WO2022048674A1

WO2022048674A1 - Server cabinet-based virtual machine management method and apparatus

Info

Publication number: WO2022048674A1
Application number: PCT/CN2021/116816
Authority: WO
Inventors: 张振中; 姜臻颖
Original assignee: 华为云计算技术有限公司
Priority date: 2020-09-07
Filing date: 2021-09-07
Publication date: 2022-03-10

Abstract

The present application relates to the field of cloud computing, and particularly relates to a server cabinet-based virtual machine management method and apparatus. The method comprises: obtaining the first cabinet power of a server cabinet; and in the case that the first cabinet power is within a preset power range, sending a first request to a virtual machine migration apparatus, the first request being used for indicating the virtual machine migration apparatus to prohibit the creation of a virtual machine in the server cabinet. According to the method, when the power of the cabinet is high, the virtual machine is not created in the cabinet any more, so that the power of the cabinet is not increased as much as possible. Hence, the risks of power failure and down of the cabinet due to overload can be avoided or reduced.

Description

Method and device for virtual machine management based on server cabinet

technical field

The present application relates to the field of cloud computing, and in particular, to a server cabinet-based virtual machine management method and device.

Background technique

Cloud computing belongs to distributed computing. You can set up multiple servers, and the servers can run multiple virtual machines (VMs) as hosts to process multiple services at the same time, or distribute processing. Larger data processing programs. Wherein, the server can be set in a cabinet, and one cabinet can be set with one or more servers.

Cloud computing has the characteristics of elastic scaling, flexible deployment, high reliability, high security and high cost performance, and has been favored by more and more users. The users and numerous richness on the cloud computing platform (also called cloud data) make the demands of resources (computing resources, storage resources, communication resources, etc.) change dynamically over time. As a result, the load of one or more virtual machines may be too high, thereby causing the power of a certain cabinet to be too high. In addition, in order to reduce deployment costs, cloud computing platforms usually have high-density cabinets. As a result, the power supply of the cabinet is challenged. When each server in a rack is running at high load, the power of the entire rack may exceed the maximum load of the rack, resulting in power loss.

Therefore, there is an urgent need for a solution to avoid power failure of the cabinet.

SUMMARY OF THE INVENTION

The embodiments of the present application provide a virtual machine management method and device based on a server cabinet, which can rapidly migrate virtual machines under the condition of ensuring business continuity, thereby reducing the power of the cabinet and reducing the power overload of the cabinet as much as possible. the risk of power outages.

In a first aspect, the present application provides a virtual machine management method based on a server cabinet, which is applied to a cabinet manager of a server cabinet. The method includes: obtaining a first cabinet power of the server cabinet; when the first cabinet power is at a preset power In the case of the range, a first request is sent to the virtual machine migration apparatus, where the first request is used to instruct the virtual machine migration apparatus to prohibit the creation of the virtual machine in the server cabinet. Wherein, in an example, the virtual machine may be a tenant virtual machine (or a user virtual machine). In another example, the server cabinet is a cabinet in a data center, and the virtual machine may include a tenant virtual machine and a management virtual machine of the data center (a virtual machine provided with a cloud management component for managing the tenant virtual machine).

That is to say, when the cabinet power of the server cabinet is relatively high, no new virtual machines can be created in the server cabinet, so that the power of the server cabinet can be prevented from further increasing while ensuring business continuity. .

In a possible implementation manner, the first request includes a server identifier of each server set in the server rack, and the first request is used to instruct the virtual machine migration apparatus to prohibit the creation of a virtual machine in any server in the server rack.

In a possible implementation manner, the method further includes: when the power of the first cabinet is greater than a preset power range, sending a second request to the virtual machine migration apparatus, where the second request is used to instruct the virtual machine migration apparatus to confirm the server The virtual machine to be migrated in the cabinet and the virtual machine to be migrated out of the server cabinet.

That is to say, in this implementation manner, when the power of the server rack is further increased, one or more virtual machines in the server rack can be migrated out, so that the power of the server rack can be reduced, and business continuity can be ensured. At the same time, it avoids or reduces the risk of power failure and downtime, improves the reliability of the data center, and guarantees the user service level agreement.

In a possible implementation manner, the second request includes the power of the first rack, the virtual machine identifier and the virtual machine load of each virtual machine that has been created in each server in the server rack.

That is, in this implementation manner, the cabinet manager may send the virtual machine identifier and virtual machine load of the virtual machine already created in the server in the server cabinet, and the power of the first cabinet to the virtual machine migration apparatus, so that the virtual machine The migration apparatus may determine the virtual machine to be migrated.

In a possible implementation manner, before sending the second request to the virtual machine migration apparatus, the method further includes: querying each server in the server cabinet for the virtual machine identifier and virtual machine load of each virtual machine that has been created .

That is, in this implementation manner, the cabinet manager can query each server in the server cabinet for the relevant information of the server to obtain the required information.

In a possible implementation manner, the method further includes: when the power of the first cabinet is greater than a preset power range, determining the virtual machine to be migrated in the server cabinet; sending a virtual machine migration request to the virtual machine migration apparatus, and the virtual machine The migration request is used to instruct the virtual machine migration apparatus to migrate the virtual machine to be migrated out of the server cabinet.

That is, in this implementation manner, when the power of the server rack is further increased, the rack manager may determine the virtual machine to be migrated in the server rack, and may request the virtual machine migration apparatus to transfer the virtual machine to be migrated from the server rack to the virtual machine to be migrated. Migrating out of the server cabinet, thereby reducing the power of the server cabinet, avoiding or reducing the risk of power failure and downtime while ensuring business continuity, improving the reliability of the data center, and ensuring the user service level agreement.

In a possible implementation manner, the method further includes: pre-recording the target power; then, determining the virtual machine to be migrated in the server rack includes: determining, according to the load of each virtual machine that has been created on each server in the server rack, determining The power of each virtual machine that has been created on each server; according to the target power, the power of the first cabinet, and the power of each virtual machine in the server cabinet, the virtual machine to be migrated is selected from the virtual machines that have been created in the server cabinet.

That is to say, in this implementation manner, the power of the virtual machine can be determined according to the power of the server and the load of the virtual machine; and then the virtual machine to be migrated can be determined according to the power of the cabinet and the power of the virtual machine.

In a possible implementation manner, selecting the virtual machine to be migrated from the virtual machines already created in the server rack according to the target power, the power of the first rack, and the power of each virtual machine in the server rack, including: obtaining the first The difference between the power of a cabinet and the target power; select at least two virtual machines whose power sum is the closest and not less than the difference as the virtual machine to be migrated; or select a virtual machine whose power is the closest and not less than the difference as the virtual machine to be migrated The virtual machine to be migrated.

In a possible implementation manner, the power of each virtual machine that has been created on each server is determined according to the load of each virtual machine that has been created on each server in the server rack, including:

in,

represents the power of virtual machine i at time t;

Represents the power of the CPU of virtual machine i at time t;

represents the CPU utilization of virtual machine i at time t; α ₁ and β ₁ represent the weight and skew of CPU utilization-power conversion, respectively;

represents the power of the memory of the virtual machine i at time t;

represents the memory overhead of virtual machine i at time t; α ₂ and β ₂ represent the weight and skew of memory overhead-power conversion, respectively;

represents the power consumption of the hard disk of virtual machine i at time t;

Indicates the total number of bytes read and written on the hard disk of virtual machine i at time t, where,

Indicates the number of bytes written to the hard disk of virtual machine i at time t,

represents the number of hard disk read bytes of virtual machine i at time t; α ₃ and β ₃ represent the total number of hard disk read and write bytes - the weight and skew of power conversion, respectively.

In a possible implementation manner, after acquiring the power of the first rack of the server rack, the method further includes: acquiring the power of the second rack of the server rack; in the case that the power of the second rack is less than the preset power range, sending the A third request for the server identification of the server provided in the server rack is sent to the virtual machine migration apparatus, the third request is used to instruct the virtual machine migration apparatus to allow the creation of a virtual machine in the server rack.

That is, in this implementation manner, when the power of the server rack is low, the virtual machine migration apparatus can be instructed to create a virtual machine in the server rack, thereby improving the resource utilization of the server rack.

In a possible implementation manner, the power of the first cabinet is measured power or predicted power.

In a possible implementation manner, the virtual machine load includes at least one of the following:

CPU utilization, memory overhead, and the total number of bytes read and written to the hard disk.

In a second aspect, an embodiment of the present application provides a virtual machine management method based on a server cabinet, which is applied to a cabinet manager of a server cabinet. The method includes: obtaining a first cabinet power of the server cabinet; When the power of the cabinet is greater than the preset power range, send a second request to the virtual machine migration apparatus, where the second request is used to instruct the virtual machine migration apparatus to confirm the virtual machine to be migrated in the server cabinet and send the virtual machine to the virtual machine migration apparatus. The virtual machine to be migrated is migrated out of the server cabinet.

That is to say, when the power of the server rack is sufficiently increased, one or more virtual machines in the server rack can be migrated out, thereby reducing the power of the server rack, avoiding or reducing the power of the server rack while ensuring business continuity. The risk of power failure and downtime is reduced, the reliability of the data center is improved, and the user service level agreement is guaranteed.

In a possible implementation manner, the second request includes the power of the first rack, the virtual machine identifier of each virtual machine that has been created in each server in the server rack, and the virtual machine load.

In a possible implementation manner, before sending the second request to the virtual machine migration apparatus, the method further includes: querying each server in the server cabinet for the virtual machine identifier and virtual machine identifier of each virtual machine that has been created machine load.

In a third aspect, an embodiment of the present application provides a virtual machine management method based on a server cabinet, which is applied to a cabinet manager of a server cabinet. The method includes: obtaining a first cabinet power of a server cabinet; When the power is greater than the preset power range, determine the virtual machine to be migrated in the server cabinet; send a virtual machine migration request to the virtual machine migration device, where the virtual machine migration request is used to instruct the virtual machine migration device to migrate the virtual machine to be migrated out of the server cabinet .

That is, when the power of the server rack is further increased, the rack manager may determine the virtual machine to be migrated in the server rack, and may request the virtual machine migration apparatus to migrate the virtual machine to be migrated out of the server rack, Therefore, the power of the server cabinet can be reduced, and while ensuring business continuity, the risk of power failure and shutdown can be avoided or reduced, the reliability of the data center is improved, and the user service level agreement is guaranteed.

In a possible implementation manner, determining the virtual machines to be migrated in the server cabinet includes: determining a first among the at least two servers according to the server power of the at least two servers in the service cabinet and the load of each virtual machine on the at least two servers The power of the first virtual machine on the server; the virtual machine to be migrated is determined according to the power of the first cabinet and the power of each virtual machine in the service cabinet.

In a fourth aspect, an embodiment of the present application provides a virtual machine management method based on a server cabinet, which is applied to a virtual machine migration apparatus. The method includes: acquiring a first request sent by a cabinet manager, where the first request includes setting a management The server identifier of each server in the server cabinet where the server is located; the creation of a virtual machine in any server in the server cabinet is prohibited according to the first request.

In a fifth aspect, an embodiment of the present application provides a virtual machine management method based on a server cabinet, which is applied to a virtual machine migration apparatus. The method includes: obtaining a second request sent by the cabinet manager, where the second request includes the cabinet manager server Each virtual machine identifier and virtual machine load of the virtual machines that have been created in each server in the rack, and the first rack power of the server rack; each server rack that has been created in each server in the server rack carried according to the second request. The virtual machine identifier and the virtual machine load of each virtual machine, and the first cabinet power of the server cabinet determine the virtual machine to be migrated in the server cabinet; the virtual machine to be migrated is migrated out of the server cabinet.

