JP2014206805A - Control device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a control device of a virtual machine capable of improving utilization efficiency of an operating environment.SOLUTION: A control device includes: a storage unit for storing a table in which performance information of a plurality of hardware resources capable of operating a plurality of virtual machines is stored; a resource measuring unit for measuring the current amount of resource consumption of the plurality of hardware resources by operating the plurality of virtual machines by the plurality of hardware resources; a resource calculation unit for calculating the amount of resource consumption of the plurality of hardware resources by operating the plurality of virtual machines after being migrated among the plurality of hardware resources by the plurality of hardware resources; and a migration unit for making the plurality of virtual machines migrate among the plurality of hardware resources when the amount of resource consumption calculated by the resource calculation unit is smaller than the current amount of resource consumption measured by the resource measuring unit.

Description

  The present invention relates to a virtual machine control device.

  Redundancy of the system configuration is effective for improving the fault tolerance of the computer system. For example, in a computer system having a high level of demand for stable operation, a system configuration is generally adopted in which a standby system that switches to an operating state when an active system stops operation due to a failure occurs. In this case, the standby system requires the same resources as the active system in case of failure. However, the standby system resource is not used during the operation of the active system, so that a surplus resource is generated.

  In recent years, with the use of virtualization technology, live migration technology is used to migrate OS and software running on a running virtual machine (VM) to another physical server without stopping. Even when the load on the operating physical server is high, another physical server having surplus resources tries to move the virtual machine to operate without stopping the service (for example, Patent Documents). 1).

JP 2011-197852 A

  However, the invention described in Patent Document 1 described above has a configuration in which the management server requests a plurality of physical servers to notify the resource usage rate when a resource shortage occurs. If a physical server having sufficient surplus resources cannot be secured, there is a possibility that migration cannot be executed safely.

  In addition, since the resource consumption of the entire system including multiple physical servers is not reduced, even if the shortage of resources can be resolved temporarily, the operating efficiency of the operating environment as a whole system There is a possibility that it cannot be raised. Here, the “system” is configured such that hardware resources such as a plurality of physical servers that operate a plurality of virtual machines are connected to a control device such as a management server so that they can communicate with each other via a network. Virtual environment.

  The present invention has been made in consideration of the above facts, and the problem is that it is possible to efficiently and safely operate live migration among a plurality of hardware resources, and to optimize resources as a whole system. To obtain a control device for a virtual machine that can increase the utilization efficiency of the operating environment.

  According to an embodiment of the present invention, a storage unit storing a first table storing performance information of a plurality of hardware resources on which one or more virtual machines can operate, and the one or more virtual machines A resource measuring unit that measures current resource consumption of the plurality of hardware resources by operating with the plurality of hardware resources; and the one or more virtual machines between the plurality of hardware resources. Resource consumption of the plurality of hardware resources by operating the one or more virtual machines after the migration to hardware resources having high performance according to the performance information among the hardware resources. And a current resource measured by the resource measuring unit. The smaller the resource consumption amount calculated by the resource calculating unit than consumption control device is provided with a migration unit for migrating the one or more virtual machines between the plurality of hardware resources.

  The storage unit stores a second table that stores information indicating whether migration is possible between the plurality of hardware resources of the one or more virtual machines, and the migration unit is configured to store the one or more virtual machines. May be determined according to the migration possibility stored in the second table when migrating between the plurality of hardware resources.

  The first table stores rank information indicating a rank based on the performance information of the plurality of hardware resources, and the resource calculation unit is configured based on the rank information stored in the first table. The plurality of hardware resources for migrating a plurality of virtual machines may be determined.

  The resource measuring unit calculates resource statistical information for calculating statistical information of temporal fluctuations of current resource consumption of the plurality of hardware resources by operating the one or more virtual machines with the plurality of hardware resources. And calculating the availability of migration between the plurality of hardware resources of the one or more virtual machines by time zone using the statistical information calculated by the resource statistical information calculation unit. You may have the calculation part classified by time zone stored in this table.

  The second table stores information indicating whether real-time migration is possible for the one or more virtual machines, and the migration unit stores each of the one or more virtual machines stored in the second table. Depending on whether real-time live migration is possible or not, migration may be performed by real-time or batch processing.

  The second table may store the availability of migration between the plurality of hardware resources of the one or more virtual machines by representing the migration priority.

