JP2009087213A - Computer reserve capacity calculating device and computer reserve capacity calculating method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide useful information when a job is allocated to any of a plurality of computers. <P>SOLUTION: The computer reserve capacity calculating device 1 is equipped with a memory device 2 and a processing device 3, wherein the memory device 2 stores the performance value information DA to indicate the performance values of different computers, the use rate information DB to indicate the resource use rate of each computer, and the threshold information DR to indicate the threshold of the resource use rate. The processing device 3 makes reference to these pieces of information DA, DB, and DR, and calculates the reserve capacity of each computer on the basis of the difference between the threshold and resource use rate and the performance value. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a technique for grasping the usage status of a computer system composed of a plurality of computers. In particular, the present invention relates to a technique for providing information useful in assigning a job to any of a plurality of computers.

  There is known a computer system in which a plurality of computers are connected via a network. In such a computer system, it is important to manage the usage status and load of the system. The following are known as techniques related to management of computer systems.

  Japanese Patent Laying-Open No. 2006-92053 (Patent Document 1) discloses a system usage rate management apparatus that calculates the resource usage rate of an entire computation node group composed of a plurality of computation nodes. According to the system usage rate management apparatus, the node maximum performance setting means sets the maximum performance of each of the plurality of calculation nodes. Here, the maximum performance means the number of requests that a node can process per unit time. Further, the resource usage rate acquisition unit acquires the resource usage rate of each of the plurality of calculation nodes. Further, the virtual usage rate calculation means calculates the resource usage rate of the calculation node group as a whole based on the maximum performance and resource usage rate of each calculation node. Thus, by calculating the resource usage rate of the entire computing node group, it is possible to quickly grasp the usage status and load of the system.

  Further, in a computer system in which a plurality of computers are connected via a network, it is important to distribute the load on each computer. The following are known as techniques related to such load balancing.

  Japanese Patent Laid-Open No. 10-161987 (Patent Document 2) describes a load distribution processing system. According to the load distribution processing system, the load status acquisition means provided in each computer acquires the load factor of its own computer. Here, the load factor is the total of the CPU usage rate by the job being executed. One computer is provided with load adjusting means. When a new job occurs in any one of the computers, the load adjustment unit detects the computer with the lowest load factor and causes the detected computer to execute the new job. In this way, an appropriate load state is realized by sequentially assigning jobs that occur one after another to the computer having the lowest load factor.

  Japanese Patent Application Laid-Open No. 11-24949 (Patent Document 3) describes a multi-server / client system including a plurality of calculation servers and one calculation client. When a calculation client requests a new calculation, the calculation client selects an assignment destination of the calculation from a plurality of calculation servers according to the following method. That is, the calculation client refers to the number of operations that each calculation server can process per unit time and the number of operations that each calculation server is currently executing. Further, the calculation client calculates the time until the operation currently being executed in each calculation server is completed by dividing the “number of operations currently being executed” by the “number of operations that can be processed per unit time”. Then, the calculation client selects the calculation server with the shortest calculated time as the assignment destination of the requested calculation. According to such a method, the arithmetic processing is assigned in a direction in which the calculation load of the calculation server becomes uniform as a whole.

  Japanese Patent Laying-Open No. 2003-46539 (Patent Document 4) describes a router load distribution method. The router includes a measurement unit that measures its own CPU usage rate and a setting unit that dynamically sets the priority. The setting means sets the priority such that the higher the CPU usage rate, the lower the priority. According to VRRP (Virtual Router Redundancy Protocol), a router with high priority is assigned to specific packet processing. Therefore, a router with a low CPU usage rate processes more packets, and the processing load is distributed among the routers.

JP 2006-92053 A Japanese Patent Laid-Open No. 10-161987 Japanese Patent Laid-Open No. 11-24949 JP 2003-46539 A

  When assigning a job to any of a plurality of computers, it is conceivable to select a computer having the lowest CPU usage rate as an assignment destination, for example, as in the related art described above. However, when computers with different performances are included in the system, simply assigning a job to a computer with the lowest CPU usage rate does not necessarily reduce the processing time of the job. This is because a high-performance computer may be able to process assigned jobs faster than other computers even if the CPU usage rate is relatively high. Therefore, it is not preferable to determine the job assignment destination based only on the CPU usage rate.

