and performance information,
A calculation step for calculating a new operation state and load distribution setting from the operation information, physical space position information and load distribution setting held by the holding means, and the temperature and performance information;
A control step of controlling the plurality of server devices based on the new operating state and load distribution setting;
A server processing distribution method comprising: A distribution step of distributing a request from the client to a plurality of server devices each responding to a request from the client, each of the computers operating by program control;
A collection step of collecting temperature and performance information of each of the plurality of server devices;
A holding step of holding operation information, physical space position information, and load distribution settings of each of the plurality of server devices in a holding unit together with the collected temperature and performance information,
A temperature / performance prediction step for predicting and calculating a temperature / performance of an operating server device when the operating state and load distribution setting of each of the plurality of server devices are updated;
A calculation step of calculating a new operating state and load distribution setting from the operation information, physical space position information and load distribution setting and temperature and performance information held by the holding unit;
A determination step for determining whether or not the new operating state and load distribution setting calculated by the calculation step are appropriate from the prediction result of the temperature / performance information predicted by the temperature / performance prediction step;
A control step of controlling the plurality of server devices based on the new operating state and load distribution setting determined to be appropriate by the determination step;
A server processing distribution method comprising: A collection step for collecting temperature and performance information of each of a plurality of server devices, each of which is a computer on which each virtual machine operates and a server program operates on the virtual machine, and responds to a request from a client;
A holding step of holding operation information, physical space position information, and load distribution settings of each of the plurality of server devices in a holding unit together with the collected temperature and performance information,
A calculation step for calculating a new operation state and load distribution setting from the operation information, physical space position information and load distribution setting held by the holding means, and the temperature and performance information;
A virtual machine control step for controlling the number of virtual machines to be executed simultaneously by the plurality of server devices based on the new operating state and load distribution setting;
A server processing distribution method comprising: A distribution step of distributing a request from the client to a plurality of server devices each responding to a request from the client, each of the computers operating by program control;
A collection step of collecting temperature and performance information of each of the plurality of server devices;
An extraction step for extracting the temperature of the server device, the performance information and the processing status from the client from the information collected by the collecting step, and the load distribution setting status of each of the server devices;
A holding step of holding a correspondence relationship between the environmental situation extracted by the extraction step and the load distribution setting situation in a holding unit;
Whether the current environmental status extracted by the extraction step is the same as or similar to the current load allocation setting status, and whether the load allocation setting status that has been properly operated in the past is held by the holding means. A search means for searching;
A control step for controlling the plurality of server devices according to the environmental status held in the hitting holding means and the load distribution setting status that has been properly operated in the past when the search by the search step is hit; ,
A server processing distribution method comprising: A collection step for collecting temperature and performance information of each of a plurality of server devices, each of which is a computer on which each virtual machine operates and a server program operates on the virtual machine, and responds to a request from a client;
An extraction step for extracting the temperature of the server device, the performance information and the processing status from the client from the information collected by the collecting step, and the load distribution setting status of each of the server devices;
A holding step of holding a correspondence relationship between the environmental status extracted by the extracting unit and the load distribution setting status in a holding unit;
Whether the current environmental status extracted by the extraction step is the same as or similar to the current load allocation setting status, and whether the load allocation setting status that has been properly operated in the past is held by the holding means. A search step to search;
When the search in the search step is hit, the server is simultaneously executed by the plurality of server devices based on the environmental status held in the hit holding means and the load distribution setting status that has been properly operated in the past. A control step for controlling the number of the virtual machines;
A server processing distribution method comprising: It is determined whether or not there is an abnormality in the temperature, load and power consumption of the server device within a certain time range with respect to the environmental status and load distribution setting status held in the holding means, and the determination A learning step of providing a flag indicating a result and machine learning a correspondence relationship between the environmental situation held in the holding unit and the load distribution setting situation;
The result of machine learning by the learning step is that the current environmental status in the search step and the current environmental status of the load allocation setting are the same or similar to each other, and the load allocation setting status that has been properly operated in the past is determined by the holding means. Used to determine whether the current environmental status and the current load distribution setting status are in good operation when searching for