  A performance information acquisition unit that acquires performance information of the plurality of hardware resources, and converts the performance information acquired by the performance information acquisition unit into performance information expressed on a scale that is unified among the plurality of hardware resources. And a performance information conversion unit stored in the first table.

  According to an embodiment of the present invention, a first table storing performance information of a plurality of hardware resources on which one or more virtual machines can operate is stored, and the one or more virtual machines are stored in the plurality of virtual machines. A current resource consumption amount of the plurality of hardware resources by operating with the hardware resources of the plurality of hardware resources, and the one or the plurality of virtual machines among the plurality of hardware resources among the plurality of hardware resources Calculating resource consumption of the plurality of hardware resources by operating the one or plurality of virtual machines after being migrated to hardware resources having high performance according to performance information, and operating the plurality of hardware resources; If the calculated resource consumption is smaller than the measured current resource consumption, the Or computer-readable storage medium storing a program for executing the be migrated to a computer a plurality of virtual machines among the plurality of hardware resources are provided.

  According to the present invention, it is possible to efficiently and safely operate live migration between a plurality of hardware resources, and to improve the utilization efficiency of the operating environment by optimizing the resources of the entire system. A control device is provided.

It is a block diagram which shows schematic structure of the system which concerns on one Embodiment of this invention. It is a figure which shows the table which the control apparatus of the virtual machine which concerns on one Embodiment of this invention has. It is a figure which shows the table which the control apparatus of the virtual machine which concerns on one Embodiment of this invention has. It is a flowchart which shows the flow of a process in the control apparatus of the virtual machine which concerns on one Embodiment of this invention. It is a figure for demonstrating the process in the control apparatus of the virtual machine which concerns on one Embodiment of this invention.

  The virtual machine control device of the present invention will be described in detail below with reference to FIGS.

  FIG. 1 is a block diagram showing a schematic configuration of a system 10 according to an embodiment of the present invention. As shown in FIG. 1, in a system 10 according to an embodiment of the present invention, a plurality of physical servers 13-1 and 13-2 and a control device 11 are connected via a network 12 so that they can communicate with each other. It has a configuration. In FIG. 1, “physical server” is illustrated as an example of hardware resources constituting the system 10, but any hardware resource capable of operating a virtual machine is not limited to a physical server, but a storage resource. It may be a device or the like.

  The physical server 13-1 is a physical machine that operates a plurality of virtual machines VM1, VM2, and VM3, and the physical server 13-2 is a physical machine that operates a plurality of virtual machines VM4, VM5, and VM6. The physical servers 13-1 and 13-2 are servers that provide Web services and the like using the virtual machines VM <b> 1 to VM <b> 6, respectively, and realize, for example, IaaS (Infrastructure as a Service). FIG. 1 illustrates a configuration including three virtual machines for one physical server. Hereinafter, the number of physical servers and virtual machines is an arbitrary number of 1 or more.

  The control device 11 includes a storage unit 11-1, a resource measurement unit 11-2, a resource calculation unit 11-3, and a migration unit 11-4. The storage unit 11-1 includes a first table that stores performance information of the plurality of physical servers 13-1 and 13-2 on which the plurality of virtual machines VM1 to VM6 can operate. In addition, the storage unit 11-1 may include a second table that stores information indicating whether or not the plurality of virtual machines VM1 to VM6 can be migrated between the plurality of physical servers 13-1 and 13-2. .

  The resource measurement unit 11-2 calculates the current resource consumption of the plurality of physical servers 13-1 and 13-2 by operating the plurality of virtual machines VM1 to VM6 on the plurality of physical servers 13-1 and 13-2. taking measurement.

  The resource calculation unit 11-3 migrates the virtual machines VM1 to VM6 after the migration of the virtual machines VM1 to VM6 between the physical servers 13-1 and 13-2 to the physical servers 13-1. The resource consumption of the plurality of physical servers 13-1 and 13-2 by operating in 13-2 is calculated.

  If the resource consumption calculated by the resource calculation unit 11-3 is smaller than the current resource consumption measured by the resource measurement unit 11-2, the migration unit 11-4 includes a plurality of virtual machines VM1 to VM6. Are migrated between the physical servers 13-1 and 13-2. At this time, the migration unit 11-4 may decide to execute the migration according to the availability of migration stored in the second table.

  The migration process performed by the control device 11 having such a configuration will be described in detail below with reference to FIGS. 2 and 3 regarding the configuration of the table included in the storage unit 11-1.