  One object of the present invention is to provide information useful in determining a job assignment destination.

  According to the present invention, information useful for assigning a job to any of a plurality of computers is provided. For this purpose, the “remaining power” of each computer is calculated. Here, the inventor of the present application paid attention to the following points. That is, when the resource usage rate of a certain computer exceeds a predetermined threshold, the effective processing performance of that computer is significantly deteriorated. Therefore, in calculating the remaining capacity of each computer, the threshold value of the resource usage rate is taken into consideration.

  In a first aspect of the present invention, a computer remaining power calculation device is provided. The computer remaining capacity calculation device includes a storage device and a processing device. The storage device stores performance value information indicating the performance value of each of a plurality of computers, usage rate information indicating the resource usage rate of each computer, and threshold information indicating a threshold value of the resource usage rate. The processing device refers to the performance value information, the usage rate information, and the threshold information, and calculates the remaining capacity of each computer based on the difference between the threshold and the resource usage rate and the performance value.

  In a second aspect of the present invention, a computer remaining power calculation method is provided. The computer capacity calculation method includes (A) obtaining the performance value of each of a plurality of computers, (B) obtaining the resource usage rate of each computer, and (C) obtaining the threshold value of the resource usage rate. And (D) calculating the remaining capacity of each computer based on the difference between the threshold value and the resource usage rate and the performance value.

  In a third aspect of the present invention, a computer remaining capacity calculation program is provided. The computer capacity calculation program (a) obtains the performance value of each of a plurality of computers, (b) obtains the resource usage rate of each computer, and (c) obtains the threshold value of the resource usage rate. And (d) calculating the remaining capacity of each computer based on the difference between the threshold value and the resource usage rate and the performance value.

  When determining an assignment destination of a certain job, it is preferable to refer to the remaining capacity of each computer calculated by the present invention. In calculating the surplus power, not only the resource usage rate such as the CPU usage rate but also the performance value is considered. Therefore, even if the performance of a plurality of computers is different from each other, the calculated remaining capacity can be a useful guideline in determining a job assignment destination.

  Furthermore, according to the present invention, in calculating the remaining capacity, in addition to the resource usage rate and the performance value of each computer, the threshold value of the resource usage rate is also taken into consideration. More specifically, the difference between the threshold value and the resource usage rate is considered. By referring to the surplus power thus calculated, it is possible to determine the job assignment destination so that the resource usage rate does not exceed the threshold. As a result, it is possible to prevent the resource usage rate of the computer to which the job has been assigned from exceeding a threshold and degrading the effective processing performance of the computer.

  As described above, according to the present invention, information useful for determining a job assignment destination is provided.

1. Computer System First, consider a computer system in which a plurality of computers are connected via a network. Examples of the computer include a server, a workstation, and a calculation node. For example, FIG. 1 schematically shows a server system in which a plurality of servers are connected via a network. This server system includes N servers M _{1 to} M _N (N is an integer of 2 or more). ID numbers 1 to N are assigned to the servers M _{1 to MN} , respectively.

  In FIG. 1, each server may be a physical computer or a virtual computer (VM) realized by a virtualization technique. The former is sometimes called a “physical server”, and the latter is sometimes called a “virtual server”. Virtual servers do not necessarily correspond one-to-one with physical servers. With the virtualization technology, a plurality of virtual servers can be operated on one physical server.

The servers M _{1 to MN} may include two or more servers having different performances. That is, heterogeneous servers may be mixed in the server system (heterogeneous environment).

Consider a case where a certain job is assigned to _{one of} a plurality of servers M _{1 to} M _N. The job here is not limited to a process as a processing unit on the OS, but may be a request to a transaction system such as a Web server. The job may be a virtual machine such as a virtual server. This is because a physical server can be regarded as a job to be processed by a physical server.

