Hot migration method for virtual machine, and related device

By obtaining virtual machine status information in the cloud computing system, selecting the appropriate target node for virtual machine hot migration, the problem of timeout and failure during the migration process is solved, and the stability and service quality of the cloud computing system are improved.

WO2024239865A1PCT designated stage expired Publication Date: 2024-11-28CLOUD INTELLIGENCE ASSETS HOLDING (SINGAPORE) PTE LTD

Patent Information

Application Number
PCT/CN2024/088675
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2023-05-22
Filing Date
2024-04-18
Publication Date
2024-11-28

AI Technical Summary

Technical Problem

In cloud computing systems, migration timeouts, migration failures, etc. are prone to occur during the hot migration of virtual machines, which affects the security and stability of the cloud computing system. How to accurately select the virtual machines and target physical nodes that need to be hot migration to achieve Effective and reliable thermal migration is an urgent problem.

Method used

By obtaining the status information of multiple virtual machines, filter out the virtual machines that need to be migrated, and based on the prediction of the migration results, select the appropriate target physical node for thermal migration. Use the prediction model trained by the integrated learning algorithm to predict the migration timeout probability and performance Loss, optimize the migration process.

Benefits of technology

It improves the reliability of hot migration, reduces the risk of migration failure and timeout, ensures the security and reliability of cloud computing systems, and provides higher quality and stable cloud services.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN2024088675_28112024_PF_FP_ABST
    Figure CN2024088675_28112024_PF_FP_ABST
Patent Text Reader

Abstract

Provided in the present description are a hot migration method for a virtual machine, and a related device, which are applied to a cloud computing system. The cloud computing system comprises a plurality of physical nodes, and a plurality of virtual machines are run on at least some of the plurality of physical nodes. The method comprises: on the basis of state information of a plurality of virtual machines, determining, from among the plurality of virtual machines, at least one virtual machine to be migrated; and predicting a hot migration result for the at least one virtual machine to be migrated, and on the basis of the predicted hot migration result, determining, from among a plurality of physical nodes, a target physical node corresponding to a target virtual machine, wherein the target virtual machine is a virtual machine to be migrated.
Need to check novelty before this filing date? Find Prior Art

Description

Virtual machine hot migration method and related equipment

[0001] This application claims priority to the Chinese patent application filed with the China Patent Office on May 22, 2023, with application number 202310581692.2 and invention name “A method for hot migration of virtual machines and related equipment”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] One or more embodiments of this specification relate to the field of cloud computing technology, and in particular, to a method for hot migration of a virtual machine and related equipment. Background Art

[0003] In cloud computing systems, a single physical node may be overloaded with multiple virtual machines, impacting the performance of each virtual machine. Hot migration allows virtual machines to be migrated from one physical node to another without shutting down or pausing them. This allows for seamless and lossless adjustments to resource allocation and load balancing within the cloud computing system, ensuring the performance of each virtual machine.

[0004] However, virtual machines are prone to migration timeouts and migration failures during live migration, which can affect the security and stability of cloud computing systems. Therefore, how to accurately select the virtual machines to be hot migrated and the target physical nodes to achieve effective and reliable hot migration is an urgent problem to be solved.

[0005] Summary of the Invention

[0006] In view of this, one or more embodiments of this specification provide a method for hot migration of a virtual machine and related devices.

[0007] In a first aspect, the present specification provides a method for hot migration of a virtual machine, which is applied to a cloud computing system, wherein the cloud computing system includes multiple physical nodes, and multiple virtual machines are running on at least some of the multiple physical nodes; the method includes: based on status information of the multiple virtual machines, determining at least one virtual machine to be migrated from the multiple virtual machines; predicting the hot migration result of the at least one virtual machine to be migrated, and based on the predicted hot migration result, determining a target physical node corresponding to the target virtual machine from the multiple physical nodes, the target virtual machine being the virtual machine to be migrated; and hot migrating the target virtual machine to the target physical node.

[0008] In the second aspect, the present specification provides a hot migration device for a virtual machine, which is applied to a cloud computing system, wherein the cloud computing system includes multiple physical nodes, and multiple virtual machines are running on at least some of the multiple physical nodes; the device includes: a determination unit, which is used to determine at least one virtual machine to be migrated from the multiple virtual machines based on status information of the multiple virtual machines; a prediction unit, which is used to predict the hot migration result of the at least one virtual machine to be migrated, and based on the predicted hot migration result, determine a target physical node corresponding to the target virtual machine from the multiple physical nodes, wherein the target virtual machine is the virtual machine to be migrated; and a hot migration unit, which is used to hot migrate the target virtual machine to the target physical node.

[0009] Accordingly, this specification also provides a computing device, comprising: a memory and a processor; the memory stores a computer program that can be executed by the processor; when the processor runs the computer program, it executes the hot migration method of the virtual machine as described in the above embodiments.

[0010] Accordingly, this specification also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for hot migration of a virtual machine as described in the above embodiments is executed. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] FIG1 is a schematic diagram of the architecture of a cloud computing system provided by an exemplary embodiment;

[0012] FIG2 is a flow chart of a method for hot migration of a virtual machine provided by an exemplary embodiment;

[0013] FIG3 is a schematic structural diagram of a thermal migration system provided by an exemplary embodiment;

[0014] FIG4 is a schematic structural diagram of a hot migration device for a virtual machine provided by an exemplary embodiment;

[0015] FIG5 is a schematic structural diagram of a server provided by an exemplary embodiment. DETAILED DESCRIPTION

[0016] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The implementations described in the following exemplary embodiments are not intended to represent all implementations consistent with one or more embodiments of this specification. Rather, they are merely examples of apparatuses and methods consistent with certain aspects of one or more embodiments of this specification, as detailed in the appended claims.

[0017] It should be noted that in other embodiments, the steps of the corresponding method are not necessarily performed in the order shown and described in this specification. In some other embodiments, the method may include more or fewer steps than those described in this specification. In addition, a single step described in this specification may be broken down into multiple steps for description in other embodiments, and multiple steps described in this specification may be combined into a single step for description in other embodiments.

[0018] The user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0019] First, some of the terms in this specification are explained to facilitate understanding by those skilled in the art.

[0020] (1) Cloud computing is an Internet-based computing model that uses virtualization technology to abstract physical resources (such as central processing units (CPUs), memory, storage, and networks) into virtual resources (such as virtual machines, virtual networks, and virtual storage) that can be allocated and used on demand. Cloud computing has the advantages of elasticity, scalability, low cost, and high efficiency, and is widely used in various industries and fields.