  2 and 3 are diagrams illustrating examples of tables included in the storage unit 11-1 of the control device 11 according to an embodiment of the present invention. FIG. 2A illustrates an example of a first table that stores performance information of the plurality of physical servers 1 to 3, and FIG. 3A illustrates information indicating whether migration of the plurality of virtual machines 1 to 3 is possible. Is an example of a second table for storing.

  Here, the performance information stored in the first table can be a numerical value defined for the response time of the hardware resource. The performance information may be, for example, the CPU clock number, spec (SPEC) value, and rPerf value of hardware resources such as the physical servers 1 to 3. The performance information may be a value arbitrarily defined by an administrator, a user, or the like.

  As shown in FIG. 2A, the first table stores performance information of the plurality of physical servers 1 to 3, rank information and status information of the plurality of physical servers 1 to 3, and the like. Also good. At this time, the performance information may include a performance value per system unit and a system performance value, as shown in FIG.

  The performance value per arithmetic unit indicates the processing capacity of each physical server 1 to 3 by a numerical value per arithmetic unit. For example, the CPU performance value may be a numerical value indicating the processing capacity per CPU core. Here, the CPU performance value may represent the processing capability of an arbitrary number of CPUs of the physical servers 1 to 3.

  The system performance value is a numerical value indicating the capacity of each physical server 1 to 3 and is a value set not only according to the number of CPUs but also according to the size of resources such as memory, network bandwidth, and hard disk. . The system performance value may be a value calculated by an administrator of the control device 11 with reference to the performance value per arithmetic unit and a performance value set for each manufacturer described later.

  The rank information of the physical servers 1 to 3 is information obtained by ranking the physical servers 1 to 3 based on the performance information of the physical servers 1 to 3. The rank information may be numerical values numbered in order from the physical servers 1 to 3 having the highest performance values. By using such rank information, it is possible to migrate virtual machines so that virtual machine processes are aggregated in preference to physical servers having high ranking orders, that is, high performance, in the migration process described later. It becomes possible. As described above, the rank information may be set as a priority for consolidating the processing of the virtual machines to the high performance physical server so that the low performance physical server can be appropriately removed. .

  The information regarding the status of each physical server 1 to 3 may store information regarding the current status of each physical server 1 to 3. For example, as shown in FIG. 2A, when a failure occurs in the physical server 3 or when a maintenance operation is performed, “1” is set in the status column by the administrator or the like. The server 3 may be set so that migration is not executed.

  In addition to the first table shown in FIG. 2A, the storage unit 11-1 stores performance values for each manufacturer of the plurality of physical servers 1 to 3, as shown in FIG. 2B. It may have a table. A performance value based on a unique standard for each manufacturer may be designed in the plurality of physical servers 1 to 3. FIG. 2B shows a table for storing information on performance values for each manufacturer.

  When the performance value is defined based on a different standard for each manufacturer for the plurality of physical servers 1 to 3, the control device 11 converts the performance value for each manufacturer into performance information expressed in a unified scale. It may be converted and stored in the first table. The control device 11 represents the performance information acquisition unit that acquires the performance information of the plurality of physical servers 1 to 3 and the performance information acquired by the performance information acquisition unit on a scale that is unified among the plurality of physical servers 1 to 3. A performance information conversion unit that converts the performance value into the first table, and converts the performance value for each manufacturer into a system performance value and stores it in the first table. May be.

  Next, with reference to FIG. 3, a table for storing information related to the plurality of virtual machines 1 to 3 will be described. As illustrated in FIG. 3A, the second table that stores information regarding the plurality of virtual machines 1 to 3 is a table that stores information indicating whether migration of the plurality of virtual machines 1 to 3 is possible.

  As illustrated in FIG. 3A, the table that stores information on the plurality of virtual machines 1 to 3 may store information on a migration method, migration availability, and migration priority.

  In the migration method column for the virtual machines 1 to 3, for example, “0” is set when migration is automatically executed in real time, and when migration is automatically started and executed by batch processing. When “1” is set and migration is performed manually by an administrator or the like, “2” is set. When visual confirmation is performed by an administrator or the like, migration is automatically started and executed by batch processing. “3” or the like may be set. Thereby, the migration method desired by the administrator can be set for each of the virtual machines 1 to 3.

  In the migration availability column, “0” can be set when migration is executed for each of the virtual machines 1 to 3, and “1” can be set when migration is not executed. Thus, for example, for the virtual machine 3 shown in FIG. 3A, even if it is set to manually execute migration in the migration method column, migration cannot be performed in the migration availability column. Can be set.