When the machine learning result indicates that the load distribution setting state held by the holding unit is not a good operation, the control step updates the load distribution setting state to a load distribution setting state that performs a good operation. The server processing distribution method according to claim 15 or 16, characterized in that:   18. The method according to any one of claims 12 to 17, further comprising a measuring step of measuring the temperature of the server device individually for each of the plurality of server devices and reflecting the measured temperature in the collecting step. The server processing distribution method according to claim 1. The calculation step includes:
When a server device having a temperature higher than a preset appropriate temperature is detected among the plurality of server devices, the physical space from the high-temperature server device among the non-operating server devices at or below the proper temperature 19. The load distribution weighting factor is calculated for sharing the processing of the high-temperature server device with a non-operating server device in the closest position. 19. Server processing distribution method. The calculation step includes:
When a server device having a temperature higher than a preset appropriate temperature is detected among the plurality of server devices, the power consumption, processing performance, and air conditioning efficiency index of each of the non-operating server devices in the proper temperature range And calculating an appropriate power-on indicator using one or more of the four parameters consisting of the distance from the high-temperature server device,
The control step includes
The non-operating server device is controlled as a server device that should be operated to share the processing of the high-temperature server device based on the power-on appropriate index. The server processing distribution method described. The calculation step includes:
Among the plurality of server devices, when a low-temperature server device that is in a state of lower temperature than a preset appropriate temperature and performs processing in a shared manner with other server devices is detected, the low-temperature server 21. The load distribution weighting factor for calculating the load distribution weighting coefficient for stopping the processing of the device and for sharing the processing of the low-temperature server device that has stopped the processing with the remaining operating server devices. The server processing distribution method according to any one of the above. The calculation step includes:
Of the plurality of server devices, when a plurality of low-temperature server devices that are in a state of temperature lower than a preset appropriate temperature and that perform processing in a shared manner with other server devices are detected, Calculate an appropriate power-off indicator using one or more of the four parameters consisting of the temperature, power consumption, processing performance, and air conditioning efficiency index of each low-temperature server device,
The control step includes
Based on the appropriate power-off indicator, a server device to be stopped is selected from the plurality of low-temperature server devices, and the remaining server devices that are in operation are assigned to the processing of the low-temperature server device to be stopped. 21. The server processing distribution method according to any one of claims 12 to 20, wherein: A distribution step of distributing a request from the client to a plurality of server devices each responding to a request from the client, each of the computers operating by program control;
A collection step of collecting temperature and performance information of each of the plurality of server devices;
A holding step of holding operation information, physical space position information, and load distribution setting of each of the plurality of server devices together with the collected temperature and performance information in a holding unit;
A calculation step for calculating a new operation state and load distribution setting from the operation information, physical space position information and load distribution setting held by the holding means, and the temperature and performance information;
A server processing distribution program that causes a control step for controlling the plurality of server devices to be executed by a computer different from the server devices based on the new operating state and load distribution setting. A distribution step of distributing a request from the client to a plurality of server devices each responding to a request from the client, each of the computers operating by program control;
A collection step of collecting temperature and performance information of each of the plurality of server devices;
A holding step of holding operation information, physical space position information, and load distribution settings of each of the plurality of server devices in a holding unit together with the collected temperature and performance information,
A temperature / performance prediction step for predicting and calculating the temperature / performance of an operating server device when the operating state and load distribution setting of each of the plurality of server devices are updated;
A calculation step of calculating a new operating state and load distribution setting from the operation information, physical space position information and load distribution setting and temperature and performance information held by the holding unit;
A determination step for determining whether or not the new operating state and load distribution setting calculated by the calculation step are appropriate from the prediction result of the temperature / performance information predicted by the temperature / performance prediction step;
And a control step of controlling the plurality of server devices based on the new operating state and load distribution setting determined to be appropriate by the determination step, by a computer different from the server device. Server processing distributed program. A collection step of collecting temperature and performance information of each of a plurality of server devices, each of which is a computer on which each virtual machine operates and a server program operates on the virtual machine, and responds to a request from a client;
A holding step of holding operation information, physical space position information, and load distribution settings of each of the plurality of server devices in a holding unit together with the collected temperature and performance information,
A calculation step for calculating a new operation state and load distribution setting from the operation information, physical space position information and load distribution setting held by the holding means, and the temperature and performance information;
A virtual machine control step for controlling the number of virtual machines to be executed simultaneously by the plurality of server devices based on the new operating state and load distribution setting is executed by a computer different from the server devices. Server processing distributed program. A distribution step of distributing a request from the client to a plurality of server devices each responding to a request from the client, each of the computers operating by program control;