In a possible implementation manner, the method further includes: pre-recording the target power of the server power; then, according to the virtual machine of each virtual machine that has been created in each server in the server cabinet carried by the second request The identification, the virtual machine load, and the first cabinet power of the server cabinet determine the virtual machine to be migrated in the server cabinet according to the second request. Determine the virtual machine to be migrated in the server in the server cabinet according to the second request, including: according to each server in the server cabinet. The load of each virtual machine that has been created, determines the power of each virtual machine that has been created on each server; based on the target power, the power of the first rack, and the power of each virtual machine in the server rack, the created power from the server rack Select the virtual machine to be migrated from the virtual machines in the virtual machine.

In a possible implementation manner, selecting the virtual machine to be migrated from the virtual machines created in the server rack according to the target power, the power of the first rack, and the power of each virtual machine in the server rack, including: obtaining the first rack The difference between the power and the target power; select at least two virtual machines whose sum of power is the closest and not less than the difference as the virtual machine to be migrated; or select a virtual machine whose power is the closest and not less than the difference as the virtual machine to be migrated The virtual machine to be migrated.

In a possible implementation manner, the power of each virtual machine that has been created on each server is determined according to the load of each virtual machine that has been created on each server in the server cabinet, including:

in,

represents the power of virtual machine i at time t;

Represents the power of the CPU of virtual machine i at time t;

represents the power of the memory of the virtual machine i at time t;

In a sixth aspect, an embodiment of the present application provides a virtual machine management method, which is applied to a virtual machine migration apparatus. The method includes: acquiring a virtual machine migration request sent by a cabinet manager, where the virtual machine migration request includes setting a virtual machine where the cabinet manager is located. The virtual machine identifier of the virtual machine to be migrated in the server cabinet; according to the virtual machine migration request, the virtual machine to be migrated is migrated out of the server cabinet.

In a seventh aspect, an embodiment of the present application provides a cabinet manager, which is configured in a server cabinet, and the cabinet manager includes: an acquisition unit, configured to obtain the first cabinet power of the server cabinet; In the case of a preset power range, a first request is sent to the virtual machine migration apparatus, where the first request is used to instruct the virtual machine migration apparatus to prohibit the creation of a virtual machine in the server cabinet.

In a possible implementation manner, the first request includes a server identifier of each server set in the server cabinet, and the first request is used to instruct the virtual machine migration apparatus to prohibit the creation of a virtual machine in any server in the server cabinet.

In a possible implementation manner, the sending unit is further configured to: when the power of the first cabinet is greater than a preset power range, send a second request to the virtual machine migration apparatus, where the second request is used to instruct the virtual machine migration apparatus to confirm The virtual machine to be migrated in the server cabinet and the virtual machine to be migrated out of the server cabinet.

In a possible implementation manner, the rack manager further includes a query unit; the query unit is further configured to: query each server in the server rack for the virtual machine identifier and virtual machine load of each virtual machine that has been created.

In a possible implementation manner, the cabinet manager further includes a determining unit; the determining unit is configured to determine the virtual machine to be migrated in the server cabinet under the condition that the power of the first cabinet is greater than the preset power range; the sending unit further uses In sending the virtual machine migration request to the virtual machine migration apparatus, the virtual machine migration request is used for instructing the virtual machine migration apparatus to migrate the virtual machine to be migrated out of the server rack.

In a possible implementation manner, the cabinet manager further includes a recording unit and a selection unit; the recording unit is used for pre-recording the target power; the determining unit is further used for recording the target power according to the Load, which determines the power of each virtual machine that has been created on each server; the selection unit is used to select from the virtual machines that have been created in the server cabinet based on the target power, the power of the first cabinet, and the power of each virtual machine in the server cabinet Select the virtual machine to be migrated.

In a possible implementation manner, the selection unit is further configured to: obtain the difference between the power of the first cabinet and the target power; and select at least two virtual machines whose sum of power is the closest and not less than the difference as the virtual machines to be migrated ; or select a virtual machine whose power is the closest and not less than the difference as the virtual machine to be migrated.

In a possible implementation manner, the determining unit is used to determine the power of each virtual machine that has been created on each server by using the following formula:

in,

represents the power of virtual machine i at time t;

Represents the power of the CPU of virtual machine i at time t;

represents the power of the memory of the virtual machine i at time t;

In a possible implementation manner, the obtaining unit is further configured to obtain the power of the second cabinet of the server cabinet; the sending unit is further configured to send the power provided in the server cabinet when the power of the second cabinet is less than a preset power range. The third request for the server identification of the server in the virtual machine migration apparatus is used to instruct the virtual machine migration apparatus to allow the virtual machine to be created in the server cabinet.

In an eighth aspect, an embodiment of the present application provides a cabinet manager, which is configured in a server cabinet. The cabinet manager includes: an obtaining unit, used for obtaining the first cabinet power of the server cabinet; When the power is greater than the preset power range, a second request is sent to the virtual machine migration apparatus, and the second request is used to instruct the virtual machine migration apparatus to confirm the virtual machine to be migrated in the server rack and migrate the to-be-migrated virtual machine out of the server rack.

In a possible implementation manner, the second request includes the power of the first rack, the virtual machine identifier of the virtual machine that has been created in each server in the server rack, and the virtual machine load.

In a possible implementation manner, the rack manager further includes: a query unit, configured to query each server in the server rack for the virtual machine identifier and virtual machine load of each virtual machine that has been created.

In a ninth aspect, an embodiment of the present application provides a cabinet manager, which is configured in a server cabinet. The cabinet manager includes: an obtaining unit, used for obtaining the first cabinet power of the server cabinet; When the power is greater than the preset power range, determine the virtual machine to be migrated in the server cabinet; the sending unit is configured to send a virtual machine migration request to the virtual machine migration device, where the virtual machine migration request is used to instruct the virtual machine migration device to transfer the virtual machine to be migrated. The machine is migrated out of the server cabinet.

In a tenth aspect, an embodiment of the present application provides an apparatus for migrating a virtual machine, including: an obtaining unit configured to obtain a first request sent by a cabinet manager, where the first request includes each data set in a server cabinet where the cabinet manager is located. The server identifier of each server; the prohibition unit is used to prohibit the creation of a virtual machine in any server in the server cabinet according to the first request.

In an eleventh aspect, an embodiment of the present application provides an apparatus for migrating a virtual machine, including: an obtaining unit configured to obtain a second request sent by a cabinet manager, where the second request includes an already created request in each server in the server cabinet Each virtual machine identifier of the virtual machine, the virtual machine load, and the first cabinet power of the server cabinet; a determining unit, used for each virtual machine that has been created in each server in the server cabinet carried according to the second request The virtual machine identifier and virtual machine load of the server cabinet, and the first cabinet power of the server cabinet determine the virtual machine to be migrated in the server cabinet; the migration unit is used for migrating the virtual machine to be migrated out of the server cabinet.

In a possible implementation manner, the virtual machine migration apparatus further includes: a recording unit and a selection unit; the recording unit is used for pre-recording the target power of the server power; The load of each virtual machine determines the power of each virtual machine that has been created on each server; the selection unit is used to select the Select the virtual machine to be migrated from the virtual machines in the virtual machine.

In a possible implementation manner, the selection unit is configured to: obtain a difference between the power of the first cabinet and the target power; select at least two virtual machines whose sum of power is closest to and not less than the difference as the virtual machines to be migrated; Or select a virtual machine with the closest power and not less than the difference as the virtual machine to be migrated.

in,

represents the power of virtual machine i at time t;

Represents the power of the CPU of virtual machine i at time t;

represents the power of the memory of the virtual machine i at time t;

In a twelfth aspect, an embodiment of the present application provides a virtual machine migration apparatus, including: an obtaining unit configured to obtain a virtual machine migration request sent by a cabinet manager, where the virtual machine migration request includes settings in a server cabinet where the cabinet manager is located The virtual machine identifier of the virtual machine to be migrated in; the migration unit is used for migrating the virtual machine to be migrated out of the server cabinet according to the virtual machine migration request.

In a thirteenth aspect, an embodiment of the present application provides a cabinet manager, including a processor, a memory, and a transceiver; the memory is used to store computer instructions; when the cabinet manager runs, the processor executes the computer instructions, so that the cabinet manager The method provided by the first aspect is performed.

In a fourteenth aspect, an embodiment of the present application provides a cabinet manager, including a processor, a memory, and a transceiver; the memory is used to store computer instructions; when the cabinet manager runs, the processor executes the computer instructions, so that the cabinet manager The method provided by the second aspect is performed.

In a fifteenth aspect, an embodiment of the present application provides a cabinet manager, including a processor, a memory, and a transceiver; the memory is used to store computer instructions; when the cabinet manager runs, the processor executes the computer instructions, so that the cabinet manager Execute the method provided by the third aspect.

In a sixteenth aspect, an embodiment of the present application provides a virtual machine migration apparatus, including a processor, a memory, and a transceiver;

The memory is used for storing computer instructions; when the virtual machine migration apparatus runs, the processor executes the computer instructions, so that the virtual machine migration apparatus executes the method provided by the fourth aspect.

In a seventeenth aspect, an embodiment of the present application provides a virtual machine migration apparatus, which is characterized by comprising a processor, a memory, and a transceiver; the memory is used to store computer instructions; when the virtual machine migration apparatus runs, the processor executes the computer The instructions cause the virtual machine migration apparatus to execute the method provided by the fifth aspect.

In an eighteenth aspect, an embodiment of the present application provides a virtual machine migration apparatus, which is characterized by comprising a processor, a memory, and a transceiver; the memory is used to store computer instructions; when the virtual machine migration apparatus runs, the processor executes the computer The instruction causes the virtual machine migration apparatus to execute the method provided by the sixth aspect.

In a nineteenth aspect, an embodiment of the present application provides a computer storage medium, where the computer storage medium includes computer instructions, and when the computer instructions are executed on an electronic device, the electronic device is made to execute the method provided in the first aspect or the method provided in the second aspect. The method provided or the method provided by the third aspect or the method provided by the fourth aspect or the method provided by the fifth method or the method provided by the sixth method.

In a twentieth aspect, an embodiment of the present application provides a computer program product. When the program code included in the computer program product is executed by a processor in an electronic device, the method provided in the first aspect or the method provided in the second aspect is implemented. The method provided by the third aspect or the method provided by the fourth aspect or the method provided by the fifth method or the method provided by the sixth method.

In a twenty-first aspect, an embodiment of the present application provides a virtual machine management system, including the cabinet manager and the virtual machine migration apparatus described above.

The virtual machine management method and device based on the service cabinet provided by the embodiments of the present application can no longer create a virtual machine in the server cabinet when the power of the server cabinet is high, so that the power of the server cabinet is not increased as much as possible; and When the power of the server rack is high enough, one or more virtual machines in the server rack can be migrated out to reduce the power of the rack; thus, the risk of power failure and shutdown of the server rack due to overload can be avoided or reduced.