[0021] (2) A physical node (or physical machine) is a hardware device in a cloud computing system (configured with physical resources such as CPU, memory, and hard disk). A physical node provides the hardware conditions for the operation of virtual machines (VMs). Each physical node can run one or more VMs. When creating a VM in a physical node, part of the memory and hard disk capacity in the physical node is used as the VM's hard disk and memory capacity. Each VM running on a physical node is also called a VM instance.

[0022] (3) Hot migration refers to the process of migrating a virtual machine instance from one physical node to another without shutting down or pausing it. Hot migration allows for seamless and lossless adjustments to resource allocation and load balancing in cloud computing systems. However, hot migration also carries certain risks, such as migration timeout, migration failure, and data loss.

[0023] Correspondingly, cold migration refers to the process of migrating a virtual machine instance from one physical node to another after shutting down or pausing it, which will not be described in detail here.

[0024] (4) Instance Performance Profile: refers to data or models that describe and measure the status and performance of a virtual machine instance. A virtual machine instance profile may include data from multiple aspects, such as the virtual machine's corresponding CPU usage, memory usage, disk usage, network bandwidth, latency, and throughput, etc. This specification does not specifically limit this. A virtual machine instance profile can be used to evaluate the quality and requirements of a virtual machine instance and provide a reference for hot migration.

[0025] (5) Ensemble learning refers to completing learning tasks by building and combining multiple machine learning models. In supervised learning algorithms for machine learning, the goal is to learn a stable model that performs well in all aspects. However, the actual situation is often not ideal, and usually only multiple preferred models can be obtained (i.e., multiple weakly supervised models, each of which may perform better in certain aspects). Ensemble learning can combine multiple weakly supervised models to obtain a better and more comprehensive strongly supervised model (or ensemble model).

[0026] As mentioned above, in a cloud computing system, a physical node may be overloaded with multiple virtual machines. These virtual machines may compete for resources on the physical node (such as CPU and memory), resulting in performance degradation for each virtual machine running on the node, which in turn can cause user data loss or service interruption, causing serious losses to users. Based on this, hot migration can be used to migrate virtual machines from their current physical node to another without shutting down or pausing them. This allows for seamless and lossless adjustment of resource allocation and load balancing in the cloud computing system, thereby ensuring the performance of each virtual machine.

[0027] However, virtual machines are prone to migration timeouts, migration failures, and data loss during live migration, which in turn affects the security and stability of cloud computing systems and reduces the quality of cloud services. Therefore, how to accurately select the virtual machines to be hot migrated and the target physical nodes to achieve effective and reliable hot migration is an urgent problem to be solved.

[0028] Based on this, this specification provides a technical solution for filtering out at least one virtual machine that needs to be hot migrated based on the status information of multiple virtual machines, and then selecting a suitable target physical node for each virtual machine to be migrated based on the hot migration prediction results of each virtual machine to be migrated, thereby realizing effective and reliable hot migration.

[0029] During implementation, the present application can obtain status information of multiple virtual machines running on physical nodes in a cloud computing system, and based on this status information, determine at least one virtual machine to be migrated from the multiple virtual machines. Then, a hot migration result of the at least one virtual machine to be migrated is predicted, and based on the predicted hot migration result, a target physical node corresponding to any target virtual machine in the at least one virtual machine to be migrated is determined from the multiple physical nodes in the cloud computing system. Furthermore, the target virtual machine is hot migrated to the target physical node.

[0030] In the above technical solution, in a cloud computing system including multiple physical nodes, the present application can first obtain the status information of each virtual machine running on at least some of the multiple physical nodes. Then, the present application can determine at least one virtual machine that needs to be hot migrated from all virtual machines based on the status information of each virtual machine. For the virtual machine to be migrated, its hot migration result can be predicted first, and then based on the predicted hot migration result, a suitable target physical node can be selected for the virtual machine to be migrated, and hot migrated to the target physical node. In this way, the present application filters out the virtual machines that need to be hot migrated through the status information of the virtual machines, and further selects a suitable target physical node for hot migration based on the hot migration prediction result of the virtual machine. This can greatly improve the reliability of hot migration, reduce the risks of hot migration failure and hot migration timeout, and thus ensure the security and reliability of the entire cloud computing system, and provide users with higher quality and more stable cloud services.

[0031] Please refer to Figure 1, which is a schematic diagram of the architecture of a cloud computing system provided by an exemplary embodiment. As shown in Figure 1, the cloud computing system may include multiple physical nodes, for example, physical node 100a, physical node 100b, and physical node 100c. In an illustrated embodiment, physical node 100a, physical node 100b, and physical node 100c can communicate with each other via a network connection. As shown in Figure 1, the cloud computing system may also include a server 200. In an illustrated embodiment, server 200 can establish a communication connection with physical node 100a, physical node 100b, and physical node 100c. For example, server 200 can communicate with physical node 100a, physical node 100b, and physical node 100c via a network connection.

[0032] In an illustrated embodiment, one or more virtual machines may be running on the physical node 100a, the physical node 100b, and the physical node 100c.

[0033] Exemplarily, as shown in FIG1 , virtual machines A, B, and C are running on the physical node 100 a , virtual machine D is running on the physical node 100 b , and virtual machines E and F are running on the physical node 100 c .

[0034] In one illustrated embodiment, server 200 can monitor the status and performance of each physical node in real time and further obtain status information of each virtual machine running on each physical node. For example, server 200 can obtain status information of virtual machines A, B, and C running on physical node 100a, status information of virtual machine D running on physical node 100b, status information of virtual machines E and F running on physical node 100c, and so on.

[0035] In an illustrated embodiment, the state information of the virtual machine may include the CPU occupancy, memory occupancy, disk occupancy, network bandwidth, network delay and throughput, etc. of the virtual machine, which is not specifically limited in this specification.

[0036] In an illustrated embodiment, the server 200 may determine at least one virtual machine to be migrated from all virtual machines based on the acquired status information of each virtual machine.

[0037] In an illustrated embodiment, the server 200 may predict the hot migration impairment rate of each virtual machine based on the acquired state information of each virtual machine. The hot migration impairment rate may be used to indicate the degree of performance loss incurred by the virtual machine during the hot migration process.

[0038] For example, a live migration impairment rate of 50% may indicate that the virtual machine will lose half of its performance after live migration, i.e., the performance of the virtual machine after live migration is half of the performance before live migration. Another example is a live migration impairment rate of 20%, which may indicate that the virtual machine will lose 20% of its performance after live migration, i.e., the performance of the virtual machine after live migration is 20% of the performance before live migration. Examples are not given here.