  When assigning jobs, it is desirable to determine the most appropriate assignment destination in consideration of the usage status of the server system. According to the embodiment of the present invention, useful information reflecting the usage status of the server system is provided. More specifically, the “remaining power” of each server is calculated as the useful information. An apparatus for calculating the surplus capacity of each server (each computer) is a “computer surplus capacity calculating apparatus” described next.

2. FIG. 2 is a block diagram showing a configuration of a computer surplus power calculation apparatus 1 according to the embodiment of the present invention. The computer remaining capacity calculation device 1 is a computer including a storage device 2 and a processing device 3. Examples of the storage device 2 include a RAM (Random Access Memory) and a HDD (Hard Disk Drive). The processing device 3 has a CPU (Central Processing Unit).

  The storage device 2 stores a performance value table (performance value information) DA, a usage rate table (usage rate information) DB, threshold data (threshold information) DR, a surplus power table (residual power information) DC, and the like.

The performance value table DA indicates the “performance value” of each server. The performance value is an integer or a real number representing a performance ratio related to a certain computer resource. Here, the computer resource indicates a disk, a memory, a network and the like in addition to the CPU. For example, “SPECint” indicating the integer arithmetic processing performance of the server CPU and “SPECfp” indicating the real arithmetic processing performance obtained by the benchmark test are publicly available. These “SPECint” and “SPECfp” can be used as the performance value of the CPU of each server. In addition, the performance value can be calculated based on the measurement result using the benchmark for a specific application. Larger performance values represent higher performance. For example, if the performance value of a computer resource server _M 1 relates, _{M 2} is respectively 2.0,5.0, server _{M 2} performance is 2.5 times the server _{M 1.}

  The usage rate table DB indicates the “resource usage rate” of each server. The resource usage rate is a ratio of computer resources used by the server per unit time. As described above, the computer resource refers to a disk, a memory, a network, and the like in addition to the CPU. Therefore, the resource usage rate includes CPU usage rate, disk usage rate, memory usage, network transfer rate, and the like. For example, when the CPU performs processing for 30% of the unit time and the remaining 70% is idle, the CPU usage rate is 0.3 (30%). The resource usage rate can be acquired by using, for example, a function built in advance in the OS operating on the server or a module uniquely built in the OS.

  The threshold data DR indicates the “threshold” of the resource usage rate described above. For example, the threshold value of the resource usage rate is set to 0.8 (80%). It is generally known that when the resource usage rate such as the CPU usage rate exceeds this threshold, the effective performance of the application operating on the server is significantly deteriorated.

  The surplus power table DC is a table indicating the “residual power” of each server, and is created by the computer surplus power calculating apparatus 1 according to the present embodiment. As will be described later, the remaining capacity of each server is calculated based on the above-described performance value, resource usage rate, and threshold value.

  Referring to FIG. 2 again, the processing device 3 includes a data input unit 10 and a remaining power calculation unit 20. The data input unit 10 includes a performance value acquisition unit 11, a usage rate acquisition unit 12, and a threshold acquisition unit 13. The data input unit 10, the performance value acquisition unit 11, the usage rate acquisition unit 12, the threshold acquisition unit 13, and the surplus power calculation unit 20 are modules that provide specific functions, respectively, and the CPU of the processing device 3 executes the program PROG. It is realized by doing. The program PROG is typically recorded on a computer-readable recording medium. The program PROG causes the computer remaining power calculation device 1 to execute the following processing.

  FIG. 3 is a flowchart showing processing by the computer remaining power calculation device 1. With reference to FIG. 1 to FIG. 3, a computer remaining power calculation method according to the present embodiment will be described.

Step S10:
The data input unit 10 receives various types of information from the outside, and creates the above-described performance value table DA, usage rate table DB, and threshold data DR stored in the storage device 2.