Description of drawings

FIG. 1 is a schematic structural diagram of a data center to which an embodiment of the present application can be applied;

FIG. 2 is a flowchart of a virtual machine management method provided by an embodiment of the present application;

3 is a flowchart of a virtual machine management method provided by an embodiment of the present application;

FIG. 4 is a flowchart of a virtual machine management method provided by an embodiment of the present application;

FIG. 5 is a schematic structural diagram of a cabinet manager provided by an embodiment of the present application;

FIG. 6 is a schematic structural diagram of a virtual machine migration apparatus provided by an embodiment of the present application;

FIG. 7 is a schematic structural diagram of a virtual machine migration apparatus provided by an embodiment of the present application;

FIG. 8 is a schematic block diagram of a cabinet manager provided by an embodiment of the present application;

FIG. 9 is a schematic block diagram of a virtual machine migration apparatus provided by an embodiment of the present application.

detailed description

The technical solutions in the embodiments of the present invention will be described below with reference to the accompanying drawings. Obviously, the described embodiments are only some of the embodiments of the present specification, but not all of the embodiments.

In the description of this specification, "one embodiment" or "some embodiments" etc. means that a particular feature, structure or characteristic described in connection with the embodiment is included in one or more embodiments of this specification. Thus, appearances of the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in other embodiments," etc. in various places in this specification are not necessarily All refer to the same embodiment, but mean "one or more but not all embodiments" unless specifically emphasized otherwise.

Wherein, in the description of this specification, unless otherwise stated, "/" means or means, for example, A/B can mean A or B; "and/or" in this document is only an association to describe the associated object Relation, it means that there can be three kinds of relations, for example, A and/or B can mean that A exists alone, A and B exist at the same time, and B exists alone. In addition, in the description of the embodiments of the present specification, "plurality" refers to two or more than two.

In the description of this specification, the terms "first" and "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implying the number of indicated technical features. Thus, a feature defined as "first" or "second" may expressly or implicitly include one or more of that feature. The terms "including", "including", "having" and their variants mean "including but not limited to" unless specifically emphasized otherwise.

With the expansion of the scale of cloud computing systems, for the consideration of deployment cost, high-density cabinets (for example, the power load of each cabinet is 10kW or more) are often used to build cloud data centers. In this way, the power supply and heat dissipation of the entire cabinet are put under pressure, and the cabinet has a higher risk of power failure.

In one solution, a cabinet power capping strategy is used to control the power of the cabinet. Specifically, when the power of the cabinet reaches a certain value, the power of one or more servers in the cabinet is directly limited, so as to prevent the risk of power failure caused by the excessive power of the cabinet. When the power of the cabinet is high, this strategy seriously affects the user experience and cannot meet the user service level agreement (SLA).

In another solution, the load of the central processing unit (CPU) of each server in the cloud data center is monitored, and the size of the CPU resources obtained by the virtual machine is dynamically adjusted by migrating or expanding the capacity to meet the needs of a single virtual machine. quality of service (QoS). This solution does not consider the power problem of the cabinets. In a large-scale data center, in a business-intensive scenario, power imbalance between different cabinets may occur. For example, some cabinets may reach the power cap or even shut down after power failure; while some cabinets have low power and waste resources.

In yet another solution, the central control node in the data center collects the power of each cabinet, the power of each server, and the load of each virtual machine in the data center, and then comprehensively considers the power of each cabinet, the power of each server, and the power of each virtual machine The load determines the migration strategy of the virtual machine load. This solution needs to centrally consider the power overhead of each cabinet, each server, and each virtual machine in the data center. The amount of data calculation is large and the calculation is complex, which makes the calculation overhead of load balancing of the cabinets large, and the delay is also large. When a certain When the power of the cabinet is too high, it is difficult to reduce the power of the cabinet in time.

The embodiment of the present application provides a virtual machine management method, which is a distributed dynamic load balancing solution. The method can determine the migration strategy of the virtual machines on the servers in the single cabinet based on the power of the single cabinet and the load of the virtual machines running on the servers in the cabinet, so as to prevent the power overload of the cabinet, reduce the risk of power failure of the cabinet, and ensure cloud computing. The reliability and business continuity of the users are guaranteed. Moreover, the method decides the migration strategy of the virtual machine based on the relevant power parameters of a single cabinet, and does not need to centrally collect, store and analyze the power parameters of all or multiple cabinets in the data center, so that the overhead of dynamic load balancing is less. and low latency.

Next, in different embodiments, the virtual machine management methods provided by the embodiments of the present application are described.

FIG. 1 shows a network architecture of a data center 100 provided by an embodiment of the present application. The data center 100 may include a plurality of cabinets such as a cabinet 110 , a cabinet 120 , and a cabinet 130 . The racks in the data center 100 may also be referred to as server racks, wherein one or more servers may be deployed in a server rack. A server can act as a host on which one or more virtual machines are created or run.

Exemplarily, as shown in FIG. 1 , it can be set that a server A1 , a server A2 , and the like can be deployed in the cabinet 110 . The server A1 may create or run a virtual machine A11, a virtual machine A12, and the like. A virtual machine A21, a virtual machine A22, etc. may be created or run on the server A2. It can be set that the server B1 , the server B2 , and the like can be deployed in the rack 120 . The server B1 may create or run a virtual machine B11, a virtual machine B12, and the like. A virtual machine B21, a virtual machine B22, etc. may be created or run on the server B2. It can be set that the server C1, the server C2, and the like can be deployed in the rack 130. The server C1 may create or run a virtual machine C11, a virtual machine C12, and the like. A virtual machine C21, a virtual machine C22, and the like may be created or run on the server C2. Other cabinets of the data center 100 may also be provided with one or more servers, and each server may run one or more virtual machines, which will not be described one by one here.

Each cabinet in the data center 100 corresponds to a cabinet manager. That is, the data center 100 may be deployed with multiple rack managers, and the multiple rack managers have a one-to-one correspondence with multiple racks in the data center 110 . For example, the rack manager 111 corresponds to the rack 110 , the rack manager 121 corresponds to the rack 120 , and the rack manager 131 corresponds to the rack 130 . Exemplarily, as shown in FIG. 1 , the cabinet manager may be deployed in a corresponding cabinet. For example, rack manager 111 may be deployed in rack 110, rack manager 121 may be deployed in rack 120, rack manager 131 may be deployed in rack 130, and so on. The cabinet manager can be used to collect relevant information of the virtual machines on the server corresponding to or in the cabinet, determine whether the cabinet power of the cabinet satisfies relevant conditions, and exchange information with the virtual machine migration device of the data center 100 . The function of the cabinet manager will be specifically introduced in the following method embodiments, which will not be repeated here.

Exemplarily, as shown in FIG. 1 , the virtual machine migration apparatus of the data center 100 may include an initial scheduling component and a secondary scheduling component. The initial scheduling component may be used to create a new virtual machine; the secondary scheduling component may be used to migrate a virtual machine created on one server (eg, server A1 ) to another server (eg, server B1 ).

The function of the virtual machine migration apparatus will be specifically introduced in the following method embodiments, which will not be repeated here.

It can be understood that the virtual machines can be divided into tenant virtual machines and management virtual machines. The tenant virtual machine may be referred to as a user virtual machine or a service virtual machine, which may be used to process services submitted or arranged by the tenant or user. A management virtual machine refers to a virtual machine used to manage tenant virtual machines and other related devices or appliances.

In some embodiments, the virtual machine migration apparatus may be a management virtual machine deployed on one or more servers in the data center 100 . The one or more servers can be set in one cabinet or distributed in multiple cabinets, and the virtual machine migration device can be set in a specific cabinet, and the location is relatively fixed, and no virtual machine migration is required.

In some embodiments, in the virtual machine management method provided by the embodiment of the present application, the tenant virtual machine can be managed.

In some embodiments, in the virtual management method provided by the embodiments of the present application, the tenant virtual machine and the management virtual machine may be managed simultaneously.

Next, the virtual machine management method provided by the embodiment of the present application is illustrated by taking the cabinet manager 111 cooperating with the virtual machine migration apparatus to manage the virtual machines in the cabinet 110 as an example.

Referring to FIG. 2 , an embodiment of the present application provides a virtual machine management method based on a server cabinet, and the method is specifically as follows.

The rack manager 111 may execute step 201 to obtain the rack power W1 of the rack 110 .

Exemplarily, it can be understood that the rack power of the rack 110 is mainly generated by devices (eg, servers, network devices (eg, network switches), cooling devices, etc.) deployed in the rack 100 . The cabinet is usually provided with a baseboard management controller (BMC), which records and stores the power of each device in the cabinet, so that the cabinet manager 111 can obtain each device in the cabinet 110 from the baseboard management controller of the cabinet 110. , and then add up to obtain the cabinet power W1.

Exemplarily, the cabinet 110 may be provided with a power sensor for monitoring the power of the cabinet 110 and transmitting the monitored power to the cabinet manager 111 . Thus, the rack manager 111 can directly obtain the rack power W1 of the rack 110 .

When or after acquiring the rack power W1 of the rack 110, the rack manager 111 may execute step 202 to determine whether the rack power W1 of the rack 110 is within a preset power range.

The preset power range is a power range preset according to the capped power of the cabinet 110 . Exemplarily, the power range may include an upper limit and a lower limit, that is, the upper limit and the lower limit of the power range may be considered to be within the power range. Exemplarily, the power range may not include an upper limit or a lower limit, that is, the upper limit and the lower limit of the power range may be considered not to be within the power range. Among them, the ceiling power of the cabinet can be understood as the maximum withstand power of the cabinet (the cooling system of the cabinet, the power supply system, etc.). Usually, the ceiling power of the cabinet can be specified when the cabinet is produced or designed.

Exemplarily, the upper limit and the lower limit of the power range may be relative values of the capped power. For example, the upper limit of the power range may be seventy percent of the capped power, and the lower limit may be fifty percent of the capped power. That is to say, in this example, the preset power range may be a power interval [50% of the capped power, 70% of the capped power] or a power interval (50% of the capped power , seventy percent of the capped power).

Exemplarily, the upper limit of the power range can be obtained by subtracting the power W2 from the capped power, and the lower limit can be obtained by subtracting the power W3 from the capped power. The power W2 and the power W3 may be preset powers. For example, power W2 may be 700 watts and power W3 may be 1000 watts. That is to say, the preset power range can be a power interval [the difference obtained by subtracting 1000 watts from the capped power, and the difference obtained by subtracting 700 watts from the capped power] or a power interval (the difference obtained by subtracting 1000 watts from the capped power) value, the difference is obtained by subtracting 700 watts from the capped power).

It should be noted that, the preset power range is only exemplified in the above description, and does not constitute a limitation. In other embodiments, the preset power range may also be in other forms or other values.

When the rack power W1 of the rack 110 is within the preset power range, the rack manager 111 may perform step 203 to send a request R1 to the virtual machine migration apparatus. Exemplarily, the request R1 may include the server identification of each server in the rack 110 . Exemplarily, the server identifier of a server is information used to uniquely identify the server in the data center 100 . In the data center 110, different servers have different server identifiers, or in other words, different server identifiers correspond to different servers. In one example, the server identification may be the server name. In one example, the server identification may be a server number. Wait, I won't list them one by one here. Exemplarily, the rack manager 111 may query the baseboard manager of each server in the rack 110 for the server identifier of the server to obtain the server identifier of the server.

In some embodiments, the cabinet power W1 is specifically multiple measured powers of the cabinet 110 , and the multiple measured powers may be actually generated or actually measured powers of the cabinet 100 at multiple moments in a preset duration, wherein , a one-to-one correspondence between multiple powers and multiple times. The preset duration may be 10 minutes, 20 minutes, or the like. That is to say, the rack manager 111 executes step 203 only when the rack power of the rack 110 is within the preset power range after being monitored for many times, so as to avoid misjudgment of hot spots triggered by instantaneous power peak jitter.