[0039] In an illustrated embodiment, the server 200 may determine, based on the predicted hot migration impairment rates of the virtual machines, a virtual machine having a hot migration impairment rate less than a impairment rate threshold as a virtual machine to be migrated.

[0040] In one embodiment, the impairment rate threshold can be used to indicate the most severe performance loss that a user or system can accept. In one embodiment, the impairment rate threshold can be set based on user needs or actual conditions. For example, the impairment rate threshold can be 50%, 45%, or 20%, etc. This specification does not impose specific limitations on this.

[0041] For example, if the user has a higher demand for cloud service quality (i.e., the user has a lower tolerance for virtual machine damage), the damage rate threshold can be set to a lower value (e.g., 10%) to ensure the performance of each virtual machine undergoing hot migration; if the user cares more about resource utilization in the cloud computing system (i.e., the user has a higher tolerance for virtual machine damage), the damage rate threshold can be set to a higher value (e.g., 60%) to enable more virtual machines to undergo hot migration, and so on. This specification does not make any specific restrictions on this.

[0042] For example, taking the damage rate threshold of 20% as an example, the server 200 predicts, based on the status information of each virtual machine obtained, that the hot migration damage rate of the above-mentioned virtual machine A is 10%, the hot migration damage rate of virtual machine B is 40%, the hot migration damage rate of virtual machine C is 30%, the hot migration damage rate of virtual machine D is 30%, the hot migration damage rate of virtual machine E is 25%, and the hot migration damage rate of virtual machine F is 60%. Therefore, the server 200 can determine the virtual machine A whose hot migration damage rate is less than the damage rate threshold as the virtual machine to be migrated.

[0043] In one illustrated embodiment, after determining at least one virtual machine to be migrated, the server 200 can predict the hot migration result of the at least one virtual machine to be migrated, and select a suitable target physical node for each virtual machine for hot migration based on the predicted hot migration result.

[0044] Exemplarily, the suitable target physical node may be a physical node with relatively idle computing resources, for example, no virtual machine is running on the target physical node, or only a very small number of virtual machines are running on the target physical node, etc. This specification does not make any specific limitation on this.

[0045] In an illustrated embodiment, the server 200 may predict the hot migration result of the at least one virtual machine to be migrated based on the current resource occupancy of the at least one virtual machine to be migrated and subsequent changes in resource occupancy.

[0046] In one illustrated embodiment, the predicted hot migration result may include a hot migration timeout probability of the virtual machine. The hot migration timeout probability may be used to indicate the possibility that the hot migration duration required by the virtual machine during the hot migration process exceeds a time threshold. The time threshold may be a hot migration duration pre-set based on user needs or the actual situation of the cloud computing system. For example, if the virtual machine fails to complete the hot migration within the time threshold, the virtual machine is very likely to fail to migrate, or the hot migration may be directly regarded as a migration failure, etc. This specification does not specifically limit this.

[0047] For example, a 100% probability of hot migration timeout may indicate that the hot migration of a virtual machine will definitely time out or is highly likely to time out, leading to migration failure. Another example is a 10% probability of hot migration timeout, indicating that the hot migration of a virtual machine will rarely time out and has a high success rate. Examples are not given here.

[0048] In one illustrated embodiment, when different target physical nodes are selected for hot migration of the virtual machine to be migrated, due to differences in the status and performance of the different target physical nodes themselves (for example, differences in the number of CPU cores, memory capacity, and the number of running virtual machines, etc.), as well as differences in the quality of the network connection between different target physical nodes and the source physical node (i.e., the physical node where the virtual machine to be migrated is currently located), the hot migration effects of the same virtual machine corresponding to different target physical nodes may also be different. Exemplarily, the hot migration effect may include, for example, the hot migration timeout probability in the above-mentioned hot migration result, and may also include the probability of the virtual machine failing on the target physical node after completing the hot migration, and the probability of further performance degradation, etc. Based on this, the server 200 may select a suitable target physical node for hot migration for the virtual machine to be migrated based on the current resource occupancy of at least one virtual machine to be migrated and subsequent changes in resource occupancy, and further in combination with the status and performance of each physical node itself, as well as the quality of the network connection between each physical node.

[0049] 1 , the server 200 may predict the hot migration result of the virtual machine A to be migrated, select the physical node 100b as the target physical node based on the predicted hot migration result, and hot migrate the virtual machine A to the physical node 100b.

[0050] In an illustrated embodiment, the physical node 100a, the physical node 100b, and the physical node 100c may be a server having the above functions or a server cluster composed of multiple servers.

[0051] In one illustrated embodiment, server 200 may be a single server having the aforementioned functions or a server cluster comprised of multiple servers. For example, if server 200 is a server cluster comprised of multiple servers, the different functions of server 200 may be implemented by different servers in the server cluster. For example, the acquisition of virtual machine status information, the prediction of hot migration results before hot migration, and the actual hot migration processing may be implemented by different servers in the server cluster, etc. This specification does not impose specific limitations on this.

[0052] In one illustrated embodiment, the server 200 may also be any one of the above-mentioned physical nodes 100a, 100b and 100c, or the server 200 may include multiple physical nodes among the above-mentioned physical nodes 100a, 100b and 100c, that is, the functions involved in the above-mentioned hot migration method may be implemented through the physical nodes of the cloud computing system itself, etc. This specification does not make any specific limitations on this.

[0053] In one illustrated embodiment, the server 200 may also be a server or server cluster independent of the cloud computing system and connected to each physical node in the cloud computing system, etc. This specification does not make specific limitations on this.

[0054] In one illustrated embodiment, the server 200 may also be a performance awareness system in a cloud environment. The performance awareness system can detect and predict potential failures or attacks on virtual machine instances in a cloud computing environment, and take timely action to ensure the availability and security of cloud services. In one illustrated embodiment, the virtual machine hot migration method provided in this application can be applied to the performance awareness system to assist in reliable and efficient hot migration of virtual machines in a cloud computing system, minimizing virtual machine failures.

[0055] Please refer to Figure 2, which is a schematic flow diagram of a method for live migration of a virtual machine, provided in an exemplary embodiment. This method can be applied to a cloud computing system. Specifically, this method can be applied to server 200 shown in Figure 1 above. Server 200 can be a single server or a server cluster consisting of multiple servers, etc., and this specification does not specifically limit this. As shown in Figure 2, this method can specifically include the following steps S101 to S103.

[0056] Step S101 : determining at least one virtual machine to be migrated from the multiple virtual machines based on state information of the multiple virtual machines.

