CN112181594A - Virtual machine live migration method, device, equipment and storage medium - Google Patents
Virtual machine live migration method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN112181594A CN112181594A CN202011035520.8A CN202011035520A CN112181594A CN 112181594 A CN112181594 A CN 112181594A CN 202011035520 A CN202011035520 A CN 202011035520A CN 112181594 A CN112181594 A CN 112181594A
- Authority
- CN
- China
- Prior art keywords
- virtual machine
- target
- live migration
- migration
- thermal
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000013508 migration Methods 0.000 title claims abstract description 300
- 230000005012 migration Effects 0.000 title claims abstract description 300
- 238000000034 method Methods 0.000 title claims abstract description 215
- 238000012549 training Methods 0.000 claims abstract description 61
- 238000012546 transfer Methods 0.000 claims abstract description 14
- 238000007906 compression Methods 0.000 claims description 20
- 230000006835 compression Effects 0.000 claims description 20
- 238000003066 decision tree Methods 0.000 claims description 12
- 238000004590 computer program Methods 0.000 claims description 10
- 230000005540 biological transmission Effects 0.000 claims description 7
- 238000007726 management method Methods 0.000 description 8
- 238000004422 calculation algorithm Methods 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 4
- 230000001133 acceleration Effects 0.000 description 2
- 238000013144 data compression Methods 0.000 description 2
- 230000006837 decompression Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000011084 recovery Methods 0.000 description 2
- 238000007635 classification algorithm Methods 0.000 description 1
- AGVAZMGAQJOSFJ-WZHZPDAFSA-M cobalt(2+);[(2r,3s,4r,5s)-5-(5,6-dimethylbenzimidazol-1-yl)-4-hydroxy-2-(hydroxymethyl)oxolan-3-yl] [(2r)-1-[3-[(1r,2r,3r,4z,7s,9z,12s,13s,14z,17s,18s,19r)-2,13,18-tris(2-amino-2-oxoethyl)-7,12,17-tris(3-amino-3-oxopropyl)-3,5,8,8,13,15,18,19-octamethyl-2 Chemical compound [Co+2].N#[C-].[N-]([C@@H]1[C@H](CC(N)=O)[C@@]2(C)CCC(=O)NC[C@@H](C)OP(O)(=O)O[C@H]3[C@H]([C@H](O[C@@H]3CO)N3C4=CC(C)=C(C)C=C4N=C3)O)\C2=C(C)/C([C@H](C\2(C)C)CCC(N)=O)=N/C/2=C\C([C@H]([C@@]/2(CC(N)=O)C)CCC(N)=O)=N\C\2=C(C)/C2=N[C@]1(C)[C@@](C)(CC(N)=O)[C@@H]2CCC(N)=O AGVAZMGAQJOSFJ-WZHZPDAFSA-M 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
- G06F2009/45562—Creating, deleting, cloning virtual machine instances
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
- G06F2009/4557—Distribution of virtual machine instances; Migration and load balancing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Stored Programmes (AREA)
Abstract
The invention discloses a method, a device, equipment and a storage medium for live migration of a virtual machine, wherein the method comprises the following steps: determining a virtual machine needing to realize the live migration as a target virtual machine, and acquiring values of various external factors influencing the live migration in a scene where the target virtual machine is located as a sample to be input; inputting the sample to be input into an intelligent model, and determining corresponding thermal migration methods and parameters output by the intelligent model as a target thermal migration method and a target parameter respectively; the intelligent model is obtained by utilizing a training sample in advance, and the training sample comprises values of various external factors in different scenes and the heat transfer method and parameters with the shortest total time consumption in the corresponding scene; and based on the target parameters, implementing corresponding live migration on the target virtual machine according to the target live migration method. The total time consumption of thermal migration can be effectively reduced, and the thermal migration efficiency is further improved.
Description
Technical Field
The invention relates to the technical field of openstack virtualization, in particular to a virtual machine live migration method, device, equipment and storage medium.