In some embodiments, the rack power W1 may be the predicted power. It can be understood that the power of the cabinet has strong self-similarity and long correlation. Therefore, the cabinet 111 can adopt a prediction model, such as a differential integrated moving average autoregressive model (autoregressive integrated moving average model, ARIMA model), according to the recently acquired A plurality of measured powers are used to predict the predicted power of the cabinet 110 at a future moment. When the predicted power is within the preset power range, the rack manager 111 may execute step 203, so that the creation of a virtual machine in the rack 110 may be prohibited in advance to prevent the power of the rack 110 from further increasing.

In some embodiments, the cabinet power W1 may be N predicted powers at N times in the future (the N times and the N predicted powers have a one-to-one correspondence). N is an integer greater than 1. Each predicted power can be predicted by a prediction model. For details, please refer to the above introduction. The cabinet power W1 being within the preset power range may specifically mean that k predicted powers among the multiple predicted powers are all within the preset power range, and k is an integer greater than 1 and less than N. That is, in this embodiment, when multiple predicted powers are predicted to be within the preset power range, the rack manager 111 may execute step 203, so that the creation of virtual machines in the rack 110 may be prohibited in advance and the instantaneous peak value may be avoided Hotspot misjudgment triggered by jitter.

The request R1 sent in step 203 may be used to instruct the virtual machine migration apparatus to prohibit the creation of a virtual machine on the server corresponding to the server identifier included in the request R1. Exemplarily, as described above, the request R1 includes the server identifier of each server in the rack 110 . It is forbidden to create a virtual machine on the server corresponding to the server identifier included in the request R1 , that is, it is forbidden to create a virtual machine on each server in the rack 110 , that is, it is forbidden to create a virtual machine on all servers in the rack 110 . In other words, the creation of virtual machines in the cabinet 110 is prohibited.

In this way, when the virtual machine migration apparatus receives the request R1, it may execute step 204, and according to the request R1, it is forbidden to create a virtual machine in any server in the rack 110. That is to say, when the virtual machine migration apparatus receives the request R1 sent by the rack manager 111 , it will no longer distribute a new virtual machine to any server in the rack 110 . Therefore, when the rack power of the rack 110 is relatively high (within a preset power range), no new virtual machines are added to the rack 110 to avoid further increase of the power of the rack 110 .

Exemplarily, in step 204 , a label prohibiting the creation of virtual machines may be added to each server in the rack 110 , or a hot spot mark may be added, so as to mark each server in the rack 110 as a hot spot, thereby prohibiting subsequent transfer to the rack 110 . Dispatch or create virtual machines on each server.

Exemplarily, step 204 may be specifically performed by an initial scheduling component in the virtual machine migration apparatus.

Exemplarily, in step 204, the creation of the tenant virtual machine in the rack 110 may be prohibited according to the request R1.

Exemplarily, in step 204, the creation of the tenant virtual machine in the rack 110 and the creation of the management virtual machine in the rack 110 may be prohibited according to the request R1.

In some embodiments, continuing to refer to FIG. 2 , when the rack power W1 of the rack 110 is not within the preset power range, the rack manager 111 may further execute step 205 to determine whether the rack power W1 of the rack 110 is greater than the preset power range power range. Wherein, greater than the preset power range may specifically refer to greater than the upper limit of the preset power range. For the preset power range, reference may be made to the above description of step 202, and details are not repeated here.

In the case that the rack power W1 of the rack 110 is greater than the preset power range, the rack manager 111 may execute step 206a to send a request R2 to the virtual machine migration apparatus.

In some embodiments, the cabinet power W1 is specifically a plurality of measured powers, and the plurality of measured powers may be actually generated or actually measured powers of the cabinet 100 at a plurality of times in a preset time period, wherein the plurality of Power corresponds to multiple moments one by one. The preset duration may be 10 minutes, 20 minutes, or the like. That is, the rack manager 111 executes step 206a only when the rack power of the rack 110 is monitored multiple times greater than the preset power range, so as to avoid hot spot misjudgment triggered by instantaneous power peak jitter.

In some embodiments, the rack power W1 may be the predicted power. The predicted power can be predicted by the prediction model. For details, please refer to the above introduction. When the predicted power is greater than the preset power range, the rack manager 111 may execute step 206a, so that the virtual machine can be migrated from the rack 110 in advance to prevent the power of the rack 110 from being too high.

In some embodiments, the cabinet power W1 may be N predicted powers at N times in the future (the N times and the N predicted powers have a one-to-one correspondence). N is an integer greater than 1. Each predicted power can be predicted by a prediction model. For details, please refer to the above introduction. That the cabinet power W1 is greater than the preset power range may specifically mean that k predicted powers among the multiple predicted powers are all greater than the preset power range, and k is an integer greater than 1 and less than N. That is, in this embodiment, when multiple predicted powers are predicted to be greater than the preset power range, the rack manager 111 may execute step 206a, so that the virtual machine can be migrated from the rack 110 in advance to avoid the power consumption of the rack 110 too high; and avoid hot spot misjudgment triggered by instantaneous peak jitter.

The request R2 may be used to instruct the virtual machine migration apparatus to confirm the virtual machine to be migrated in the cabinet 110 and to migrate the confirmed virtual machine to be migrated out of the cabinet 110 . In this way, when the virtual machine migration apparatus receives the request R2, it may execute step 207a to confirm the virtual machine to be migrated in the cabinet 110 and migrate the confirmed virtual machine to be migrated out of the cabinet 110 according to the request R2. Exemplarily, step 207a may be specifically performed by a secondary scheduling component in the virtual machine migration apparatus.

In some embodiments, request R2 may include the server identification and server power of each server in rack 110 , and include the virtual machine identification and virtual machine load of the virtual machines that have been created in each server in rack 110 , and rack 110 The cabinet power W1. That is, in this example, request R2 may include server IDs and server powers for all servers in rack 110 , virtual machine IDs and virtual machine loads for all created virtual machines in rack 110 , and rack power W1 for rack 110 . The virtual machine created in the server may also be referred to as a virtual machine on the server or a virtual machine already existing on the server, which refers to a virtual machine whose host is the server.

Next, the meaning of each piece of information included in R2 will be described with an example.

For the server identifier of the server and the cabinet power W1, reference may be made to the above description, which will not be repeated here.

Exemplarily, it can be understood that for the server power, it can be composed of static power and dynamic power. Among them, static power refers to the power or overhead generated by the server without virtual machines running on it. That is, static power refers to the power or overhead of a server to keep the server running without running a virtual machine. In other words, static power is an inexorable cost of running a server, even if it's not running virtual machines. Typically, static power is fixed, or floats less. Dynamic power refers to the power that the server increases on the basis of static power when running virtual machines. That is, dynamic power is primarily the power or overhead generated by running virtual machines on a server. In other words, the dynamic power of a server may be the sum of the power generated by the operation of each of the plurality of virtual machines on the server. Wherein, the power generated by the operation of each virtual machine can be obtained from the load conversion of the virtual machine, which will be described in detail below.

Exemplarily, the virtual machine identifier of a virtual machine may be information that uniquely identifies the virtual machine in the data center 100 . In the data center 110, different virtual machines have different virtual machine identifiers, or in other words, different virtual machine identifiers correspond to different virtual machines. In one example, the virtual machine identification may be a virtual machine name. In one example, the virtual machine identification may be a virtual machine number. Wait, I won't list them one by one here.

Exemplarily, it can be understood that when a virtual machine is running, its power overhead mainly comes from the overhead of component loads such as virtual CPU, virtual memory, and virtual hard disk (or disk). The CPU utilization of the virtual machine, the memory overhead, and the total number of bytes read and written on the hard disk (or disk) can be collectively referred to as the virtual machine load of the virtual machine. That is, the virtual machine load may include one or more of CPU utilization, memory overhead, total number of bytes read and written to the hard disk (or disk), and the like. In an example, the memory overhead may specifically refer to the number of missing times of the last level cache (last level cache, LLC) of the virtual machine due to context switching. In one example, the total number of bytes read and written to the hard disk (or disk) may refer to the incoming and outgoing traffic of the network card of the virtual machine.

The meaning of each information included in R2 is described above. Next, we describe how to obtain each piece of information.

In some embodiments, before the rack manager 111 performs step 206a, the rack manager 111 may further perform: query each server in the rack 110 for the server identification and server power, and query each server in the rack 110 that has been created The virtual machine ID and virtual machine load of the virtual machine.

Exemplarily, the rack manager 111 may query the baseboard manager of each server in the rack 110 for the server identifier of the server.

Exemplarily, the rack manager 111 may query the baseboard management controller of each server in the rack 110 for the server power of the server to obtain the server power of the server. Exemplarily, each server may be provided with a power sensor. The power sensor may monitor the server power of the server and communicate the monitored server power to the rack manager 111 . Thus, the rack manager 111 can obtain server power for each server in the rack 110 .

Exemplarily, the server power included in the request R2 consists of static power and dynamic power.

Exemplarily, the server power included in the request R2 may specifically be the dynamic power of the server. It can be understood that the static power of the server is relatively fixed, and the static power of each server in the rack 110 can be preset in the rack manager 111 . When the server power of a server is inquired, the rack manager 111 may subtract the static power of the server from the server power of the server to obtain the dynamic power of the server.

Exemplarily, the cabinet manager 111 may query the load (CPU utilization, memory overhead, hard disk (or disk) read and write total bytes) of the virtual machine from the log file generated by each virtual machine on each server. number, etc.) and the virtual machine ID.

Exemplarily, each server can be provided with a virtual machine information collector, which can be used to collect dynamic information of each virtual machine on the server (for example, CPU utilization, memory overhead, total number of bytes read and written on the hard disk (or disk), etc. load information), and store the collected dynamic information in the local database (database, DB) of the server for the cabinet manager 111 to query. Thus, the rack manager 111 can query the load and the virtual machine identifiers of the virtual machines on the server from the local database of the server.

Through the above solution, the rack manager 111 can query to obtain the identifier and server power of each server in the rack 110 , as well as the virtual machine identifier and virtual machine load of the virtual machine that has been created in each server in the rack 110 , and use these information It is carried in the request R2 and sent to the virtual machine migration apparatus.

Continuing to refer to FIG. 2 , the virtual machine migration apparatus may execute step 207 a to confirm the virtual machine to be migrated in the cabinet 110 and migrate the to-be-migrated virtual machine out of the cabinet 110 according to the request R2 .

Specifically, in step 207a, according to the server identifier and server power of each server in the rack 110 included in the request R2, the virtual machine identifier and virtual machine load of the created virtual machine in each server, and The rack power W1 of the rack 110 determines the virtual machine to be migrated.

Exemplarily, in step 207a, the virtual machine load of the virtual machines created in each server in the rack 110 may be converted into virtual machine power. Exemplarily, each of the at least two servers may be determined according to the server power of the at least two servers in the cabinet 110 and the virtual machine load of each virtual machine on each of the two servers in the instruction. The virtual machine power of the virtual machine. Specifically, the conversion coefficient between the virtual machine load and the virtual machine power can be determined according to the server power of at least two servers in the cabinet 110 and the virtual machine load of each virtual machine on each of the two servers in the instruction, Then, according to the conversion system and the virtual machine load of any virtual machine, the virtual machine power of the virtual machine is determined. The virtual machine power of the virtual machine may be simply referred to as the power of the virtual machine.