  In the priority column, the migration priority can be set. For example, as illustrated in FIG. 3A, when the priority order of the virtual machine 1 is “1” and the priority order of the virtual machine 2 is “2”, among the plurality of virtual machines 1 to 3, It is possible to decide to migrate the virtual machine 1 with priority over the virtual machine 2.

  As shown in FIGS. 3B and 3C, the storage unit 11-1 stores a table number for referring to the resource usage rate (%) of the plurality of virtual machines 1 to 3, and And a table for storing the resource usage rate (%) for each hour.

  The resource usage rate reference table number shown in FIG. 3B corresponds to the table number shown in FIG. Each of the virtual machines 1 to 3 has a predetermined resource usage pattern for each time zone, and each table shown in FIG. 3C calculates an average value of the resource usage rates for each time zone. It represents. By referring to the resource usage rate table for each time zone shown in FIG. 3C, it is possible to determine the time zone in which the surplus resource exists and execute the migration.

  The resource usage rate data for each time slot shown in FIG. 3C may be calculated by statistically analyzing the resource consumption measured by the resource measuring unit 11-2 of the control device 11. Hereinafter, information obtained by statisticizing such resource consumption is referred to as statistical information. The statistical information is not limited to the resource usage rate data for each time zone shown in FIG. 3C, but includes data such as the number of CPUs used and the CPU usage time for each time zone of each virtual machine 1-3. May be.

  Such statistical information is calculated by the resource statistical information calculation unit of the resource measurement unit 11-2. The resource measuring unit 11-2 calculates the statistical information of the time variation of the current resource consumption of the plurality of physical servers 1 to 3 by operating the plurality of virtual machines 1 to 3 on the plurality of physical servers 1 to 3. Using the statistical information calculated by the resource statistical information calculation unit and the resource statistical information calculation unit, the availability of migration between the plurality of physical servers 1 to 3 of the plurality of virtual machines 1 to 3 for each time zone is calculated, You may have the calculation part classified by time zone stored in 2 tables.

  As described above, the control device 11 performs migration of the plurality of virtual machines 1 to 3 between the plurality of physical servers 1 to 3 using the information stored in the table included in the storage unit 11-1.

  Hereinafter, the migration processing by the control device 11 will be described in more detail with reference to FIG.

  FIG. 4 is a flowchart showing the flow of migration processing by the control apparatus 11 according to an embodiment of the present invention.

  As illustrated in FIG. 4, the control device 11 first acquires performance data including the performance information of each physical server 1 to 3 included in the system 10 (S210). In S210, the performance information acquisition unit of the control device 11 acquires the performance value, rank information, current status information, and the like of each physical server 1 to 3 stored in the first table.

  Next, the control device 11 measures data on the current resource consumption of each of the virtual machines 1 to 3 (S220). The resource measuring unit 11-2 of the control device 11 measures the current resource consumption of the plurality of physical servers 1 to 3 by operating the plurality of virtual machines 1 to 3 on the plurality of physical servers 1 to 3.

  Next, the control device 11 executes comparison data processing (S230). In S230, the resource calculation unit 11-3 of the control device 11 first transfers the plurality of virtual machines 1 to 3 after the migration of the plurality of virtual machines 1 to 3 between the plurality of physical servers 1 to 3. The resource consumption of the plurality of physical servers 1 to 3 is calculated by operating the servers 1 to 3.

  At this time, the resource calculation unit 11-3 is based on the rank information of the physical servers 1 to 3 stored in the first table and the priority order of the virtual machines 1 to 3 stored in the second table. The combination of the virtual machines 1 to 3 that execute migration and the physical servers 1 to 3 that are migration destinations and the migration order pattern may be determined. The virtual machines 1 to 3 that execute the migration may be determined to execute the migration in descending order of priority based on the priority information of the virtual machines 1 to 3 stored in the second table. . Further, the migration destination physical servers 1 to 3 may be determined as the migration destination from the one with the highest ranking ranking based on the rank information of the physical servers 1 to 3 stored in the first table.

  The resource consumption of each of the physical servers 1 to 3 after migration for each pattern determined in this way is calculated by simulation.