[0057] In an illustrated embodiment, a cloud computing system includes multiple physical nodes, and multiple virtual machines may run on at least some of the multiple physical nodes.

[0058] For example, each virtual machine may be configured with one or more virtual CPUs (vCPUs), which will not be described in detail here.

[0059] In one illustrated embodiment, during the operation of the virtual machine, the present application can detect the status and performance of each physical node in real time, as well as the status and performance of each virtual machine running on each physical node in real time, thereby obtaining status information of multiple virtual machines.

[0060] In one embodiment, the virtual machine's state information may include the virtual machine's CPU usage, memory usage, disk usage, network bandwidth, network latency, and throughput, etc., which is not specifically limited in this specification. In some possible embodiments, the virtual machine's state information may also include any other possible information for describing the virtual machine's state, which is not specifically limited in this specification.

[0061] In an illustrated embodiment, the present application may determine at least one virtual machine to be migrated from the multiple virtual machines based on the acquired state information of the multiple virtual machines.

[0062] In one illustrated embodiment, the present application can predict the hot migration damage rates of multiple virtual machines based on the acquired status information of multiple virtual machines, and determine at least one virtual machine to be migrated from the multiple virtual machines based on the hot migration damage rates of the multiple virtual machines.

[0063] The hot migration impairment rate can be used to indicate the degree of performance loss incurred by a virtual machine during the hot migration process. For example, a hot migration impairment rate of 50% can indicate that the virtual machine will lose half of its performance after hot migration, meaning that the performance of the virtual machine after hot migration is half of its performance before hot migration. Another example, a hot migration impairment rate of 10% can indicate that the virtual machine will lose 10% of its performance after hot migration, meaning that the performance of the virtual machine after hot migration is 10% of its performance before hot migration. Examples are not given here.

[0064] For example, the higher the CPU occupancy of the virtual machine, the greater its hot migration damage rate may be; the higher the memory occupancy of the virtual machine, the greater its hot migration damage rate may be; the more serious the network delay of the virtual machine, the greater its hot migration damage rate may be, and so on. This specification does not make specific limitations on this.

[0065] In one illustrated embodiment, in the process of determining at least one virtual machine to be migrated from the multiple virtual machines based on the hot migration damage rates of the multiple virtual machines, the present application can specifically determine a virtual machine among the multiple virtual machines whose hot migration damage rate is less than a damage rate threshold as the virtual machine to be migrated, so as to screen out at least one virtual machine to be migrated from the multiple virtual machines.

[0066] In one embodiment, the above-mentioned damage rate threshold can be set according to user needs or actual conditions, and this specification does not specifically limit this. For example, the damage rate threshold can be 50%, 45%, or 20%, etc., and this specification does not specifically limit this.

[0067] For example, if the user has a higher demand for cloud service quality (i.e., the user has a lower tolerance for virtual machine damage), the damage rate threshold can be set to a lower value (e.g., 5%) to ensure the performance of each virtual machine that performs hot migration; if the user cares more about resource utilization in the cloud computing system (i.e., the user has a higher tolerance for virtual machine damage), the damage rate threshold can be set to a higher value (e.g., 60%) to enable more virtual machines to perform hot migration, and so on. This specification does not make specific restrictions on this.

[0068] In an illustrated embodiment, the state information of the virtual machine may include an instance portrait of the virtual machine.

[0069] In one embodiment, the present application can obtain an instance profile of each virtual machine based on the various states and performance data detected during the operation of the virtual machine. In one embodiment, the instance profile of the virtual machine can include data or models used to describe the state of the virtual machine, such as the aforementioned CPU usage, memory usage, disk usage, network bandwidth, network latency, and throughput, etc.

[0070] In an illustrated embodiment, the present application may also generate a damage label and a migratable label corresponding to each virtual machine based on the instance portrait of each virtual machine. Exemplarily, the present application may predict the hot migration damage rate of each virtual machine based on the instance portrait of each virtual machine, and generate a damage label and a migratable label corresponding to each virtual machine based on the hot migration damage rate of each virtual machine. Among them, the damage label can be used to indicate the hot migration damage rate of the virtual machine, and the migratable label is used to indicate that the virtual machine can be hot migrated. In an illustrated embodiment, when the hot migration damage rate of a virtual machine is less than a damage rate threshold, the virtual machine may have a corresponding migratable label.

[0071] In one illustrated embodiment, the present application can determine the virtual machines with migratable labels (i.e., virtual machines whose hot migration damage rate is less than the damage rate threshold) as virtual machines to be migrated based on the damage labels and migratable labels corresponding to each virtual machine, and so on. This specification does not make specific limitations on this.

[0072] Step S102 : predicting a hot migration result of the at least one virtual machine to be migrated, and determining a target physical node corresponding to a target virtual machine from among the multiple physical nodes based on the predicted hot migration result, the target virtual machine being the virtual machine to be migrated.

[0073] In one illustrated embodiment, after determining at least one virtual machine to be migrated from multiple virtual machines, the present application can predict the hot migration result of the at least one virtual machine to be migrated. Based on the predicted hot migration result, a target physical node corresponding to any target virtual machine in the at least one virtual machine to be migrated is determined from multiple physical nodes included in the cloud computing system. Subsequently, the target virtual machine can be hot migrated from its source physical node to its corresponding target physical node.

[0074] In one illustrated embodiment, the present application can predict the hot migration result of at least one virtual machine to be migrated based on the current resource usage of at least one virtual machine to be migrated and subsequent resource usage changes. Exemplarily, the resource usage changes can include the resource usage of the virtual machine over a period of time in the future.

[0075] Illustratively, the resource occupancy of the virtual machine may include the occupancy of CPU, memory, network and other resources of the physical node aggregated by the virtual machine, etc., which is not specifically limited in this specification.

[0076] In an illustrated embodiment, the present application may first obtain historical resource occupancy of at least one virtual machine to be migrated, and based on the historical resource occupancy, predict the resource occupancy of the at least one virtual machine to be migrated.

[0077] In one illustrated embodiment, when predicting the resource usage of at least one virtual machine to be migrated, the present application may specifically include predicting the resource usage of the at least one virtual machine to be migrated within a preset duration after the current moment. For example, the preset duration can be set based on user needs or actual circumstances, such as 10 minutes, 5 minutes, or 20 minutes, and is not specifically limited in this specification.

[0078] Exemplarily, the historical resource occupancy status may be all historical resource occupancy status from the creation of the virtual machine to the current moment, or the historical resource occupancy status may be the historical resource occupancy status of the virtual machine within a period of time before the current moment (for example, 10 minutes or 15 minutes, etc.), etc. This specification does not make specific limitations on this.