In one example, a load power conversion model may be constructed, which may consist of the following equations (1), (2).

Among them, P _static represents the static power of the server;

Represents the dynamic power of the server at time t; M represents the set of virtual machines on the server; i represents an element in M, that is, i represents a virtual machine on the server.

represents the power generated by virtual machine i at time t; F ^it represents the load of virtual machine i at time t; α and β represent the weight and skew of load-power conversion, respectively. Among them, the weight and skew of the load-power conversion can be collectively referred to as the conversion coefficient between the virtual machine load and the virtual machine power.

The request R2 includes the server power of each server in the cabinet 110 (when the server power is composed of static power and dynamic power, the virtual machine migration apparatus can obtain the dynamic power through formula (1), where the static power can be preset at In the virtual machine migration device; when the server power is dynamic power, the virtual machine migration device can directly obtain the dynamic power) and the virtual machine load of each virtual machine on each server. The values of α and β can be calculated by using the server power of the at least two servers in the cabinet 110 and the virtual machine load of the virtual machines on the at least two servers and using formulas (1) and (2). Thus, the virtual machine power of the virtual machine can be calculated by using the known values of α and β and the virtual machine load of each virtual machine, so as to realize the conversion from load to power.

In a specific example, it can be set that the virtual machine load is represented by CPU utilization, memory overhead, and the total number of bytes read and written on the hard disk (or disk), and the load power conversion model can be composed of the following formula.

in,

represents the power of virtual machine i at time t;

represents the power of the CPU of virtual machine i at time t;

represents the power of the memory of the virtual machine i at time t;

Represents the memory overhead of virtual machine i at time t (specifically, it can be the number of misses of LLC); α ₂ , β ₂ represent the weight and skew of memory overhead-power conversion, respectively;

represents the power of the hard disk (or disk) of the virtual machine i at time t;

Indicates the total number of bytes read and written on the hard disk (or disk) of virtual machine i at time t, where,

Indicates the number of bytes written to the hard disk (or disk) of virtual machine i at time t (specifically, it can be the incoming traffic of the network card),

Represents the number of bytes read by the hard disk (or disk) of virtual machine i at time t (specifically, it can be the outgoing traffic of the network card); α ₃ and β ₃ respectively represent the total number of bytes read and written by the hard disk (or disk) - the weight of power conversion and skew.

Using the server power of the at least three servers in the cabinet 110 and the virtual machine load of the virtual machines on the at least three servers, and using formulas (1), (2a), (2b), (2c), (2d), the calculation Get the values of α ₁ , β ₁ , α ₂ , β ₂ , α ₃ , and β ₃ . Therefore, the known values of α ₁ , β ₁ , α ₂ , β ₂ , α ₃ , β ₃ and the virtual machine load of each virtual machine can be used to calculate the virtual machine power of the virtual machine to achieve the load conversion to power.

In some embodiments, according to the above-mentioned formulas (1), (2a), (2b), (2c), (2d) according to the experimental value of the cabinet power, the experimental value of the server power, and the experimental value of the virtual machine load in advance , and determine the values of α ₁ , β ₁ , α ₂ , β ₂ , α ₃ , and β ₃ . The request R2 sent in step 206a may include the rack power W1, the virtual machine identifier and the virtual machine load of each virtual machine that has been created in each server in the rack 110. In step 207a, the power of the virtual machine may be determined according to the load of each virtual machine using the above formulas (1), (2a), (2b), (2c), and (2d). Exemplarily, before the rack manager 111 performs step 206a, the rack manager 111 may further perform: query each server in the query rack 110 for the virtual machine identifier and virtual machine load of the created virtual machine. For details, please refer to the above introduction, which will not be repeated here.

In the above manner, the virtual machine load can be converted into virtual machine power, so that the virtual machine power to be migrated from the cabinet 110 can be determined according to the virtual machine power of each virtual machine in the cabinet 110 and the cabinet power W1 of the cabinet 110 In order to reduce the rack power of the rack 110 by reducing the number of virtual machines in the rack 110 . The virtual machine to be migrated from the cabinet 110 may also be referred to as the virtual machine to be migrated from the cabinet 110 .

Next, an example presents a scenario for determining the virtual machine to migrate.

In some embodiments, the virtual machine migration apparatus may pre-record the target power of the rack 110 . The target power is the target power of the cabinet power adjustment. In this embodiment of the present application, the purpose of determining the virtual machine to be migrated in the rack 110 and migrating the virtual machine to be migrated from the rack 110 in subsequent steps is to reduce the rack power of the rack 110 to below the target power. Exemplarily, the target power may be a preset value, which is less than or equal to the lower limit of the preset power range described above.

In step 207a, the virtual machine migration apparatus may subtract the target power from the cabinet power W1 to obtain a power difference value E1; and then determine the virtual machine to be migrated among the virtual machines on all servers in the cabinet 110 according to the power difference value E1. Specifically, it may be determined from the virtual machines in the cabinet 110 that the virtual machines whose virtual machine power is not less than the power difference value E1, or a plurality of virtual machines whose power sum is not less than the power difference value E1 can be determined, The virtual machine or the multiple virtual machines whose power of the virtual machine is not less than the power difference value E1 is determined as the virtual machine to be migrated. In one example, at least two virtual machines whose power sum is closest to and not less than the power difference E1 can be selected as the virtual machines to be migrated; or one virtual machine whose power is the closest and not less than the power difference E1 can be selected as the virtual machine to be migrated. The virtual machine to be migrated.

Next, it demonstrates concretely.

In some embodiments, when a single virtual machine whose virtual machine power is not less than the power difference value E1 exists among the virtual machines in the cabinet 110 , that is, when the virtual machine power of one virtual machine in the cabinet 110 is not less than the power difference value E1 When the value is E1, the virtual machine can be determined as the virtual machine to be migrated.

In some embodiments, when the virtual machine power of each of the plurality of virtual machines in the cabinet 110 is not less than the power difference E1, the virtual machine with the smallest memory overhead among the plurality of virtual machines may be determined as the waiting Migrate virtual machines. Thereby, migration energy consumption can be saved.

In some embodiments, when there is no single virtual machine whose virtual machine power is not less than the power difference E1 in the rack 110, that is, when the virtual machine power of each virtual machine in the rack 110 is less than the power difference E1 At the time, it may be determined from the virtual machines in the cabinet 110 that a plurality of virtual machines whose power summation of the virtual machines is not less than the power difference value E1 are virtual machines to be migrated.

It can be understood that there may be multiple virtual machine combinations in the cabinet 110 whose power sum of the virtual machines is not less than the power difference value E1. That is, there may be a plurality of virtual machine combinations in the cabinet 110, wherein the sum of the virtual machine powers of the virtual machines in each virtual machine combination is not less than the power difference value E1. In one example, from the multiple virtual machine combinations, it may be determined that the combination containing the least number of virtual machines is the virtual machine to be migrated. In one example, from the multiple virtual machine combinations, a combination with the smallest sum of memory overheads may be determined as the virtual machine to be migrated.

In some embodiments, the virtual machine to be migrated may be determined by determining the model of the virtual machine to be migrated. The determination model of the virtual machine to be migrated may be composed of formulas (3), (4), and (5).

sum(P _VM1 , P _VM2 , ..., P _VMn ) ≥ power difference E1 (3)

Min(sum(mem(VM1), mem(VM2), ..., mem(VMn))) (4)

Among them, P _cabinet represents the cabinet power of the cabinet 110 ; P _VM1 , P _VM2 , _. Virtual machine power; mem(VM1), mem(VM2), ..., mem(VMn) respectively represent the memory overhead of virtual machine 1, the memory overhead of virtual machine 2, ..., the memory overhead of virtual machine n in the cabinet 110.

Formula (3) indicates that the virtual machine power of the virtual machine to be migrated should not be less than the power difference value E1. Formula (3) may have multiple solutions, each solution represents one or more virtual machines, and the virtual power sum of the one or more virtual machines is not less than the power difference value E1. The formula (4) indicates that the solution with the smallest sum of memory overheads is determined from the multiple solutions of the formula (3) as the virtual machine to be migrated. Therefore, the virtual machine memory that needs to be copied during virtual machine migration can be minimized.

In step 207a, when or after the virtual machine to be migrated is determined, the virtual machine migration apparatus can migrate the virtual machine to be migrated out of the cabinet 110, for example, can be migrated to another cabinet. In an example, returning to FIG. 1 , the virtual machine A11 and the virtual machine A21 in the rack 110 may be set as the virtual machines to be migrated, and the rack 120 is not marked as a hotspot. The virtual machine migration apparatus may migrate the virtual machine A11 and the virtual machine A21 to the servers in the cabinet 120 .

In some embodiments, when the rack power W1 of the rack 110 is greater than the preset power range, the rack manager 111 may determine the virtual machine to be migrated in the rack 110 . Exemplarily, the above formulas (1), (2a), (2b), (2c), (2d), determine the power of each virtual machine in the rack 110. Exemplarily, the above formulas (1), (2a), (2b), (2c), (2d) and predetermined α ₁ , β ₁ can be used according to the virtual machine identifier and the virtual machine load of each virtual machine. The values of , α ₂ , β ₂ , α ₃ , and β ₃ determine the power of each virtual machine in the rack 110 . For details, reference may be made to the above description of steps 206a and 207a, which will not be repeated here.

When or after determining the virtual machine to be migrated, the cabinet manager 111 may execute step 206b to send a virtual machine migration request to the virtual machine migration apparatus. The virtual machine migration request may include a virtual machine identifier of the virtual machine to be migrated. The virtual machine migration apparatus may execute step 207b to migrate the virtual machine to be migrated out of the cabinet 110 according to the virtual machine migration request. For details, reference may be made to the above description of step 207a, which will not be repeated here.

In some embodiments, when the rack power W1 of the rack 110 is greater than the preset power range, the rack manager 110 may also perform step 203 , and then the virtual machine migration apparatus may perform step 204 .

In some embodiments, after the rack manager 111 acquires the rack power W1 of the rack, the rack manager 111 may further perform step 208 to acquire the rack power W2 of the rack 110 . For the method of obtaining the cabinet power W2, please refer to the above description of the obtaining method of the cabinet power W1.

After acquiring the rack power W2, the rack manager 111 may further execute step 209 to determine whether the rack power W2 is less than a preset power range. The fact that the cabinet power W2 is less than the preset power range specifically means that the cabinet power W2 is less than the lower limit of the preset power range. For the preset power range, reference may be made to the above description of step 202, which will not be repeated here.

It can be understood that through steps 206a, 207a, or steps 206b, 207b, one or more virtual machines in the cabinet 110 are migrated out. As a result, the power overhead of the rack 110 may be reduced. Moreover, as the running state of the virtual machines in the cabinet 110 changes, for example, some virtual machines no longer process service data or process service data is reduced, the power consumption of the cabinet 110 may be further reduced. Therefore, it may happen that the cabinet power W2 is smaller than the preset power range.

In the case that the rack power W2 is less than the preset power range, the rack manager 111 may execute step 210 to send a request R3 to the virtual machine migration apparatus. Request R3 may include a server identification for each server in rack 110 . The request R3 is used to instruct the virtual machine migration apparatus to allow the creation of a virtual machine in any server in the rack 110 .

Upon receiving the request R3, the virtual machine migration apparatus may execute step 211 to allow the virtual machine to be created in the cabinet 110 according to the request R3. That is, upon or after receiving the request R3, the virtual machine migration apparatus (specifically, the initial scheduling component and/or the secondary scheduling component) may create a virtual machine in the server in the cabinet 110 .