Specifically, the performance value acquisition unit 11 acquires the performance value A _i of each server M _i (i = 1 to N). For example, the user inputs the performance value A _i by using an input device (not shown), and the performance value acquisition unit 11 receives the input performance value A _i . Then, the performance value acquisition unit 11 records the received performance value A _i in the performance value table DA. In this way, the performance value table DA indicating the correspondence relationship between the server M _i and the performance value A _i is created (step S11).

Further, the usage rate acquisition unit 12 acquires the resource usage rate B _i of each server M _i at a certain point in time. Resource Utilization B _i is obtained by, for example, pre-installed functionality OS that operates on a server M _i. Each server M _i transmits the obtained resource usage rate B _i to the computer remaining capacity calculation device 1, and the usage rate acquisition unit 12 receives the resource usage rate B _i from each server M _i . Utilization acquisition unit 12, a resource utilization ratio B _i of each server M _i, can also be obtained from the server system running in real time. Alternatively, the usage rate acquisition unit 12 may acquire the resource usage rate B _i of each server M _i by reading a log file that records the operation results. The usage rate acquisition unit 12 records the acquired resource usage rate B _i in the usage rate table DB. In this way, the usage rate table DB indicating the correspondence between the server M _i and the resource usage rate B _i is created (step S12).

  Further, the threshold acquisition unit 13 acquires a threshold R of the resource usage rate. For example, the user inputs an appropriate threshold value R for server operation by using an input device (not shown). Alternatively, an appropriate threshold value R may be dynamically input from an external system. The threshold acquisition unit 13 receives the input threshold R, and stores the received threshold R in the storage device 2 as threshold data DR (step S13).

  Note that the order of steps S11 to S13 is arbitrary. Typically, first, the performance value table DA and the threshold data DR are created, and the usage rate table DB is updated in real time.

Step S20:
The remaining capacity calculation unit 20 calculates the remaining capacity C _i of each server M _i . The calculated remaining capacity C _i of each server M _i is referred to when determining the job assignment destination. Therefore, it is preferable that the remaining force C _i is estimated as accurately as possible.

For example, there is a method that considers that the lower the CPU usage rate, the greater the remaining capacity. However, in a heterogeneous environment, simply assigning a job to a server with the lowest CPU usage rate does not necessarily reduce the processing time of the job. This is because a high-performance server may process assigned jobs faster than other servers even if the CPU usage rate is relatively high. Therefore, in this embodiment, the performance value A _i is taken into account for calculating the remaining power C _i . In other words, the remaining capacity C _i includes the performance value A _i as a parameter.

Further, when the resource usage rate _Bi of a certain server exceeds a predetermined threshold value R, the effective processing performance of that server is significantly deteriorated. Therefore, in the present embodiment, the difference (R−B _i ) between the threshold R and the resource usage rate B _i is also taken into account for calculating the remaining capacity C _i . In other words, the remaining capacity C _i includes a difference (R−B _i ) between the threshold R and the resource usage rate B _i as a parameter.

According to the present embodiment, the remaining power calculation unit 20 calculates the remaining power C _i of each server M _i based on the performance value A _i and the difference (R−B _i ). For this purpose, the surplus power calculation unit 20 refers to the performance value table DA, the usage rate table DB, and the threshold data DR stored in the storage device 2. Then, the remaining power calculation unit 20 calculates the remaining power C _i of each server M _i according to the following equation (1), for example.

C _i = A _i × (R−B _i ) (1)

That is, the remaining power calculation unit 20 calculates the remaining power C _i by multiplying the difference between the threshold value R and the resource usage rate B _i by the performance value A _i . Then, the remaining force calculation unit 20 creates a remaining force table DC by using the calculated remaining force C _i . The surplus power table DC indicates the correspondence between the server M _i and the calculated surplus power C _i and is stored in the storage device 2. When determining an assignment destination of a certain job, the surplus power table DC may be referred to.