[0079] In one illustrated embodiment, the present application can predict the hot migration result of at least one virtual machine to be migrated based on the current resource occupancy of at least one virtual machine to be migrated and the predicted resource occupancy of at least one virtual machine to be migrated within a preset time period after the current moment.

[0080] In one illustrated embodiment, the predicted hot migration result may include a hot migration timeout probability for the virtual machine. This hot migration timeout probability may be used to indicate the likelihood that the hot migration duration required by the virtual machine during the hot migration process will exceed a time threshold. This time threshold may be a pre-set hot migration time. If the virtual machine fails to complete the hot migration within the time threshold, the virtual machine is likely to have failed to migrate. Alternatively, the hot migration may be directly deemed to have failed, and so on. This specification does not specifically limit this.

[0081] For example, a 0% probability of hot migration timeout can indicate that the probability of hot migration timeout is almost zero. Another example is a 100% probability of hot migration timeout, which can indicate that the hot migration of the virtual machine will definitely time out or is very likely to time out, leading to migration failure. Another example is a 10% probability of hot migration timeout, which can indicate that the hot migration of the virtual machine will almost never time out, indicating a high success rate for hot migration. Examples are not given here.

[0082] In one illustrated embodiment, the present application can also select a suitable target physical node for the virtual machine to be migrated based on the hot migration results obtained by the above prediction, and further combined with the status and performance of each physical node itself, as well as the network connection quality between each physical node. For details, please refer to the description of the corresponding embodiment of Figure 1 above, which will not be repeated here.

[0083] In an illustrated embodiment, the present application may predict the hot migration result of at least one virtual machine to be migrated based on a pre-trained prediction model.

[0084] In one illustrated embodiment, the present application can input the current resource occupancy of the virtual machine to be migrated and the predicted resource occupancy of the virtual machine to be migrated within a preset time period after the current moment into a pre-trained prediction model to predict the hot migration result of the virtual machine to be migrated.

[0085] It should be noted that this specification does not specifically limit the training method and model type of the prediction model.

[0086] In one illustrated embodiment, the prediction model can be an integrated model that integrates at least two machine learning models for hot migration result prediction, obtained through training based on an ensemble learning algorithm. Accordingly, when predicting hot migration results using this integrated model, the present application can combine the hot migration timeout probabilities output by the at least two machine learning models in the integrated model to further obtain a more accurate hot migration timeout probability, thereby providing reliable data support for the subsequent implementation of hot migration.

[0087] In one illustrated embodiment, the at least two machine learning models mentioned above may be basic models, such as decision trees, support vector machines, neural networks, etc., which are not specifically limited in this specification.

[0088] In an illustrated embodiment, before predicting the hot migration results of each virtual machine, the present application may first sort at least one virtual machine to be migrated.

[0089] In one embodiment, the present application may predict the hot migration result of each virtual machine in the at least one virtual machine to be migrated in sequence according to the sorted order, and select a suitable target physical node for each virtual machine in sequence.

[0090] In one illustrated embodiment, the present application may also predict the hot migration results of a portion of the top-ranked virtual machines in the at least one virtual machine in the sorted order, and sequentially select a suitable target physical node for each virtual machine in the portion. Accordingly, the target virtual machine to be migrated may be any virtual machine in the portion of virtual machines.

[0091] Exemplarily, the portion of virtual machines may be a preset proportion of the top-ranked virtual machines of the at least one virtual machine, such as the top 50% of the at least one virtual machine, or the top 60% of the at least one virtual machine, etc., which is not specifically limited in this specification. Exemplarily, the portion of virtual machines may also be a preset number of the top-ranked virtual machines of the at least one virtual machine, such as the top 10 virtual machines, or the top 8 virtual machines, etc., which is not specifically limited in this specification.

[0092] In an illustrated embodiment, the present application may sort at least one virtual machine to be migrated based on a preset sorting reference factor.

[0093] In one illustrated embodiment, the preset sorting reference factors include a combination of one or more factors shown below: the current resource occupancy of the virtual machine, the resource occupancy of the virtual machine within a preset time period after the current moment, the sensitivity of the virtual machine to changes in the operating environment, the hot migration damage rate of the virtual machine, the user damage preference corresponding to the virtual machine, the damage rate threshold corresponding to the virtual machine, the user quota corresponding to the virtual machine, the total operating time of the virtual machine, etc. This specification does not make specific limitations on this.

[0094] A virtual machine's sensitivity to operating environment changes can affect the feasibility and user experience of live migration. Generally speaking, virtual machines with low sensitivity to operating environment changes are more adaptable to migration, while more sensitive virtual machines have difficulty adapting, experience greater performance loss after migration, and are more likely to experience failures. Operating environment changes can primarily include differences in the operating environment of the virtual machine on the source and target physical nodes before and after live migration.

[0095] The user's loss preference refers to the user's tolerance and preference for potential functional or performance loss during live migration. This preference reflects the user's emphasis on service quality and the trade-offs they make. Generally speaking, users with a lower tolerance for virtual machine performance loss care more about service quality, and their preset loss rate threshold will be lower accordingly. Users with a higher tolerance for virtual machine performance loss care more about resource utilization, and their preset loss rate threshold will be higher accordingly.

[0096] Among them, the user quota corresponding to the virtual machine can be used to measure the total number of virtual machine instances operated and maintained by the user in different windows (for example, 1 hour, 24 hours, 30 days, 90 days, etc.), and how many virtual machine instances are affected in performance.

[0097] The total running time of a virtual machine refers to the duration from creation to destruction (i.e., the lifetime of a virtual machine). This reflects the lifecycle and operational characteristics of a virtual machine and influences the necessity and urgency of virtual machine migration. Generally, hot migration of a virtual machine with a shorter total running time (such as a newly created virtual machine) has a smaller impact on users; hot migration of a virtual machine with a longer total running time has a greater impact on users.

[0098] In one embodiment shown, a priority sorting algorithm can also be designed to assign a weight value to each factor based on the importance and influence of each of the above factors on hot migration. Then, the total score of each virtual machine to be migrated is calculated based on the score of each virtual machine to be migrated on each factor and the weight value of each factor. For example, the less sensitive the virtual machine is to the operation and maintenance environment, the higher the score on this factor; the shorter the total running time of the virtual machine, the higher the score on this factor; the smaller the hot migration damage rate of the virtual machine, the higher the score on this factor, and so on. The examples are not listed here. Finally, the virtual machines to be migrated can be sorted according to the rule of total score from high to low. It can be understood that the virtual machine with a higher total score tends to have a higher hot migration success rate and is more suitable for hot migration.