  Next, the migration unit 11-4 of the control device 11 includes the current resource consumption measured by the resource measurement unit 11-2, and the resource consumption after migration calculated by the resource calculation unit 11-3. Compare As a result of the comparison processing, when the calculated resource consumption after migration is smaller than the current resource consumption, the combination of the virtual machines 1 to 3 and the migration destination physical servers 1 to 3 and the order of migration Determine the pattern.

  Next, the control device 11 determines whether migration is possible (S240). The migration unit 11-4 of the control device 11 determines whether the virtual machines 1 to 3 and the migration destination physical servers 1 to 3 determined in S230 are based on the information indicating whether migration is possible, which is stored in the second table. It is determined whether migration is actually possible for each combination and migration order pattern.

  If it is determined in S240 that migration is possible, the control device 11 determines whether or not migration processing in real time is possible (S250). At this time, the migration unit 11-4 may determine whether or not migration processing in real time is possible based on the migration method information of the virtual machines 1 to 3 stored in the second table. For example, if automatic execution is set as the migration method for all the migration target virtual machines 1 to 3, migration in real time may be executed.

  Also, by using the information stored in the tables shown in FIGS. 3B and 3C, the resource usage rate for each time zone of each of the virtual machines 1 to 3 is referred to determine whether real-time migration is possible. May be. For example, when the resource usage rate of the virtual machines 1 to 3 to be migrated is higher than a preset threshold value, the real-time migration may not be executed. In addition to the information on the migration method stored in the second table described above, the information on the resource usage rate for each time described above may be referred to, and the execution of migration in real time or batch processing may be determined.

  If it is determined in S250 that real-time migration is not possible, migration in batch processing is executed in S260. Migration by batch processing may be executed manually by an administrator or the like during a time period when each of the virtual machines 1 to 3 is not used, and is automatically started and executed according to a preset schedule. It may be a thing.

  In S250, when it is determined that real-time migration is possible, the control device 11 executes real-time migration.

  As described above, according to the control device 11 of the present invention, since the configuration for determining whether or not execution is possible in real time is performed before the migration is performed, the live migration is efficiently performed between the plurality of physical servers 1 to 3. It becomes possible to operate safely.

  As described above, according to the migration process by the control device 11 of the present invention, the processes of the virtual machines 1 to 3 are aggregated in the physical servers 1 to 3 having high performance, and the resource consumption of the entire system is reduced before the migration. It is possible to increase the resource utilization efficiency of the entire system by reducing the resource consumption amount of the system. Moreover, after consolidating the processing to the physical servers 1 to 3 having high performance among the plurality of physical servers 1 to 3, it is possible to remove the physical servers 1 to 3 in which the virtual machines 1 to 3 no longer operate. It becomes.

  Hereinafter, as an embodiment of the present invention, a process for migrating virtual machines A to E between a plurality of storage apparatuses 1 and 2 as hardware resources on which virtual machines A to E can operate will be described with reference to FIG. .

  FIG. 5 shows a process of migrating a virtual machine B and a virtual machine E between the storage apparatus 1 and the storage apparatus 2 having different CPU performance values by the virtual machine control apparatus 11 according to an embodiment of the present invention. It is a figure for demonstrating.

  The storage apparatus 1 has a CPU performance value “a”, and the storage apparatus 2 has a CPU performance value “b = 2a”. As shown in FIG. 5A, the virtual machines A and B operate in the storage apparatus 1 before the migration execution, and as shown in FIG. 5B, the virtual machines A and B operate in the storage apparatus 2 before the migration execution. Machines C, D and E are operating.

  At this time, the performance information acquisition unit of the control device 11 indicates that the CPU performance value of the storage device 1 is “a” and the CPU performance value of the storage device 2 is “b” from the first table of the storage unit 11-1. = 2a ”is acquired (S210).

  Next, the resource measuring unit 11-2 of the control device 11 operates the plurality of virtual machines A to E on the plurality of storage devices 1 and 2, respectively, thereby causing the current resource consumption of the plurality of storage devices 1 and 2 to operate. Is measured (S220).

  As shown in FIG. 5A, by the resource measuring unit 11-2, the virtual machine A executes a process that takes 60 minutes with 2 CPU cores in the storage apparatus 1 having the CPU performance value “a” before the migration. The virtual machine B is measured to execute a process that takes 40 minutes with 6 CPU cores. In addition, in the storage device 2 with the CPU performance value “b = 2a”, the resource measuring unit 11-2 executes a process that takes 60 minutes with 2 CPU cores, and the virtual machine D performs 20 minutes with 2 CPU cores. It is measured that the virtual machine E executes such a process with 0.5 CPU core for 30 minutes.