Background
With the development of cloud computing and virtualization technologies, deployment of cloud services by using openstack (a cloud computing management platform) is becoming more and more widely applied to data centers and cluster scenes; virtualization is one of indispensable key technologies for constructing a cloud infrastructure, more virtual machines are abstractly simulated in a set of computer resources, and each virtual machine can be used as an independent terminal to be added into a cloud distributed system; virtualization therefore has great advantages in terms of efficient use of resources, dynamic deployment, and high reliability.
The live migration is to move the running virtual machine between different physical machines, during which the link from the client cannot be disconnected, and the memory, storage and network links of the virtual machine are all transferred from the source end to the destination end. The migration of the virtual machine simplifies the maintenance management of the system, improves the load balance of the system, enhances the fault tolerance of the load and optimizes the power management of the system. Therefore, how to reduce the total time consumption of the thermal migration and improve the thermal migration efficiency is a problem to be solved by those skilled in the art.
Disclosure of Invention
The invention aims to provide a virtual machine live migration method, a virtual machine live migration device, virtual machine live migration equipment and a virtual machine storage medium, which can effectively reduce the total time consumption of live migration and improve the live migration efficiency.
In order to achieve the above purpose, the invention provides the following technical scheme:
a virtual machine live migration method comprises the following steps:
determining a virtual machine needing to realize the live migration as a target virtual machine, and acquiring values of various external factors influencing the live migration in a scene where the target virtual machine is located as a sample to be input;
inputting the sample to be input into an intelligent model, and determining corresponding thermal migration methods and parameters output by the intelligent model as a target thermal migration method and a target parameter respectively; the intelligent model is obtained by utilizing a training sample in advance, and the training sample comprises values of various external factors in different scenes and the heat transfer method and parameters with the shortest total time consumption in the corresponding scene;
and based on the target parameters, implementing corresponding live migration on the target virtual machine according to the target live migration method.
Preferably, before the intelligent model is obtained by training using the training samples and the label values, the method further includes:
deleting or adding a live migration method in a preset white list under the instruction of a user, and determining each live migration method contained in the white list as the live migration method contained in the label value.
Preferably, determining the corresponding thermal migration method and parameter output by the intelligent model as a target thermal migration method and a target parameter respectively includes:
if the intelligent model outputs a thermal migration method and a thermal migration parameter, determining that the thermal migration method and the thermal migration parameter output by the intelligent model are a target thermal migration method and a target parameter respectively; and if the intelligent model outputs a plurality of thermal migration methods and parameters, determining the thermal migration method with the highest priority in the plurality of thermal migration methods and parameters and the corresponding parameter as a target thermal migration method and a target parameter respectively.
Preferably, the method further comprises the following steps:
and receiving a priority updating instruction input by an external user, and updating the priority of each live migration method and the corresponding parameter based on the priority updating instruction.
Preferably, the external factors include a service scenario, a load type, a bandwidth, and a configuration of a host; the values of the service scenes comprise industries, the values of the load types comprise CPU intensive type, I/O intensive type, memory intensive type, network intensive type and mixed type, the values of the bandwidths comprise bandwidth transmission speed and bandwidth stability, and the configured values of the host machine comprise whether the host machine where the corresponding virtual machine is located supports a network card and whether hardware compression is supported.
Preferably, the training by using the training samples and the label values to obtain the intelligent model comprises:
and training a pre-established decision tree by using the training samples and the label values to obtain the intelligent model.
Preferably, after the target virtual machine is implemented with the corresponding live migration according to the target live migration method based on the target parameter, the method further includes:
and sending the information of completing the hot migration of the target virtual machine to a corresponding management terminal.