[0099] In this way, the present application can prioritize hot migration result prediction and hot migration implementation for virtual machines with higher hot migration success rates and more suitable for hot migration by sorting at least one virtual machine to be migrated, while hot migration prediction and hot migration implementation can be performed later for virtual machines with lower rankings and lower hot migration success rates, or even not performed at all. This allows for the hot migration of virtual machines to be performed as much and successfully as possible while ensuring user experience and service quality, thereby improving resource utilization and saving costs.

[0100] Step S103: hot migrate the target virtual machine to the target physical node.

[0101] In one illustrated embodiment, after determining a corresponding target physical node for any target virtual machine among the at least one virtual machine to be migrated, the target virtual machine can be hot migrated from its source physical node to its corresponding target physical node. Ultimately, all of the at least one virtual machine to be migrated are hot migrated to their respective target physical nodes.

[0102] In an illustrated embodiment, after a target virtual machine is hot migrated from its source physical node to its corresponding target physical node, the service of the virtual machine may be restored on the target physical node.

[0103] In one illustrated embodiment, during the hot migration of a target virtual machine, hot migration parameters of the target virtual machine can be adjusted in real time based on the predicted hot migration timeout probability of the target virtual machine. The hot migration parameters can be used to control the hot migration duration of the target virtual machine, thereby minimizing hot migration timeouts while not affecting the operation of other virtual machines.

[0104] In one illustrated embodiment, the target virtual machine's live migration parameters may include a data transmission rate limit corresponding to the target virtual machine during the live migration process. It should be noted that the purpose of limiting the data transmission rate of a virtual machine is to prevent the virtual machine from occupying excessive network bandwidth during live migration, thereby affecting the normal operation of other virtual machines or applications.

[0105] Please refer to Figure 3, which is a schematic diagram of the structure of a hot migration system provided by an exemplary embodiment. In an illustrated embodiment, the hot migration system 20 may include the server 200 shown in Figure 1 above. Specifically, as shown in Figure 3, the hot migration system 20 may include a detection module 201, a portrait module 202, a screening module 203, a prediction module 204, a selection module 205, and a migration module 206. In an illustrated embodiment, a communication connection can be established between the detection module 201, the portrait module 202, the screening module 203, the prediction module 204, the selection module 205, and the migration module 206 based on a wired or wireless manner.

[0106] In one illustrated embodiment, detection module 201 is configured to detect the status and performance of each physical node in a cloud computing environment, as well as the status and performance of each virtual machine running on the physical node, in real time, and transmit the detected data to profiling module 202. Detection module 201 may also be configured to detect network traffic and resource usage of each physical node and virtual machine, etc., which is not specifically limited in this specification.

[0107] In an illustrated embodiment, the portrait module 202 is used to receive the detection data about each virtual machine sent by the detection module 201, and to profile each virtual machine based on the detection data to obtain an instance portrait of each virtual machine, and to predict the hot migration damage rate of each virtual machine based on the instance portrait of each virtual machine. In an illustrated embodiment, the portrait module 202 can also generate a damage label and a migratable label corresponding to each virtual machine based on the hot migration damage rate of each virtual machine. Among them, the damage label can be used to indicate the hot migration damage rate of the virtual machine, and the migratable label is used to indicate that the virtual machine can be hot migrated. In an illustrated embodiment, when the hot migration damage rate of a virtual machine is less than a damage rate threshold, the virtual machine can have a corresponding migratable label.

[0108] In one illustrated embodiment, screening module 203 is configured to receive the hot migration impairment rates of each virtual machine sent by profiling module 202 and, based on the hot migration impairment rates of each virtual machine, screen at least one virtual machine to be migrated from a large number of virtual machines. Exemplarily, screening module 203 may screen virtual machines with a migratable tag (i.e., virtual machines with a hot migration impairment rate less than a impairment rate threshold) as virtual machines to be migrated.

[0109] In one embodiment, the screening module 203 may also be used to sort the at least one virtual machine to be migrated that has been screened. In one illustrated embodiment, the screening module 203 may sort the at least one virtual machine to be migrated based on various factors, such as the current resource usage of the virtual machine and subsequent changes in resource usage, the hot migration damage rate of the virtual machine, the user quota corresponding to the virtual machine, and the total operating time of the virtual machine.

[0110] In an illustrated embodiment, the prediction module 204 is configured to predict the hot migration result of each virtual machine in the at least one virtual machine to be migrated in sequence according to the sorted order, and send the predicted hot migration result to the selection module 205 .

[0111] In an illustrated embodiment, the prediction module 204 may predict the hot migration result of the virtual machine based on a pre-trained prediction model. In an illustrated embodiment, the hot migration result may include a hot migration timeout probability of the virtual machine.

[0112] In an illustrated embodiment, the selection module 205 is configured to select a suitable target physical node for each virtual machine in sequence based on the hot migration results of each virtual machine predicted by the prediction module 204 and in a sorted order.

[0113] In one illustrated embodiment, migration module 206 is configured to migrate each virtual machine to be migrated from its corresponding source physical node to its corresponding target physical node and to restore the service corresponding to the virtual machine on the target physical node. During the hot migration process, migration module 206 may also adjust the hot migration parameters of each virtual machine in real time based on the hot migration timeout probability of each virtual machine, etc., although this specification does not specifically limit this.

[0114] The implementation process of the functions and effects of the above-mentioned modules can be specifically referred to the description of the embodiments corresponding to Figures 1 and 2 above, and will not be repeated here.

[0115] In one illustrated embodiment, the detection module 201, profiling module 202, screening module 203, selection module 204, prediction module 205, and migration module 206 may be implemented via software, hardware, or a combination thereof, which is not specifically limited in this specification. For example, each of the modules may reside on a different server, or may be different programs running on the same server, etc., which is not specifically limited in this specification.

[0116] In summary, in a cloud computing system comprising multiple physical nodes, the present application can first obtain the status information of each virtual machine running on at least some of the multiple physical nodes. Then, the present application can determine at least one virtual machine that needs to be hot migrated from all virtual machines based on the status information of each virtual machine. For the virtual machine to be migrated, its hot migration result can be predicted first, and then based on the predicted hot migration result, a suitable target physical node can be selected for the virtual machine to be migrated, and hot migrated to the target physical node. In this way, the present application filters out the virtual machines that need to be hot migrated through the status information of the virtual machines, and further selects a suitable target physical node for hot migration based on the hot migration prediction result of the virtual machine, which can greatly improve the reliability of hot migration, reduce the risks of hot migration failure and hot migration timeout, and thus ensure the security and reliability of the entire cloud computing system, and provide users with higher quality and more stable cloud services.