  The resource measurement unit 11-2 may calculate the resource consumption of the entire system using these pieces of measurement information. Here, for example, as shown in FIGS. 5A and 5B, the resource consumption of the entire system can be expressed as 55 when the processing amount of 10 minutes by one CPU core is 1.

  Next, the resource calculation unit 11-3 of the control device 11 executes comparison data processing (S230). First, the resource calculation unit 11-3 performs migration based on rank information of the storage apparatuses 1 and 2 stored in the first table and information such as the priorities of the virtual machines A to E stored in the second table. The combination of the virtual machines A to E that execute and the migration destination storage apparatuses 1 and 2 and the pattern of the migration order are determined.

  For example, in the second table, as the priority information of the virtual machines A to E, the virtual machine A and the virtual machine C require a longer processing time than the processing time of other virtual machines. It is assumed that the priority for execution is set low. On the other hand, the virtual machine B and the virtual machine E are set to have a high priority. In addition, since the CPU performance value of the storage apparatus 2 is higher than the CPU performance value of the storage apparatus 1 in the first table, the processes of the virtual machines A to E are aggregated in preference to the storage apparatus 2. Assume that rank information is set. Based on these pieces of information, a combination of virtual machines A to E that execute a plurality of migrations and migration destination storage apparatuses 1 and 2 and a pattern of the migration order are determined.

  The resource consumption of each of the storage apparatuses 1 and 2 after migration for each pattern thus determined is calculated by simulation. For example, the virtual machine E operating on the storage apparatus 2 is migrated to the storage apparatus 1 as indicated by the arrow A1 in FIG. 5, and the virtual machine B operating on the storage apparatus 1 is indicated by the arrow A2 in FIG. Is determined to be migrated to the storage apparatus 2. When migration is executed based on this pattern, as shown in FIG. 5C, the virtual machine E migrated to the storage apparatus 1 has 0.5 CPU cores in the storage apparatus 2 with the CPU performance value “b = 2a”. When processing that took 30 minutes was executed, the same processing can be completed in 30 minutes using one CPU core in the storage device 1 with a CPU performance value “a” that is half of the CPU performance value “b”. It becomes. Further, as shown in FIG. 5D, the virtual machine B migrated to the storage apparatus 2 is executing the process that takes 40 minutes with 6 CPU cores in the storage apparatus 1 having the CPU performance value “a”. The storage apparatus 2 having a CPU performance value “b” that is twice the CPU performance value “a” can be completed in 20 minutes using 6 CPU cores.

  If the resource consumption after migration calculated by such a simulation is expressed by the same method as the above-described current resource consumption, as shown in FIGS. The processing amount for 10 minutes can be represented as 43 by 1.

  Next, the migration unit 11-4 of the control device 11 includes the current resource consumption measured by the resource measurement unit 11-2, and the resource consumption after migration calculated by the resource calculation unit 11-3. Compare As a result of the comparison processing, since the calculated resource consumption 43 after migration is smaller than the current resource consumption 55, the virtual machine E is migrated to the storage apparatus 1 and the virtual machine B is A pattern to be migrated to the storage device 2 is determined.

  Next, the control device 11 determines whether migration is possible (S240). The migration unit 11-4 of the control device 11 determines whether migration is actually possible for the determined pattern based on the information indicating whether migration is possible stored in the second table. Further, the migration unit 11-4 determines to execute the migration process in real time or batch process based on the migration method information of the virtual machines 1 and 2 stored in the second table (S250). As described above, the migration process can be set in advance to start with a batch process or the like during a time period when the processes of the virtual machines A to E are not executed. Accordingly, the migration process can be executed safely when the virtual machines A to E are not operating.

  When the migration process is executed, as shown in FIGS. 5C and 5D, the storage apparatus 2 can create a resource unused state for 8 CPU cores after the first 20 minutes. Further, the storage apparatus 1 can create a non-use state of resources of 7 CPU cores or more. Therefore, the entire system can generate more surplus resources than before the migration. Further, the virtual machines A to E can be migrated to other storage apparatuses 2 with high performance values so that the storage apparatus 1 with low performance values can be appropriately removed. Accordingly, by consolidating the processes of the virtual machines B to D in the storage apparatus 2 having a higher performance value than the storage apparatus 1, it becomes possible to increase the resource usage efficiency of the entire system. In addition, it is possible to remove the storage apparatus 1 in which the virtual machines A to E no longer operate after aggregation.