As an example, consider a server system including _two servers M ₁ and M ₂ . In this example, the performance values A ₁ and A ₂ of the servers M ₁ and M ₂ are 3 and 5, respectively. Further, the resource usage rates B ₁ and B ₂ of the servers M ₁ and M ₂ are 0.2 and 0.4, respectively. In terms of performance values, towards the server M ₂ it is more sophisticated, considered preferable as an allocation destination. On the other hand, from the viewpoint of resource utilization towards the server M ₁ is a lower load, considered preferable as an allocation destination. As described above, the optimum assignment destination cannot be objectively determined only from the performance value or the resource usage rate. On the other hand, the remaining capacity C _{i according} to the present embodiment has a performance value A _i and a difference (R−B _i ) as parameters, and can be a useful guideline for determining an allocation destination.

When the threshold value R is set to 0.6, according to the above formula (1), the remaining powers C ₁ and C ₂ of the servers M ₁ and M ₂ are calculated as 1.2 and 1.0, respectively. In this case, towards the server M ₁ has a larger margin. Here, consider a job assignment. It is assumed that when a server with a performance value of 3.5 has processed the job in the past, the resource usage rate was 0.3 (30%). Therefore, the minimum necessary capacity for processing the job is estimated to be 1.05 (= 3.5 × 0.3), which is the product of the performance value and the resource usage rate. Required margin (1.05) is because it is less reserve capacity of the server _{M 1} (1.2), the job can be determined to be assigned to the server _{M 1.} On the other hand, the required margin (1.05) because it is over the spare capacity of the server _{M 2} (1.0), the job can be judged not assigned to the server _{M 2.} If the relevant job is assigned to a server _{M 2,} resource utilization ratio _{B 2} servers _{M 2} is 0.21 (= 1.05 / 5) by increasing, 0.61 (= 0.4 + 0.21 ) In other words, resource utilization ratio _{B 2} servers _{M 2} will exceed the threshold value (0.6). This leads to significant degradation of the processing performance of the server M _2. In other words, the allocation destination can be determined so that the resource usage rate does not exceed the threshold by referring to the remaining capacity C _{i according} to the present embodiment. As a result, it is possible to prevent deterioration in processing performance.

The remaining force C _i greatly depends on the set value of the threshold value R. For example, if the threshold value R is 0.7, margin _C 1, _{C 2} of the server _M 1, _{M 2} is calculated both equal 1.5. Moreover, when the threshold value R is 0.8, the surplus forces C ₁ and C ₂ of the servers M ₁ and M ₂ are calculated as 1.8 and 2.0, respectively. In this case, contrary to the case the threshold R is 0.6, towards the server M ₂ has a larger margin. Since the threshold value R is set to an appropriate value for server operation, the calculated surplus capacity C _i is also a value that matches the server operation. That is, according to the present embodiment, it can be said that an appropriate guideline suitable for server operation is provided. Considering the threshold value R for the calculation of the surplus power C _i is also important from this viewpoint.

As described above, according to this embodiment, useful information (remaining capacity C _i ) is provided when determining the job assignment destination. By referring to the information, it is possible to determine a computer that can process a newly generated job, and to prevent deterioration in processing performance of the computer.

3. Planning Tool The computer capacity calculation apparatus 1 according to the present embodiment can be used as a “planning tool” for planning when and which server processes a job.

3-1. First Example FIG. 4 is a block diagram showing a first example of a planning tool to which the computer remaining power calculation device 1 is applied. A description overlapping with the above description is omitted as appropriate.

As shown in FIG. 4, the planning tool includes a storage device 2, a processing device 3, and an output device 4. The output device 4 is a device for outputting a margin information to the user relating to the available capacity C _i which is calculated (presented). For example, the output device 4 is a display device for displaying the remaining power information related to the calculated reserve capacity C _i. The processing device 3 includes an output unit 30 in addition to the data input unit 10 and the remaining power calculation unit 20 described above. The output unit 30 is also realized by the CPU of the processing device 3 executing the program PROG.