A virtual machine live migration apparatus, comprising:
an acquisition module to: determining a virtual machine needing to realize the live migration as a target virtual machine, and acquiring values of various external factors influencing the live migration in a scene where the target virtual machine is located as a sample to be input;
a determination module to: inputting the sample to be input into an intelligent model, and determining corresponding thermal migration methods and parameters output by the intelligent model as a target thermal migration method and a target parameter respectively; the intelligent model is obtained by utilizing a training sample in advance, and the training sample comprises values of various external factors in different scenes and the heat transfer method and parameters with the shortest total time consumption in the corresponding scene;
a migration module to: and based on the target parameters, implementing corresponding live migration on the target virtual machine according to the target live migration method.
A virtual machine live migration apparatus, comprising:
a memory for storing a computer program;
a processor configured to implement the steps of the virtual machine live migration method as described in any one of the above when the computer program is executed.
A computer readable storage medium having stored thereon a computer program which, when executed by a processor, carries out the steps of the virtual machine live migration method of any one of the preceding claims.
The invention provides a method, a device, equipment and a storage medium for live migration of a virtual machine, wherein the method comprises the following steps: determining a virtual machine needing to realize the live migration as a target virtual machine, and acquiring values of various external factors influencing the live migration in a scene where the target virtual machine is located as a sample to be input; inputting the sample to be input into an intelligent model, and determining corresponding thermal migration methods and parameters output by the intelligent model as a target thermal migration method and a target parameter respectively; the intelligent model is obtained by utilizing a training sample in advance, and the training sample comprises values of various external factors in different scenes and the heat transfer method and parameters with the shortest total time consumption in the corresponding scene; and based on the target parameters, implementing corresponding live migration on the target virtual machine according to the target live migration method. According to the method and the device, the intelligent model is obtained by training the values of all external factors in the scene where the virtual machine is located under different scenes and the thermal migration method and the parameters with the shortest total time consumption under the scene, so that when the thermal migration of any virtual machine is realized, the values of all external factors in the scene where the any virtual machine is located can be input into the intelligent model, the thermal migration method and the parameters output by the intelligent model are obtained, the thermal migration method and the parameters with the shortest total time consumption during the thermal migration can be realized under the scene where the any virtual machine is located, the thermal migration of the any virtual machine can be realized according to the thermal migration method and the parameters, the total time consumption of the thermal migration can be effectively reduced, and the thermal migration efficiency can be further improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
Fig. 1 is a flowchart of a virtual machine live migration method according to an embodiment of the present invention;
fig. 2 is a diagram illustrating a virtual machine live migration method according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of a virtual machine live migration apparatus according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, a flowchart of a virtual machine live migration method according to an embodiment of the present invention is shown, where the method includes:
s11: determining a virtual machine needing to realize the live migration as a target virtual machine, and acquiring values of various external factors influencing the live migration in a scene where the target virtual machine is located as a sample to be input.
The virtual machine live migration method provided by the embodiment of the invention can be a corresponding virtual machine live migration device; any virtual machine needing to realize live migration can be used as a target virtual machine, so that after the target virtual machine is determined, corresponding live migration is realized for the target virtual machine; live Migration, namely Live Migration, also called dynamic Migration and real-time Migration, namely virtual machine storage/recovery, generally completely stores the running state of the whole virtual machine, and can quickly recover to the original hardware platform or even different hardware platforms; after recovery, the virtual machine is still running smoothly and the user does not perceive any differences.
After the target virtual machine is determined, values of various external factors in a scene where the target virtual machine is located can be obtained, wherein the external factors are factors which can affect the live migration of the target virtual machine in the scene where the target virtual machine is located, such as a service scene, a load type, a bandwidth, configuration of a host where the target virtual machine is located, and other settings performed according to actual needs are within the protection scope of the present invention. After the values of the external factors in the scene of the target virtual machine are obtained, the values of the factors can be combined into a corresponding one-dimensional vector, the one-dimensional vector is a sample which needs to be input into the intelligent model and can be called as a sample to be input, and the dimension of the one-dimensional vector is the number of terms of the external factors.