[0117] Corresponding to the implementation of the above-described method flow, embodiments of this specification also provide a virtual machine hot migration device. Please refer to FIG4 , which is a schematic diagram of the structure of a virtual machine hot migration device provided by an exemplary embodiment. This device 30 can be applied to a cloud computing system, which may include multiple physical nodes, at least some of which have multiple virtual machines running on them. As shown in FIG4 , this device 30 includes:

[0118] A determination unit 301 is used to determine at least one virtual machine to be migrated from the multiple virtual machines based on the status information of the multiple virtual machines; a prediction unit 302 is used to predict the hot migration result of the at least one virtual machine to be migrated, and based on the predicted hot migration result, determine a target physical node corresponding to the target virtual machine from the multiple physical nodes, wherein the target virtual machine is the virtual machine to be migrated; and a hot migration unit 303 is used to hot migrate the target virtual machine to the target physical node.

[0119] In an illustrated embodiment, the determination unit 301 is specifically used to: predict the hot migration damage rate of the multiple virtual machines based on the status information of the multiple virtual machines; wherein the hot migration damage rate is used to indicate the degree of performance loss generated by the virtual machines during the hot migration process; and determine the virtual machines among the multiple virtual machines whose hot migration damage rate is less than the damage rate threshold as the virtual machines to be migrated.

[0120] In an illustrated embodiment, the prediction unit 302 is specifically configured to: predict the resource occupancy of the at least one virtual machine to be migrated based on the historical resource occupancy of the at least one virtual machine to be migrated;

[0121] Based on the current resource occupancy of the at least one virtual machine to be migrated and the predicted resource occupancy, a hot migration result of the at least one virtual machine to be migrated is predicted.

[0122] In an illustrated embodiment, the prediction unit 302 is specifically used to: sort the at least one virtual machine to be migrated based on a preset sorting reference factor; and predict the hot migration results of some of the virtual machines ranked at the top of the at least one virtual machine to be migrated in sequence according to the sorted order.

[0123] In one illustrated embodiment, the preset sorting reference factors include a combination of one or more factors shown below: the current resource usage and predicted resource usage of the virtual machine, the sensitivity of the virtual machine to changes in the operating environment, the hot migration damage rate of the virtual machine, the user damage preference corresponding to the virtual machine, the user quota corresponding to the virtual machine, and the total operating time of the virtual machine.

[0124] In one illustrated embodiment, the predicted hot migration result includes a hot migration timeout probability; wherein the hot migration timeout probability is used to indicate the possibility that the hot migration duration required by the virtual machine during the hot migration process exceeds a duration threshold.

[0125] In one illustrated embodiment, the hot migration unit 303 is specifically used to: adjust the hot migration parameters of the target virtual machine based on the predicted hot migration timeout probability of the target virtual machine during the hot migration of the target virtual machine to the target physical node; wherein the hot migration parameters are used to control the hot migration duration of the target virtual machine.

[0126] In an illustrated embodiment, the hot migration parameters of the target virtual machine include a data transmission rate limit corresponding to the target virtual machine during the hot migration process.

[0127] In one illustrated embodiment, the prediction unit 302 is specifically used to: predict the hot migration result of at least one virtual machine to be migrated based on a pre-trained prediction model; the prediction model is an integrated model trained based on an integrated learning algorithm that integrates at least two machine learning models for hot migration result prediction.

[0128] In an illustrated embodiment, the state information includes a combination of one or more of the following data: CPU occupancy, memory occupancy, disk occupancy, network bandwidth, network latency, and throughput of the virtual machine.

[0129] The implementation process of the functions and effects of each unit in the above-mentioned device 30 is specifically described in the description of the corresponding embodiments of Figures 1 to 3 above, and will not be repeated here. It should be understood that the above-mentioned device 30 can be implemented by software, or by hardware or a combination of software and hardware. Taking software implementation as an example, as a device in the logical sense, it is formed by the processor (CPU) of the device where it is located reading the corresponding computer program instructions into the memory and running them. From the hardware level, in addition to the CPU and memory, the device where the above-mentioned device is located usually also includes other hardware such as chips for wireless signal transmission and reception, and / or other hardware such as boards and cards for implementing network communication functions.

[0130] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical modules, that is, they may be located in one place or distributed across multiple network modules. Some or all of the units or modules may be selected according to actual needs to achieve the purpose of the scheme of this specification. Those of ordinary skill in the art can understand and implement the present invention without inventive effort.

[0131] The devices, units, and modules described in the above embodiments may be implemented by computer chips or entities, or by products having certain functions. A typical implementation device is a computer, which may be in the form of a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email transceiver, game console, tablet computer, wearable device, or any combination of these devices.

[0132] Corresponding to the above-mentioned method embodiment, the embodiment of this specification also provides a server. Please refer to Figure 5, which is a structural diagram of a server provided by an exemplary embodiment. Exemplarily, the server can be the server 200 shown in Figure 1 above. Exemplarily, the server 200 can be any physical node among the physical node 100a, physical node 100b, physical node 100c, etc. in the cloud computing system shown in Figure 1 above, or the server 200 can also be any other possible server or server cluster in the cloud computing system for executing the above-mentioned hot migration method of the virtual machine, etc., which is not specifically limited in this specification. As shown in Figure 5, the server can include a processor 1001 and a memory 1002, and can further include an input device 1004 (such as a keyboard, etc.) and an output device 1005 (such as a display, etc.). The processor 1001, the memory 1002, the input device 1004 and the output device 1005 can be connected by a bus or other means. As shown in Figure 5, the memory 1002 includes a computer-readable storage medium 1003, which stores a computer program that can be executed by the processor 1001. The processor 1001 can be a general-purpose processor, a microprocessor, or an integrated circuit for controlling the execution of the above method embodiment. When the processor 1001 runs the stored computer program, it can execute the various steps of the hot migration method of the virtual machine in the embodiment of this specification, including: obtaining the status information of the multiple virtual machines, and determining at least one virtual machine to be migrated from the multiple virtual machines based on the status information; predicting the hot migration result of the at least one virtual machine to be migrated, and based on the predicted hot migration result, determining the target physical node corresponding to any target virtual machine in the at least one virtual machine to be migrated from the multiple physical nodes; hot migrating the target virtual machine to the target physical node, and so on.

[0133] For a detailed description of each step of the above-mentioned virtual machine live migration method, please refer to the previous content and will not be repeated here.