Exemplarily, from the above, in step 204, the virtual machine migration apparatus labels each server in the rack 110 as a hotspot. The request R3 can be understood as a hotspot clearing message, which can cause the virtual machine machine migration apparatus to clear the hotspot tag of the server in the cabinet 110 or prohibit the creation of a virtual machine tag.

Therefore, with the virtual machine management method provided by the embodiment of the present application, when the power of the cabinet is high, no virtual machine is created in the cabinet, so that the power of the cabinet is not increased as much as possible; and the power of the cabinet is sufficient When the power is high, one or more virtual machines in the cabinet are migrated out to reduce the power of the cabinet; when the power of the cabinet is low, the virtual machine migration device can be instructed to distribute virtual machines to it to increase the resources in the cabinet. utilization. Therefore, the risk of power failure and shutdown of the cabinet due to overload can be avoided or reduced without reducing the resource utilization rate of the cabinet.

Referring to FIG. 3 , an embodiment of the present application provides a virtual machine management method, and the method is as follows.

The rack manager 111 may execute step 301 to obtain the rack power W1 of the rack 110 . For details, reference may be made to the above description of step 201 in FIG. 2 for implementation, which will not be repeated here.

The rack manager 111 may execute step 302 to determine whether the rack power W1 is greater than a preset power range. For details, reference may be made to the above description of step 205 in FIG. 2 for implementation, which will not be repeated here.

In the case that the cabinet power W1 is greater than the preset range of power, the rack manager 111 may perform step 303 to send a request R2 to the virtual machine migration apparatus. The request R2 is used to instruct the virtual machine migration apparatus to confirm the virtual machine to be migrated in the cabinet 110 and migrate the virtual machine to be migrated out of the cabinet 110 . For details, reference may be made to the above description of step 206a in FIG. 2 for implementation.

In some embodiments, request R2 may include the server identification and server power of each server in rack 110, the virtual machine identification and virtual machine load of the virtual machines that have been created in each server in rack 110, and rack power W1. For details, reference may be made to the above description of step 206a in FIG. 2 for implementation.

In some embodiments, before executing step 303, the rack manager 111 may query each server in the rack 110 for the created virtual machine identifier and virtual machine load. For details, reference may be made to the above description of step 206a in FIG. 2 for implementation.

When or after receiving the request R2, the virtual machine migration apparatus may execute step 304 to confirm the virtual machine to be migrated in the cabinet 110 and migrate the virtual machine to be migrated out of the cabinet 110 according to the request R2. For details, reference may be made to the above description of step 207a in FIG. 2 , which will not be repeated here.

In some embodiments, as shown in FIG. 3 , when the rack power W1 is not greater than the preset power range, the rack manager 111 may further execute 305 to determine whether the rack power W1 is within the preset power range. For details, reference may be made to the above description of step 202 in FIG. 2 , which will not be repeated here.

When the rack power W1 is within the preset power range, the rack manager 111 may execute step 306 to send a request R1 to the virtual machine migration apparatus. Exemplarily, the request R1 may include the server identification of each server in the rack 110 . For details, refer to the above description of step 203 in FIG. 2 .

When or after receiving the request R1, the virtual machine migration apparatus may execute step 307, and according to the request R1, prohibit the creation of a virtual machine in any server of the rack 110. For details, reference may be made to the above description of step 204 in FIG. 2 , which will not be repeated here.

In some embodiments, as shown in FIG. 3 , after performing step 301, the rack manager 111 may further perform step 308 to obtain the rack power W2 of the rack 110; and perform step 309 to determine whether the rack power W2 is less than the preset power Scope. For details, reference may be made to the above description of step 208 and step 209 in FIG. 2 .

When the rack power W2 is less than the preset power range, the rack manager 111 may perform step 310 to send a request R3 to the virtual machine migration apparatus, where the request R3 may include the server identifier of each server in the rack 110 . The request R3 is used to instruct the virtual machine migration apparatus to allow the creation of virtual machines in the rack 110 . For details, reference may be made to the above description of step 210 in FIG. 2 .

When or after receiving the request R3, the virtual machine migration apparatus may execute step 311 to allow the virtual machine to be created in the cabinet 110 according to the request R3. For details, reference may be made to the above description of step 211 in FIG. 2 , which will not be repeated here.

Therefore, with the virtual machine management method provided by the embodiment of the present application, when the power of the cabinet is high enough, one or more virtual machines in the cabinet can be migrated out to reduce the power of the cabinet; the power of the cabinet can be high. When the power of the cabinet is low, no more virtual machines will be created in the cabinet, so that the power of the cabinet will not increase as much as possible; and when the power of the cabinet is low, the virtual machine migration device can be instructed to distribute virtual machines to it, so as to increase the resources in the cabinet. utilization. Therefore, the risk of power failure and shutdown of the cabinet due to overload can be avoided or reduced without reducing the resource utilization rate of the cabinet.

Referring to FIG. 4 , an embodiment of the present application provides a virtual machine management method, and the method is as follows.

The rack manager 111 may execute step 401 to obtain the rack power W1 of the rack 110 . For details, reference may be made to the above description of step 201 in FIG. 2 for implementation, which will not be repeated here.

The rack manager 111 may execute step 402 to determine whether the rack power W1 is greater than a preset power range. For details, reference may be made to the above description of step 205 in FIG. 2 for implementation, which will not be repeated here.

In the case that the cabinet power W1 is greater than the preset range of power, the cabinet manager 111 may perform step 403 to determine the virtual machine to be migrated in the cabinet 110 . Exemplarily, the rack manager 111 may determine, according to the server power of the at least two servers in the rack 110 and the load of each virtual machine on the at least two servers, the power of the first virtual machine on the first server of the at least two servers. power. With reference to the way of determining the power of the first virtual machine, the power of other virtual machines in the cabinet 110 may be determined. Then, the virtual machines to be migrated in the rack 110 are determined according to the rack power W1 and the power of each virtual machine in the rack 110 . For details, reference may be made to the above description of step 207a in FIG. 2 , which will not be repeated here.

When or after determining the virtual machine to be migrated, the cabinet manager 111 may execute step 404 to send a virtual machine migration request to the virtual machine migration apparatus, where the virtual machine migration request includes the virtual machine identifier of the virtual machine to be migrated. For details, reference may be made to the above description of step 206b in FIG. 2 , which will not be repeated here.

When or after receiving the virtual machine migration request, the virtual machine migration apparatus may execute step 405, and may migrate the virtual machine to be migrated out of the cabinet 110 according to the virtual machine migration request. For details, reference may be made to the above description of step 207b in FIG. 2 , which will not be repeated here.

In some embodiments, as shown in FIG. 4 , when the rack power W1 is not greater than the preset power range, the rack manager 111 may further execute 406 to determine whether the rack power W1 is within the preset power range. For details, reference may be made to the above description of step 202 in FIG. 2 , which will not be repeated here.

When the rack power W1 is within the preset power range, the rack manager 111 may execute step 407 to send a request R1 to the virtual machine migration apparatus. Exemplarily, the request R1 may include the server identification of each server in the rack 110 . For details, refer to the above description of step 203 in FIG. 2 .

When or after receiving the request R1, the virtual machine migration apparatus may execute step 408 to prohibit the creation of a virtual machine in any server of the rack 110 according to the request R1. For details, reference may be made to the above description of step 204 in FIG. 2 , which will not be repeated here.

In some embodiments, as shown in FIG. 4 , after performing step 401, the rack manager 111 may further perform step 409 to obtain the rack power W2 of the rack 110; and perform step 410 to determine whether the rack power W2 is less than the preset power Scope. For details, reference may be made to the above description of step 208 and step 209 in FIG. 2 .

When the rack power W2 is less than the preset power range, the rack manager 111 may perform step 411 to send a request R3 to the virtual machine migration apparatus, where the request R3 may include the server identifier of each server in the rack 110 . The request R3 is used to instruct the virtual machine migration apparatus to allow the creation of virtual machines in the rack 110 . For details, reference may be made to the above description of step 210 in FIG. 2 .

When or after receiving the request R3, the virtual machine migration apparatus may execute step 412 to allow the virtual machine to be created in the cabinet 110 according to the request R3. For details, reference may be made to the above description of step 211 in FIG. 2 , which will not be repeated here.

It should be noted that, the virtual machine management method provided by the embodiment of the present application is described above by taking the virtual machine management in the cabinet 110 as an example. Other cabinets in the data center 100 may refer to the virtual machine management mode of the cabinet 110 to perform virtual machine management. This manual will not repeat them one by one.

Referring to FIG. 5 , an embodiment of the present application provides a cabinet manager, and the cabinet manager can be configured in a server cabinet, such as the cabinet 110 shown in FIG. 1 . As shown in Figure 5, the cabinet manager may include:

an obtaining unit 510, configured to obtain the first cabinet power of the server cabinet;

The sending unit 520 is configured to send a first request to the virtual machine migration apparatus when the power of the first cabinet is within a preset power range, where the first request is used to instruct the virtual machine migration apparatus to prohibit the Create a virtual machine on any server in the server rack.

The functions of each functional unit of the cabinet manager provided in this embodiment of the present application may be implemented with reference to the above method embodiments shown in FIG. 2 , and details are not described herein again.

With the cabinet manager provided in the embodiment of the present application, when the power of the cabinet is high, no virtual machine is created in the cabinet, so that the power of the cabinet is not increased as much as possible.

Continuing to refer to FIG. 5 , an embodiment of the present application provides a rack manager, and the rack manager may be configured in a server rack, such as the rack 110 shown in FIG. 1 . As shown in Figure 5, the cabinet manager may include:

The sending unit 520 is configured to send a second request to the virtual machine migration apparatus when the power of the first cabinet is greater than a preset power range, where the second request is used to instruct the virtual machine migration apparatus to confirm the server The virtual machine to be migrated in the cabinet and the virtual machine to be migrated out of the server cabinet.

The functions of each functional unit of the cabinet manager provided in this embodiment of the present application may be implemented with reference to the above method embodiments shown in FIG. 3 , and details are not described herein again.

With the cabinet manager provided in the embodiment of the present application, when the power of the cabinet is high enough, one or more virtual machines in the cabinet can be migrated out, so as to reduce the power of the cabinet.

The embodiment of the present application provides a cabinet manager (not shown), and the cabinet manager can be configured in a server cabinet, such as the cabinet 110 shown in FIG. 1 . Enclosure Manager includes:

an obtaining unit for obtaining the first cabinet power of the server cabinet;

a determining unit, configured to determine the virtual machine to be migrated in the server cabinet when the power of the first cabinet is greater than the preset power range;

The sending unit is configured to send a virtual machine migration request to the virtual machine migration apparatus, where the virtual machine migration request is used to instruct the virtual machine migration apparatus to migrate the virtual machine to be migrated out of the server cabinet.

The functions of each functional unit of the cabinet manager provided in this embodiment of the present application may be implemented with reference to the above method embodiments shown in FIG. 4 , and details are not described herein again.

Referring to FIG. 6, an embodiment of the present application provides a virtual machine migration apparatus, including:

an obtaining unit 610, configured to obtain a first request sent by the cabinet manager, where the first request includes the server identifier of each server set in the server cabinet where the cabinet manager is located;

The prohibiting unit 620 is configured to prohibit the creation of a virtual machine in any server of the server rack according to the first request.