  As described above, according to the migration process by the control device 11 according to an embodiment of the present invention, the performance is high among the plurality of hardware resources such as the physical servers 1 to 3 and the storage devices 1 and 2 described above. Since processing by the virtual machines A to E can be concentrated on hardware resources, the resource consumption of the entire system can be reduced more than the resource consumption before migration. Therefore, it is possible to achieve the effect of improving the resource utilization efficiency of the entire system. Further, as shown as an example in FIG. 5, by reducing the number of CPUs used as the entire system, for example, it is possible to achieve an effect of reducing the fee for the software license added according to the number of CPUs.

  In the embodiment described above with reference to FIG. 5, the migration processing using the CPU performance values of the storage apparatuses 1 and 2 as hardware resource performance information has been described as an example. The same applies to resources such as memory, network bandwidth, and hard disk.

  As described above, according to the migration process by the control device 11 according to an embodiment of the present invention, live migration can be efficiently and safely operated between a plurality of hardware resources, and the resource as a whole system is optimized. Therefore, it is possible to increase the usage efficiency of the operating environment.

DESCRIPTION OF SYMBOLS 10 System 11 Control apparatus 12 Network 13 Physical server 11-1 Memory | storage part 11-2 Resource measurement part 11-3 Resource calculation part 11-4 Migration part

Claims (8)

A storage unit that stores a first table that stores performance information of a plurality of hardware resources capable of operating one or a plurality of virtual machines;
A resource measuring unit that measures current resource consumption of the plurality of hardware resources by operating the one or more virtual machines with the plurality of hardware resources;
The one or more virtual machines after the migration of the one or more virtual machines to hardware resources having high performance according to the performance information among the plurality of hardware resources between the plurality of hardware resources. A resource calculation unit for calculating resource consumption of the plurality of hardware resources by operating with a plurality of hardware resources;
A migration unit that migrates the one or more virtual machines between the plurality of hardware resources if the resource consumption calculated by the resource calculation unit is smaller than the current resource consumption measured by the resource measurement unit And a control device. The storage unit stores a second table that stores information indicating whether migration is possible between the plurality of hardware resources of the one or more virtual machines,
2. The migration unit according to claim 1, wherein when the one or more virtual machines are migrated between the plurality of hardware resources, the migration unit determines whether to perform migration according to whether or not the migration is stored in the second table. Control device. The first table stores rank information indicating a rank based on performance information of the plurality of hardware resources,
3. The control device according to claim 1, wherein the resource calculation unit determines the plurality of hardware resources for migrating the one or more virtual machines based on rank information stored in the first table. The resource measuring unit calculates resource statistical information for calculating statistical information of temporal fluctuations of current resource consumption of the plurality of hardware resources by operating the one or more virtual machines with the plurality of hardware resources. Part
Using the statistical information calculated by the resource statistical information calculation unit, whether or not migration between the plurality of hardware resources of the one or more virtual machines for each time zone is calculated and stored in the second table. The control device according to claim 1, further comprising a time zone calculation unit. The second table stores information indicating whether real-time migration is possible for the one or more virtual machines,
The control device according to claim 2, wherein the migration unit performs migration by real time or batch processing depending on whether or not live migration of each of the one or more virtual machines stored in the table is possible.   The control device according to claim 2, wherein the second table represents whether or not migration between the plurality of hardware resources of the one or more virtual machines is represented by a priority of migration. A performance information acquisition unit that acquires performance information of the plurality of hardware resources;
A performance information conversion unit that converts performance information acquired by the performance information acquisition unit into performance information expressed on a scale that is unified among the plurality of hardware resources and stores the performance information in the first table. The control device according to any one of 1 to 6. Storing a first table storing performance information of a plurality of hardware resources on which one or a plurality of virtual machines can operate;
Measuring current resource consumption of the plurality of hardware resources by operating the one or more virtual machines with the plurality of hardware resources;
The one or more virtual machines after the migration of the one or more virtual machines to hardware resources having high performance according to the performance information among the plurality of hardware resources between the plurality of hardware resources. Calculating resource consumption of the plurality of hardware resources by operating with a plurality of hardware resources;
A program for causing a computer to migrate the one or more virtual machines between the plurality of hardware resources if the calculated resource consumption is smaller than the measured current resource consumption. A stored computer-readable storage medium.

Images