A collection step of collecting temperature and performance information of each of the plurality of server devices;
An extraction step for extracting the temperature of the server device, the performance information and the processing status from the client from the information collected by the collecting step, and the load distribution setting status of each of the server devices;
A holding step of holding a correspondence relationship between the environmental situation extracted by the extraction step and the load distribution setting situation in a holding unit;
Whether the current environmental status extracted by the extraction step is the same as or similar to the current load allocation setting status, and whether the load allocation setting status that has been properly operated in the past is held by the holding means. A search means for searching;
A control step for controlling the plurality of server devices according to the environmental status held in the hitting holding means and the load distribution setting status that has been properly operated in the past when the search by the search step is hit; Is executed by a computer different from the server device. A collection step of collecting temperature and performance information of each of a plurality of server devices, each of which is a computer on which each virtual machine operates and a server program operates on the virtual machine, and responds to a request from a client;
An extraction step for extracting the temperature of the server device, the performance information and the processing status from the client from the information collected by the collecting step, and the load distribution setting status of each of the server devices;
A holding step of holding a correspondence relationship between the environmental status extracted by the extracting unit and the load distribution setting status in a holding unit;
Whether the current environmental status extracted by the extraction step is the same as or similar to the current load allocation setting status, and whether the load allocation setting status that has been properly operated in the past is held by the holding means. A search step to search;
When the search in the search step is hit, the server is simultaneously executed by the plurality of server devices based on the environmental status held in the hit holding means and the load distribution setting status that has been properly operated in the past. A server processing distribution program that causes a control step for controlling the number of virtual machines to be executed by a computer different from the server device. It is determined whether or not there is an abnormality in the temperature, load and power consumption of the server device within a certain time range with respect to the environmental status and load distribution setting status held in the holding means, and the determination A learning step of providing a flag indicating a result and machine learning a correspondence relationship between the environmental situation held in the holding unit and the load distribution setting situation;
The result of machine learning by the learning step is that the current environmental status in the search step and the current environmental status of the load allocation setting are the same or similar to each other, and the load allocation setting status that has been properly operated in the past is determined by the holding means. Used to determine whether the current environmental status and the current load distribution setting status are in good operation when searching for
In the control step, when the machine learning result indicates that the load distribution setting status held by the holding means is not a good operation, the load distribution setting status is updated to a load distribution setting status that performs a good operation. 28. The server processing distribution program according to claim 26 or 27.   29. The method according to any one of claims 23 to 28, further comprising a measuring step of measuring the temperature of the server device individually for each of the plurality of server devices and reflecting the measured temperature in the collecting step. A server processing distribution program according to any one of the preceding claims. The calculation step includes:
When a server device having a temperature higher than a preset appropriate temperature is detected among the plurality of server devices, the physical space from the high-temperature server device among the non-operating server devices at or below the proper temperature 30. A load distribution weighting factor is calculated, which causes a non-operating server device in a closest position to share the processing of the high-temperature server device. Server processing distributed program. The calculation step includes:
When a server device having a temperature higher than a preset appropriate temperature is detected among the plurality of server devices, the power consumption, processing performance, and air conditioning efficiency index of each of the non-operating server devices in the proper temperature range And calculating an appropriate power-on indicator using one or more of the four parameters consisting of the distance from the high-temperature server device,
The control step includes
30. The server device according to any one of claims 23 to 29, wherein the non-operating server device is controlled as a server device to be operated to share the processing of the high-temperature server device based on the power-on appropriate index. The server processing distribution program described. The calculation step includes:
Among the plurality of server devices, when a low-temperature server device that is in a state of lower temperature than a preset appropriate temperature and performs processing in a shared manner with other server devices is detected, the low-temperature server 32. The load distribution weighting coefficient is calculated, wherein the processing of the device is stopped, and the remaining server devices that are in operation share the processing of the low-temperature server device that has stopped processing. The server processing distribution program according to any one of the above. The calculation step includes:
Of the plurality of server devices, when a plurality of low-temperature server devices that are in a state of temperature lower than a preset appropriate temperature and that perform processing in a shared manner with other server devices are detected, Calculate an appropriate power-off indicator using one or more of the four parameters consisting of the temperature, power consumption, processing performance, and air conditioning efficiency index of each low-temperature server device,
The control step includes
Based on the appropriate power-off indicator, a server device to be stopped is selected from the plurality of low-temperature server devices, and the remaining server devices that are in operation are assigned to the processing of the low-temperature server device to be stopped. 32. The server processing distribution program according to any one of claims 23 to 31.