The functions of each functional unit of the virtual machine migration apparatus provided in the embodiment of the present application may be implemented with reference to the above method embodiments shown in FIG. 2 , and details are not described herein again.

With the virtual machine migration apparatus provided by the embodiment of the present application, when the power of the cabinet is high, no virtual machine is created in the cabinet, so that the power of the cabinet is not increased as much as possible.

Referring to FIG. 7, an embodiment of the present application provides a virtual machine migration apparatus, including:

Obtaining unit 710, configured to obtain a second request sent by the cabinet manager, the second request requesting the virtual machine identifier and virtual machine load of each virtual machine that has been created in each server in the server cabinet, and the server cabinet The power of the first cabinet;

A determining unit 720, configured to carry the virtual machine identifier and virtual machine load of each virtual machine that has been created in each server in the server rack, and the first rack power of the server rack carried according to the second request determining the virtual machine to be migrated in the server cabinet;

The migration unit 730 is configured to migrate the virtual machine to be migrated out of the server cabinet.

The functions of each functional unit of the virtual machine migration apparatus provided in the embodiment of the present application may be implemented with reference to the above method embodiments shown in FIG. 3 , and details are not described herein again.

With the virtual machine migration apparatus provided in the embodiment of the present application, when the power of the cabinet is high enough, one or more virtual machines in the cabinet can be migrated out, so as to reduce the power of the cabinet.

An embodiment of the present application provides a virtual machine migration apparatus (not shown), including:

an obtaining unit, configured to obtain a virtual machine migration request sent by the cabinet manager, where the virtual machine migration request includes a virtual machine identifier of the virtual machine to be migrated set in the server cabinet where the cabinet manager is located;

The migration unit is configured to migrate the virtual machine to be migrated out of the server cabinet according to the virtual machine migration request.

The functions of each functional unit of the virtual machine migration apparatus provided in this embodiment of the present application may be implemented with reference to the above method embodiments shown in FIG. 4 , and details are not described herein again.

Referring to FIG. 8 , an embodiment of the present application provides a cabinet manager. The cabinet manager includes a processor 810 , a memory 820 , and a transceiver 830 . Memory 820 stores instructions that are executable by processor 810 . When the instruction is executed by the processor 810, the cabinet manager may perform the operations performed by the cabinet manager 111 in the above method embodiments. Specifically, the processor 810 may perform data processing operations, and the transceiver 830 may perform data transmission and/or reception operations.

The specific implementation of each component/device of the cabinet manager in this embodiment of the present application may be implemented with reference to each method embodiment shown in FIG. 2 or FIG. 3 or FIG. 4 , and details are not described herein again.

The cabinet manager provided in this embodiment of the present application can migrate one or more virtual machines out of the cabinet when the power of the cabinet is high enough to reduce the power of the cabinet; or, when the power of the cabinet is high, no longer Create virtual machines in enclosures so that the power of the enclosures does not increase as much as possible.

Referring to FIG. 9 , an embodiment of the present application provides a virtual machine migration apparatus. The virtual machine migration apparatus includes a processor 910 , a memory 920 , and a transceiver 930 . Memory 920 stores instructions that are executable by processor 910 . When the instruction is executed by the processor 910, the virtual machine migration apparatus may perform the operations performed by the virtual machine migration apparatus in the foregoing method embodiments. Specifically, the processor 910 may perform data processing operations, and the transceiver 930 may perform data transmission and/or reception operations.

The specific implementation of each component/device of the virtual machine migration apparatus according to the embodiment of the present application may be implemented with reference to each method embodiment shown in FIG. 2 or FIG. 3 or FIG. 4 , and details are not repeated here.

The virtual machine migration apparatus provided by the embodiment of the present application can migrate one or more virtual machines out of the cabinet when the power of the cabinet is high enough to reduce the power of the cabinet; or, when the power of the cabinet is high, it can not Then create virtual machines in the cabinet so that the power of the cabinet will not increase as much as possible.

It can be understood that, the processor in the embodiment of the present application may be a central processing unit (central processing unit, CPU), and may also be other general-purpose processors, digital signal processors (digital signal processors, DSP), application-specific integrated circuits ( application specific integrated circuit, ASIC), field programmable gate array (FPGA) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. A general-purpose processor may be a microprocessor or any conventional processor.

The method steps in the embodiments of the present application may be implemented in a hardware manner, or may be implemented in a manner in which a processor executes software instructions. Software instructions can be composed of corresponding software modules, and software modules can be stored in random access memory (RAM), flash memory, read-only memory (ROM), programmable read-only memory (programmable rom) , PROM), erasable programmable read-only memory (erasable PROM, EPROM), electrically erasable programmable read-only memory (electrically EPROM, EEPROM), registers, hard disks, removable hard disks, CD-ROMs or known in the art in any other form of storage medium. An exemplary storage medium is coupled to the processor, such that the processor can read information from, and write information to, the storage medium. Of course, the storage medium can also be an integral part of the processor. The processor and storage medium may reside in an ASIC.

In the above-mentioned embodiments, it may be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented in software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present application are generated. The computer may be a general purpose computer, special purpose computer, computer network, or other programmable device. The computer instructions may be stored in or transmitted over a computer-readable storage medium. The computer instructions can be sent from one website site, computer, server, or data center to another website site by wire (eg, coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (eg, infrared, wireless, microwave, etc.) , computer, server or data center. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, a data center, or the like that includes an integration of one or more available media. The usable media may be magnetic media (eg, floppy disks, hard disks, magnetic tapes), optical media (eg, DVDs), or semiconductor media (eg, solid state disks (SSDs)), and the like.

It can be understood that, the various numbers and numbers involved in the embodiments of the present application are only for the convenience of description, and are not used to limit the scope of the embodiments of the present application.

Claims

A virtual machine management method based on a server cabinet, characterized in that it is applied to a cabinet manager of a server cabinet, and the method includes:

obtaining the first cabinet power of the server cabinet;

When the power of the first rack is within a preset power range, send a first request to the virtual machine migration apparatus, where the first request is used to instruct the virtual machine migration apparatus to prohibit the creation of a virtual machine in the server rack .
The method according to claim 1, wherein the first request includes a server identifier of each server set in the server cabinet, and the first request is used to instruct the virtual machine migration apparatus to prohibit the Create a virtual machine on any server in the server rack.
The method according to claim 1 or 2, wherein the method comprises:

In the case that the power of the first rack is greater than the preset power range, send a second request to the virtual machine migration apparatus, where the second request is used to instruct the virtual machine migration apparatus to confirm the waiting list in the server rack A virtual machine is migrated and the to-be-migrated virtual machine is migrated out of the server cabinet.
The method according to claim 3, wherein the second request comprises the power of the first rack, the virtual machine identifier of each virtual machine that has been created in each server in the server rack, and the virtual machine load.
The method according to claim 4, wherein before the sending the second request to the virtual machine migration apparatus, the method further comprises:

Each server in the server rack is queried for the virtual machine identifier and virtual machine load of each virtual machine that has been created.
The method according to claim 1 or 2, wherein the method further comprises:

In the case that the power of the first cabinet is greater than the preset power range, determine the virtual machine to be migrated in the server cabinet;

Sending a virtual machine migration request to the virtual machine migration apparatus, where the virtual machine migration request is used to instruct the virtual machine migration apparatus to migrate the to-be-migrated virtual machine out of the server cabinet.
The method according to claim 6, wherein the method further comprises:

Pre-record target power;

Then, the determining of the virtual machine to be migrated in the server cabinet includes:

Determine the power of each virtual machine that has been created on each server according to the load of each virtual machine that has been created on each server in the server cabinet;

According to the target power, the power of the first rack, and the power of each virtual machine in the server rack, the virtual machine to be migrated is selected from the virtual machines that have been created in the server rack.
The method according to claim 7, wherein, according to the target power, the power of the first rack, and the power of each virtual machine in the server rack, the virtual machine created from the server rack is Select the virtual machine to be migrated from the machine, including:

obtaining the difference between the power of the first cabinet and the target power;

Select at least two virtual machines whose sum of power is closest to and not less than the difference as the virtual machines to be migrated; or

A virtual machine whose power is closest to and not less than the difference is selected as the to-be-migrated virtual machine.
The method according to claim 7 or 8, wherein each virtual machine that has been created on each server is determined according to the load of each virtual machine that has been created on each server in the server rack machine power, including:

in,
represents the power of virtual machine i at time t;
Represents the power of the CPU of virtual machine i at time t;
represents the CPU utilization of virtual machine i at time t; α 1 and β 1 represent the weight and skew of CPU utilization-power conversion, respectively;
represents the power of the memory of the virtual machine i at time t;
represents the memory overhead of virtual machine i at time t; α 2 and β 2 represent the weight and skew of memory overhead-power conversion, respectively;
represents the power consumption of the hard disk of virtual machine i at time t;
Indicates the total number of bytes read and written on the hard disk of virtual machine i at time t, where,
Indicates the number of bytes written to the hard disk of virtual machine i at time t,
represents the number of hard disk read bytes of virtual machine i at time t; α 3 and β 3 represent the total number of hard disk read and write bytes - the weight and skew of power conversion, respectively.
The method according to claim 1, wherein after obtaining the first rack power of the server rack, the method further comprises:

obtaining the power of the second cabinet of the server cabinet;

In the case that the power of the second cabinet is less than the preset power range, send a third request including the server identifier of the server set in the server cabinet to the virtual machine migration apparatus, where the third request is used to indicate The virtual machine migration apparatus allows virtual machines to be created in the server rack.
The method according to claim 1, wherein the first cabinet power is measured power or predicted power.
A virtual machine management method based on a server cabinet, characterized in that it is applied to a cabinet manager of a server cabinet, and the method includes:

obtaining the first cabinet power of the server cabinet;

When the power of the first rack is greater than a preset power range, send a second request to the virtual machine migration apparatus, where the second request is used to instruct the virtual machine migration apparatus to confirm the virtual machine to be migrated in the server rack machine and migrate the virtual machine to be migrated out of the server cabinet.
The method of claim 12, wherein the second request includes the power of the first rack, a virtual machine identifier of each virtual machine that has been created in each server in the server rack, and a virtual machine load.
The method according to claim 13, wherein before the sending the second request to the virtual machine migration apparatus, the method further comprises:

Each server in the server rack is queried for the virtual machine identifier and virtual machine load of each virtual machine that has been created.
The method according to any one of claims 12 to 14, wherein after acquiring the first rack power of the server rack, the method further comprises:

obtaining the power of the second cabinet of the server cabinet;

In the case that the power of the second cabinet is less than the preset power range, send a third request including the server identifier of the server set in the server cabinet to the virtual machine migration apparatus, where the third request is used to indicate The virtual machine migration apparatus allows virtual machines to be created in the server rack.
A virtual machine management method based on a server cabinet, characterized in that it is applied to a virtual machine migration apparatus, and the method includes:

obtaining a first request sent by the cabinet manager, where the first request includes the server identifier of each server set in the server cabinet where the cabinet manager is located;

The creation of a virtual machine in any server of the server rack is prohibited according to the first request.
A virtual machine management method based on a server cabinet, characterized in that it is applied to a virtual machine migration apparatus, and the method includes:

Obtain a second request sent by the rack manager, where the second request includes the virtual machine identifier and virtual machine load of each virtual machine that has been created in each server in the server rack, and the first rack power of the server rack ;

According to the virtual machine identifier and virtual machine load of each virtual machine that has been created in each server in the server rack, and the first rack power of the server rack carried by the second request The virtual machine to be migrated;