FIG. 5 is a flowchart showing processing by the planning tool. As described above, the data input unit 10 receives various data, and stores the performance value table DA, the usage rate table DB, and the threshold data DR in the storage device 2 (step S10). At this time, if the planning tool is connected to the operating server system via the network, the data input unit 10 can receive the resource usage rate B _i of each server M _i from the server system in real time. Next, the remaining power calculation unit 20 refers to the performance value table DA, the usage rate table DB, and the threshold data DR, calculates the remaining power C _i , and creates a remaining power table DC (step S20). Next, the output unit 30 reads the remaining power table DC from the storage device 2 and outputs the remaining power information related to the remaining power C _i to the output device 4. In the first example, the output unit 30 outputs the contents of the entire surplus power table DC to the output device 4. The output device 4 presents the contents of the remaining capacity table DC to the user (step S30).

Next, a specific processing example will be described. As shown in FIG. 4, the server system includes three servers M ₁ , M ₂ , and M ₃ . The server _{M 1, M 2, M 3} , respectively ID number 1-3 is assigned. The CPU performance values A ₁ , A ₂ and A ₃ of the servers M ₁ , M ₂ and M ₃ are 2.0, 3.0 and 5.0, respectively. The CPU usage rates B ₁ , B ₂ , and B ₃ at a certain point in time for the servers M ₁ , M ₂ , and M ₃ are 0.25 (25%), 0.40 (40%), and 0.60 (60%), respectively. ). Further, the threshold R of the CPU usage rate is set to 0.80 (80%). In this case, the performance value table DA, the usage rate table DB, the threshold data DR, and the surplus power table DC are as shown in FIG. The remaining powers C ₁ , C ₂ , and C ₃ of the servers M ₁ , M ₂ , and M ₃ are calculated as 1.1, 1.2, and 1.0, respectively.

The output device 4, for example, a display device displays the contents of the surplus table DC. For example, the name and ID number of each server M _i and the corresponding remaining power C _i are displayed. The user refers to the information displayed can be the newly generated job each server M _i is determined whether feasible. For example, assume that the same type of job has occurred in the past, and the minimum available capacity for processing the job is 1.05. In this case, it is understood that the servers M ₁ and M ₂ can process the job, but the server M ₃ cannot. Therefore, it is possible to plan the assignment of jobs to the server M ₂ having the maximum margin for example.

Further, the surplus capacity C _i may be referred to when the virtual server is moved (transferred) between physical servers. For example, it is assumed that physical servers other than the physical servers M ₁ , M ₂ , and M ₃ are overloaded. It is assumed that a virtual server requiring a surplus capacity of 1.15 is operating on the physical server. In this case, it is understood that migratable the virtual server to a physical server M _2. The overload state can be resolved by the migration of the virtual server. In this way, it is possible to create a server migration plan by referring to the surplus power C _i .

3-2. Second Example FIG. 7 is a block diagram showing a second example of a planning tool to which the computer remaining power calculation device 1 is applied. A description overlapping with the description of the first example is omitted as appropriate. In the second example, the processing device 3 includes a remaining power calculation unit 20 ′ instead of the remaining power calculation unit 20. In addition to the function of the remaining power calculating unit 20, the remaining power calculating unit 20 ′ has a sorting function. That is, the remaining power calculation unit 20 ′ calculates the remaining power C _i of each server M _i and sorts a plurality of servers based on the calculated remaining power C _i . In spare capacity table DC is created as a result, multiple entries are sorted based on the order of magnitude of the margin C _i. For example, the plurality of entries are sorted in descending order of the remaining capacity C _i .

Next, consider the same specific example as described in the first example. FIG. 8 shows a reserve table DC created in the second example. As described above, the remaining powers C ₁ , C ₂ , and C ₃ of the servers M ₁ , M ₂ , and M ₃ are calculated as 1.1, 1.2, and 1.0, respectively. Therefore, as a result of the sorting, the entries in the surplus power table DC are in the order of the servers M ₂ , M ₁ and M ₃ . The output device 4, for example, the display device displays the sorted surplus table DC.

The sorting process described above, even when not allocate jobs to the server M ₂ with a maximum margin for some reason, it is possible to quickly identify the next allocation candidate.