[0134] In summary, in a cloud computing system comprising multiple physical nodes, the present application can first obtain the status information of each virtual machine running on at least some of the multiple physical nodes. Then, the present application can determine at least one virtual machine that needs to be hot migrated from all virtual machines based on the status information of each virtual machine. For the virtual machine to be migrated, its hot migration result can be predicted first, and then based on the predicted hot migration result, a suitable target physical node can be selected for the virtual machine to be migrated, and the virtual machine can be hot migrated to the target physical node. In this way, the present application filters out the virtual machines that need to be hot migrated through the status information of the virtual machines, and further selects a suitable target physical node for hot migration based on the hot migration prediction result of the virtual machine, which can greatly improve the reliability of hot migration, reduce the risks of hot migration failure and hot migration timeout, and thus ensure the security and reliability of the entire cloud computing system, and provide users with higher quality and more stable cloud services.

[0135] Corresponding to the above method embodiments, embodiments of this specification also provide a computer-readable storage medium storing a computer program. When executed by a processor, these computer programs perform the steps of the virtual machine live migration method described in the embodiments of this specification. For details, please refer to the description of the embodiments corresponding to Figures 1-3 above, and will not be repeated here.

[0136] The above description is only a preferred embodiment of this specification and is not intended to limit this specification. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of this specification should be included in the scope of protection of this specification.

[0137] In a typical configuration, a terminal device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0138] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0139] Computer-readable media include permanent and non-permanent, removable and non-removable media that can store information using any method or technology. The information can be computer-readable instructions, data structures, program modules or other data.

[0140] Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory media such as modulated data signals and carrier waves.

[0141] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0142] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems, or computer program products. Thus, the embodiments of this specification may take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware. Furthermore, the embodiments of this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

Claims

1. A method for hot migration of a virtual machine, applied to a cloud computing system, wherein the cloud computing system includes a plurality of physical nodes, and a plurality of virtual machines are running on at least some of the physical nodes; the method comprises: Based on the state information of the multiple virtual machines, determining at least one virtual machine to be migrated from the multiple virtual machines; Predicting a hot migration result of the at least one virtual machine to be migrated, and determining a target physical node corresponding to a target virtual machine from among the multiple physical nodes based on the predicted hot migration result, the target virtual machine being the virtual machine to be migrated; Hot migrate the target virtual machine to the target physical node.

2. The method according to claim 1, wherein determining at least one virtual machine to be migrated from the multiple virtual machines based on the state information of the multiple virtual machines comprises: Based on the status information of the multiple virtual machines, predict the hot migration damage rate of the multiple virtual machines; wherein the hot migration damage rate is used to indicate the degree of performance loss generated by the virtual machine during the hot migration process; A virtual machine among the multiple virtual machines whose hot migration damage rate is less than a damage rate threshold is determined as a virtual machine to be migrated.

3. The method according to claim 2, wherein the predicting the hot migration result of the at least one virtual machine to be migrated comprises: Predicting the resource occupancy of the at least one virtual machine to be migrated based on the historical resource occupancy of the at least one virtual machine to be migrated; Based on the current resource occupancy of the at least one virtual machine to be migrated and the predicted resource occupancy, a hot migration result of the at least one virtual machine to be migrated is predicted.

4. The method according to claim 3, wherein predicting the hot migration result of the at least one virtual machine to be migrated comprises: Sorting the at least one virtual machine to be migrated based on a preset sorting reference factor; According to the sorted order, the hot migration results of some virtual machines ranked top in the at least one virtual machine to be migrated are predicted in sequence.

5. The method according to claim 4, wherein the preset ranking reference factor comprises a combination of one or more factors shown below: The current resource occupancy and predicted resource occupancy of the virtual machine, the sensitivity of the virtual machine to changes in the operating environment, the hot migration damage rate of the virtual machine, the user damage preference corresponding to the virtual machine, the user quota corresponding to the virtual machine and the total operating time of the virtual machine.

6. The method according to claim 1, wherein the predicted hot migration result includes a hot migration timeout probability; wherein: The hot migration timeout probability is used to indicate the possibility that the hot migration duration required by the virtual machine during the hot migration process exceeds a duration threshold.

7. The method according to claim 6, wherein hot migrating the target virtual machine to the target physical node comprises: During hot migration of the target virtual machine to the target physical node, the hot migration parameters of the target virtual machine are adjusted based on the predicted hot migration timeout probability of the target virtual machine; wherein the hot migration parameters are used to control the hot migration duration of the target virtual machine.

8. The method according to claim 1, wherein predicting the hot migration result of the at least one virtual machine to be migrated comprises: Predicting a hot migration result of the at least one virtual machine to be migrated based on a pre-trained prediction model; The prediction model is an integrated model that is trained based on an integrated learning algorithm and integrates at least two machine learning models for predicting thermal migration results.

9. The method according to any one of claims 1 to 8, wherein the status information comprises a combination of one or more data shown below: The CPU occupancy, memory occupancy, disk occupancy, network bandwidth, network latency and throughput of the virtual machine.

10. A virtual machine hot migration device, applied to a cloud computing system, wherein the cloud computing system includes a plurality of physical nodes, and a plurality of virtual machines are running on at least some of the plurality of physical nodes; the device includes: a determining unit, which determines at least one virtual machine to be migrated from the multiple virtual machines based on the state information of the multiple virtual machines; a prediction unit, configured to predict a hot migration result of the at least one virtual machine to be migrated, and determine a target physical node corresponding to a target virtual machine from among the multiple physical nodes based on the predicted hot migration result, the target virtual machine being the virtual machine to be migrated; A hot migration unit is used to hot migrate the target virtual machine to the target physical node.

11. A computing device comprising: Memory and processor; The memory stores a computer program executable by the processor; When the processor runs the computer program, the processor performs the method according to any one of claims 1 to 10.

12. A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method according to any one of claims 1 to 10.

Citation Information

Patent Citations

  • IaaS cloud platform-oriented method and system for optimizing resource scheduling efficiency

    CN108595250A

  • Virtual machine migration method and device and computer equipment

    CN110162379A

  • Fault detection method and device

    CN113326158A

  • Virtual machine live migration method and related equipment

    CN116700890A

  • Virtual machine hot migration method and apparatus, and system

    US20190188024A1

Cited By

  • Intelligent scheduling management system for private cloud resources based on hyper-converged architecture

    CN120011088A

  • Page storage method, virtual machine system, device, medium and product

    CN120144063A

  • A page storage method, virtual machine system, device, medium and product

    CN120144063B

  • Load balancing method and device, electronic equipment and storage medium

    CN120610780A