Migrate the virtual machine to be migrated out of the server cabinet.
The method of claim 17, wherein the method further comprises:

Pre-recording the target power of the server power;

Then, according to the virtual machine identifier and virtual machine load of each virtual machine that has been created in each server in the server cabinet, and the first cabinet power of the server cabinet carried by the second request. The virtual machines to be migrated in the server cabinets described above include:

Determine the power of each virtual machine that has been created on each server according to the load of each virtual machine that has been created on each server in the server cabinet;

According to the target power, the power of the first rack, and the power of each virtual machine in the server rack, the virtual machine to be migrated is selected from the virtual machines that have been created in the server rack.
19. The method of claim 18, wherein, according to the target power, the power of the first rack, and the power of each virtual machine in the server rack, the virtual machine created from the server rack is Select the virtual machine to be migrated from the machine, including:

obtaining the difference between the power of the first cabinet and the target power;

Select at least two virtual machines whose sum of power is closest to and not less than the difference as the virtual machines to be migrated; or

A virtual machine whose power is closest to and not less than the difference is selected as the to-be-migrated virtual machine.
The method according to claim 18 or 19, wherein each virtual machine that has been created on each server is determined according to the load of each virtual machine that has been created on each server in the server rack machine power, including:

in,
represents the power of virtual machine i at time t;
Represents the power of the CPU of virtual machine i at time t;
represents the CPU utilization of virtual machine i at time t; α 1 and β 1 represent the weight and skew of CPU utilization-power conversion, respectively;
represents the power of the memory of the virtual machine i at time t;
represents the memory overhead of virtual machine i at time t; α 2 and β 2 represent the weight and skew of memory overhead-power conversion, respectively;
represents the power consumption of the hard disk of virtual machine i at time t;
Indicates the total number of bytes read and written on the hard disk of virtual machine i at time t, where,
Indicates the number of bytes written to the hard disk of virtual machine i at time t,
represents the number of hard disk read bytes of virtual machine i at time t; α 3 and β 3 represent the total number of hard disk read and write bytes - the weight and skew of power conversion, respectively.
A cabinet manager, characterized in that it is configured in a server cabinet, and the cabinet manager includes:

an obtaining unit, configured to obtain the first cabinet power of the server cabinet;

a sending unit, configured to send a first request to a virtual machine migration apparatus when the power of the first cabinet is within a preset power range, where the first request is used to instruct the virtual machine migration apparatus to prohibit the server Create a virtual machine in the enclosure.
The rack manager according to claim 21, wherein the first request comprises a server identifier of each server set in the server rack, and the first request is used to instruct the virtual machine migration apparatus Creation of virtual machines on any of the servers in the server rack is prohibited.
The cabinet manager according to claim 21 or 22, wherein the sending unit is further configured to:

In the case that the power of the first rack is greater than the preset power range, send a second request to the virtual machine migration apparatus, where the second request is used to instruct the virtual machine migration apparatus to confirm the waiting list in the server rack A virtual machine is migrated and the to-be-migrated virtual machine is migrated out of the server cabinet.
The rack manager of claim 23, wherein the second request includes the power of the first rack, a virtual machine identifier of each virtual machine that has been created in each server in the server rack, and virtual machine load.
The cabinet manager according to claim 24, wherein the cabinet manager further comprises a query unit; the query unit is further configured to: query each server in the server cabinet for each virtual machine that has been created The virtual machine ID and virtual machine load of the machine.
The cabinet manager according to claim 21 or 22, wherein the cabinet manager further comprises a determination unit;

the determining unit, configured to determine the virtual machine to be migrated in the server cabinet when the power of the first cabinet is greater than the preset power range;

The sending unit is further configured to send a virtual machine migration request to the virtual machine migration apparatus, where the virtual machine migration request is used to instruct the virtual machine migration apparatus to migrate the to-be-migrated virtual machine out of the server cabinet.
The cabinet manager according to claim 26, wherein the cabinet manager further comprises a recording unit and a selection unit;

The recording unit is used for pre-recording the target power;

The determining unit is further configured to determine the power of each virtual machine that has been created on each server according to the load of each virtual machine that has been created on each server in the server cabinet;

The selection unit is configured to select the virtual machine to be migrated from the virtual machines that have been created in the server rack according to the target power, the power of the first rack, and the power of each virtual machine in the server rack .
The cabinet manager according to claim 27, wherein the selection unit is further configured to:

obtaining the difference between the power of the first cabinet and the target power;

Select at least two virtual machines whose sum of power is closest to and not less than the difference as the virtual machines to be migrated; or

A virtual machine whose power is closest to and not less than the difference is selected as the to-be-migrated virtual machine.
The cabinet manager according to claim 27 or 28, wherein the determining unit is configured to use the following formula to determine the power of each virtual machine that has been created on each server:

in,
represents the power of virtual machine i at time t;
Represents the power of the CPU of virtual machine i at time t;
represents the CPU utilization of virtual machine i at time t; α 1 and β 1 represent the weight and skew of CPU utilization-power conversion, respectively;
represents the power of the memory of the virtual machine i at time t;
represents the memory overhead of virtual machine i at time t; α 2 and β 2 represent the weight and skew of memory overhead-power conversion, respectively;
represents the power consumption of the hard disk of virtual machine i at time t;
Indicates the total number of bytes read and written on the hard disk of virtual machine i at time t, where,
Indicates the number of bytes written to the hard disk of virtual machine i at time t,
represents the number of hard disk read bytes of virtual machine i at time t; α 3 and β 3 represent the total number of hard disk read and write bytes - the weight and skew of power conversion, respectively.
The rack manager of claim 21, wherein:

the obtaining unit is further configured to obtain the second cabinet power of the server cabinet;

The sending unit is further configured to send a third request including the server identifier of the server set in the server cabinet to the virtual machine migration apparatus when the power of the second cabinet is less than the preset power range, The third request is used to instruct the virtual machine migration apparatus to allow the virtual machine to be created in the server rack.
The cabinet manager according to claim 21, wherein the first cabinet power is measured power or predicted power.
A cabinet manager, characterized in that it is configured in a server cabinet, and the cabinet manager includes:

an obtaining unit, configured to obtain the first cabinet power of the server cabinet;

a sending unit, configured to send a second request to a virtual machine migration apparatus when the power of the first cabinet is greater than a preset power range, where the second request is used to instruct the virtual machine migration apparatus to confirm the server cabinet The virtual machine to be migrated in and migrate the virtual machine to be migrated out of the server cabinet.
The rack manager of claim 32, wherein the second request includes the first rack power, a virtual machine identifier of each virtual machine that has been created in each server in the server rack, and virtual machine load.
The cabinet manager according to claim 33, wherein the cabinet manager further comprises: a query unit, configured to query each server in the server cabinet for the virtual machine identifier of each virtual machine that has been created and virtual machine load.
The cabinet manager of any one of claims 32 to 34, wherein:

the obtaining unit is further configured to obtain the second cabinet power of the server cabinet;

The sending unit is further configured to send a third request including the server identifier of the server set in the server cabinet to the virtual machine migration apparatus when the power of the second cabinet is less than the preset power range, The third request is used to instruct the virtual machine migration apparatus to allow the virtual machine to be created in the server rack.
A virtual machine migration device, comprising:

an obtaining unit, configured to obtain a first request sent by the cabinet manager, where the first request includes a server identifier of each server set in the server cabinet where the cabinet manager is located;

A prohibiting unit, configured to prohibit the creation of a virtual machine in any server of the server cabinet according to the first request.
A virtual machine migration device, comprising:

an obtaining unit, configured to obtain a second request sent by the cabinet manager, where the second request includes the virtual machine identifier and virtual machine load of each virtual machine that has been created in each server in the server cabinet, and the server cabinet The power of the first cabinet;

a determining unit, configured to determine according to the virtual machine identifier and virtual machine load of each virtual machine that has been created in each server in the server rack and the first rack power of the server rack carried by the second request the virtual machine to be migrated in the server cabinet;

A migration unit, configured to migrate the virtual machine to be migrated out of the server cabinet.
The virtual machine migration apparatus according to claim 37, wherein the virtual machine migration apparatus further comprises: a recording unit and a selection unit;

The recording unit is used for pre-recording the target power of the server power;

The determining unit is configured to determine the power of each virtual machine that has been created on each server according to the load of each virtual machine that has been created on each server in the server cabinet;

The selection unit is configured to select the virtual machine to be migrated from the virtual machines that have been created in the server rack according to the target power, the power of the first rack, and the power of each virtual machine in the server rack .
The virtual machine migration apparatus according to claim 38, wherein the selection unit is used for:

obtaining the difference between the power of the first cabinet and the target power;

Select at least two virtual machines whose sum of power is closest to and not less than the difference as the virtual machines to be migrated; or

A virtual machine whose power is closest to and not less than the difference is selected as the to-be-migrated virtual machine.
The virtual machine migration apparatus according to claim 38 or 39, wherein the determining unit is configured to use the following formula to determine the power of each virtual machine that has been created on each server:

in,
represents the power of virtual machine i at time t;
Represents the power of the CPU of virtual machine i at time t;
represents the CPU utilization of virtual machine i at time t; α 1 and β 1 represent the weight and skew of CPU utilization-power conversion, respectively;
represents the power of the memory of the virtual machine i at time t;
represents the memory overhead of virtual machine i at time t; α 2 and β 2 represent the weight and skew of memory overhead-power conversion, respectively;
represents the power consumption of the hard disk of virtual machine i at time t;
Indicates the total number of bytes read and written on the hard disk of virtual machine i at time t, where,
Indicates the number of bytes written to the hard disk of virtual machine i at time t,
represents the number of hard disk read bytes of virtual machine i at time t; α 3 and β 3 represent the total number of hard disk read and write bytes - the weight and skew of power conversion, respectively.
A cabinet manager, characterized in that it includes a processor, a memory, and a transceiver;

the memory for storing computer instructions;

When the rack manager is running, the processor executes the computer instructions to cause the rack manager to perform the method of any one of claims 1-11.
A cabinet manager, characterized in that it includes a processor, a memory, and a transceiver;

the memory for storing computer instructions;

When the rack manager is running, the processor executes the computer instructions to cause the rack manager to perform the method of any of claims 12-15.
A virtual machine migration device, comprising a processor, a memory, and a transceiver;

the memory for storing computer instructions;

When the virtual machine migration apparatus is running, the processor executes the computer instructions, causing the virtual machine migration apparatus to perform the method of claim 16 .
A virtual machine migration device, comprising a processor, a memory, and a transceiver;

the memory for storing computer instructions;

When the virtual machine migration apparatus is running, the processor executes the computer instructions, so that the virtual machine migration apparatus executes the method of any one of claims 17-20.
A computer storage medium, characterized in that the computer storage medium comprises computer instructions, when the computer instructions are executed on an electronic device, the electronic device is made to perform the method or the method of any one of claims 1-11. The method of any of claims 12-15.
A computer storage medium, characterized in that the computer storage medium comprises computer instructions, when the computer instructions are executed on an electronic device, the electronic device causes the electronic device to perform the method of claim 16 or claims 17-20 The method of any one.
A computer program product. When the program code contained in the computer program product is executed by a processor in an electronic device, the electronic device can execute the method of any one of claims 1-11 or any one of claims 12-15. one of the methods described.
A computer program product, when the program code contained in the computer program product is executed by a processor in an electronic device, the electronic device can execute the method of claim 16 or the method of any one of claims 17-20. method.