The surplus power is calculated with respect to a certain computer resource (CPU in this example). On the other hand, the server has resources such as a disk and a memory in addition to the CPU, and job processing generally requires a plurality of computer resources. For example, it is assumed that for a job processing, a remaining disk capacity of 0.3 is required in addition to a remaining CPU capacity of 0.8. At this time, if the server M ₂ holds the remaining capacity of the disk at 0.3 or more, the job can be assigned to the server M ₂ . However, if disk margin of server M ₂ is less than 0.3, it is reasonable to consider server M ₁ is the next candidate as assigned.

Alternatively, consider a case where a job is a virtual server (virtual machine) assigned to a physical server. There may be a case where there is an upper limit on the total number of virtual servers allocated on a physical server due to limitations of virtualization software that implements virtualization. For example, it is assumed that the upper limit number of virtual servers has already been assigned to the server M ₂ having the maximum remaining capacity. In this case, although the server M ₂ has sufficient capacity, a new virtual server cannot be assigned to the server M ₂ . Therefore, it is reasonable to consider server M ₁ is the next candidate as assigned.

3-3. Third Example FIG. 9 is a block diagram showing a third example of a planning tool to which the computer remaining power calculation device 1 is applied. The description overlapping with the description of the first and second examples is omitted as appropriate. In the third example, the processing device 3 includes an output unit 30 ′ instead of the output unit 30. The output unit 30 ′ reads the surplus power table DC from the storage device 2, and extracts the server having the maximum surplus power C _i from the surplus power table DC. Then, the output unit 30 ′ selectively outputs only the contents of the entry relating to the server having the maximum remaining capacity to the output device 4.

The output device 4 selectively presents the name and ID number of the server having the maximum available power and the maximum available power to the user. In the above embodiment, the display device selectively displays only the server M ₂ and the maximum available capacity 1.2. Consider a case in which the minimum remaining power required for processing a newly generated job is a value (for example, 1.3) greater than the maximum remaining power 1.2. In this case, by comparing these two numerical values, it is immediately determined that there is no server that can process the job.

  Thus, according to the third example, only the server having the maximum remaining capacity and the maximum remaining capacity are presented to the user. As a result, it is possible to quickly grasp the optimal server as a job assignment destination. Alternatively, it is possible to immediately recognize that there is no server that can process the job.

4). Load Balancing Device The computer surplus computing device 1 according to the present embodiment can also be used as a “load balancing device” for distributing the load of each server. The load balancer is applied to a server system and assigns a job such as a request or a virtual machine to a server with a low load in real time.

  FIG. 10 is a block diagram illustrating an example of a load balancer to which the computer surplus calculator 1 is applied. A description overlapping with the above description is omitted as appropriate. As shown in FIG. 10, the load balancer is connected to the server system via a network. The processing device 3 of the load balancer includes a job allocation unit 40 in addition to the data input unit 10 and the remaining power calculation unit 20 described above. The job assignment unit 40 is also realized by the CPU of the processing device 3 executing the program PROG.

FIG. 11 is a flowchart showing processing by the load balancer. As described above, the data input unit 10 receives various data, and stores the performance value table DA, the usage rate table DB, and the threshold data DR in the storage device 2 (step S10). At this time, the data input unit 10 receives the resource usage rate B _i of each server M _i from the server system in real time. Next, the remaining power calculation unit 20 refers to the performance value table DA, the usage rate table DB, and the threshold data DR, calculates the remaining power C _i , and creates a remaining power table DC (step S20).

The load balancer monitors the generation of a new job processing request (step S41). If a new job processing request has not occurred (step S41; No), the processing returns to step S10. Specifically, the data input unit 10 updates the usage rate table DB based on the resource usage rate B _i received from the server system in real time. Further, the surplus power calculation unit 20 updates the surplus power table DC.

When the load balancer receives a new job (step S41; Yes), the job allocation unit 40 refers to the remaining power table DC stored in the storage device 2 and determines an allocation destination of the job (step S42). . For example, the job assignment unit 40 of the server system, the calculated reserve capacity C _i selects the largest server as an allocation destination of the job. Then, the job assignment unit 40 assigns the job to the selected server (step S43). In the above embodiment, the job assigning unit 40 assigns the job to the server M _2. Server M ₂ executes the assigned job (request or virtual servers).

  As described above, the newly generated job is allocated to an appropriate server having sufficient remaining capacity by the load balancer. As a result, the load on each server is distributed. Note that a representative one of a plurality of servers constituting the server system may play the role of a load distribution apparatus.

  The computer remaining power calculation device 1 according to the present embodiment can be applied to various systems and cases other than the examples described above.

FIG. 1 is a schematic diagram illustrating a configuration of a server system. FIG. 2 is a block diagram showing the configuration of the computer remaining power calculation apparatus according to the embodiment of the present invention. FIG. 3 is a flowchart showing processing by the computer remaining power calculation apparatus according to the embodiment of the present invention. FIG. 4 is a block diagram illustrating a first example of the planning tool according to the embodiment of the present invention. FIG. 5 is a flowchart showing processing by the planning tool. FIG. 6 is a conceptual diagram illustrating an example of a performance value table, a usage rate table, threshold data, and a power table. FIG. 7 is a block diagram showing a second example of the planning tool according to the embodiment of the present invention. FIG. 8 is a conceptual diagram illustrating an example of a reserve table. FIG. 9 is a block diagram showing a third example of the planning tool according to the embodiment of the present invention. FIG. 10 is a block diagram illustrating an example of a load distribution apparatus according to an embodiment of the present invention. FIG. 11 is a flowchart showing processing by the load balancer.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Computer surplus power calculation apparatus 2 Storage apparatus 3 Processing apparatus 4 Output apparatus 10 Data input part 11 Performance value acquisition part 12 Usage rate acquisition part 13 Threshold acquisition part 20, 20 'Surplus power calculation part 30, 30' Output part 40 Job allocation part DA Performance value table DB usage rate table DR threshold data DC surplus power table PROG program A _i performance value B _i resource usage rate R threshold C _i surplus power

Claims (10)

A storage device storing performance value information indicating the performance values of each of the plurality of computers, usage rate information indicating the resource usage rate of each of the computers, and threshold information indicating a threshold value of the resource usage rate;
A processor that refers to the performance value information, the usage rate information, and the threshold value information, and calculates a remaining capacity of each of the computers based on a difference between the threshold value and the resource usage rate and the performance value. Computer capacity calculation device. The computer remaining power calculation device according to claim 1,
The processing device calculates the remaining capacity of each of the computers by multiplying the difference between the threshold value and the resource usage rate by the performance value. The computer remaining power calculation device according to claim 1 or 2,
Further, a computer surplus power calculation device comprising an output device that outputs surplus power information related to the calculated surplus power. The computer remaining power calculation device according to claim 3,
The surplus power information is a surplus power table indicating a correspondence relationship between the plurality of computers and the calculated surplus power. The computer remaining power calculation device according to claim 4,
The surplus power table entries are sorted based on the order of the calculated surplus power. The computer remaining power calculation device according to claim 3,
The surplus power information is a computer surplus power calculation device that selectively indicates a computer having the maximum calculated surplus power among the plurality of computers and the maximum surplus power. The computer remaining power calculation device according to any one of claims 3 to 6,
The output device is a display device that displays the remaining power information. The computer remaining power calculation device according to claim 1 or 2,
The processing device allocates a job to any of the plurality of computers by referring to the calculated surplus power. (A) obtaining each performance value of a plurality of computers;
(B) obtaining a resource usage rate of each of the computers;
(C) obtaining a threshold value of the resource usage rate;
(D) calculating a remaining capacity of each computer based on a difference between the threshold value and the resource usage rate and the performance value. (A) obtaining a performance value of each of a plurality of computers;
(B) obtaining a resource usage rate of each of the computers;
(C) obtaining a threshold for resource utilization;
(D) A computer remaining capacity calculation program that causes a computer to calculate the remaining capacity of each computer based on a difference between the threshold value and the resource usage rate and the performance value.

Images