S12: inputting a sample to be input into an intelligent model, and determining a corresponding thermal migration method and parameters output by the intelligent model as a target thermal migration method and target parameters respectively; the intelligent model is obtained by utilizing a training sample in advance for training, and the training sample comprises values of various external factors in different scenes and the heat transfer method and parameters with the shortest total time consumption in the corresponding scenes.
Inputting a sample to be input into an intelligent model, wherein the intelligent model learns the sample to be input so as to obtain and output a thermal migration method corresponding to the sample to be input and parameters of the thermal migration method; wherein, the live migration method may include a multithread compression migration method, a post-copy migration method, an xbzlle migration method (an xbzlle migration method, that is, an Xor Binary Zero Run-Length-Encoding, which performs an exclusive or operation on the modified page and a previously transmitted page, and the result of the exclusive or operation is a changed data block, and then transmits the part of the difference data to the target, and the target receives the data and then applies the data to the corresponding page), an automatic convergence migration method, and an RDMA migration method (RDMA technology, that is, Remote Direct Memory Access, which is a new Direct Memory Access technology, RDMA allows the computer to directly Access the Memory of the computer without being processed by the processor, RDMA quickly moves the data from a system to the Memory of a Remote system without any influence on the operating system), and the parameters are parameters used by the live migration method, if the live migration method is a multi-thread compression migration method, the corresponding parameters may include the number of threads, etc., and if the live migration method is an automatic convergence migration method, the corresponding parameters may include an initial vCPU drop frequency and an incremental vCPU drop frequency (reducing the operating frequency or the execution time of the vCPU, thereby reducing the overall operating time of the vCPU, and reducing the generation of the dirty page rate), etc., and if the live migration method is an xbzlle migration method, the corresponding parameters may include the cache size, etc.
It should be noted that the intelligent model is obtained by training a corresponding classification network in advance; the training sample comprises a characteristic attribute and a corresponding label value, wherein the characteristic attribute is the value of each external factor in any scene, and the label value is the heat transfer method and the parameter with the shortest total time consumption corresponding to the characteristic attribute. Specifically, based on values of various external factors in different scenes, testing total consumption of different live migration methods and different parameters of the live migration methods in the live migration of any virtual machine, and selecting the live migration method with the shortest total consumption in the live migration as the optimal live migration method in the scene of the value corresponding to each group of external factors so as to establish a training sample; specifically, each scene (different extrinsic factor values) is used as a feature attribute (or a training sample) of the training sample, the thermal migration method and parameter with the shortest total time consumption in each scene are used as a label value of the feature attribute in the scene, a certain amount of training data is established and then sent to a pre-established classification network, and an optimal thermal migration intelligent model is obtained. In addition, the total consumed time is the total time required from the beginning of the live migration of the virtual machine to the end of the live migration, the training sample may be a one-dimensional vector composed of values of each external factor, the values in the training sample and the external factors corresponding to the values in the sample to be input are the same, and the value arrangement sequence corresponding to each external factor in the training sample and the sample to be input is also the same.
S13: and based on the target parameters, implementing corresponding live migration on the target virtual machine according to the target live migration method.
After determining the live migration method and the parameters used by the live migration method, the target virtual machine may be live migrated using the corresponding parameters according to the live migration method.
According to the method and the device, the intelligent model is obtained by training the values of all external factors in the scene where the virtual machine is located under different scenes and the thermal migration method and the parameters with the shortest total time consumption under the scene, so that when the thermal migration of any virtual machine is realized, the values of all external factors in the scene where the any virtual machine is located can be input into the intelligent model, the thermal migration method and the parameters output by the intelligent model are obtained, the thermal migration method and the parameters with the shortest total time consumption during the thermal migration can be realized under the scene where the any virtual machine is located, the thermal migration of the any virtual machine can be realized according to the thermal migration method and the parameters, the total time consumption of the thermal migration can be effectively reduced, and the thermal migration efficiency can be further improved.
Before training by using a training sample and a label value to obtain an intelligent model, the virtual machine live migration method provided by the embodiment of the invention may further include:
deleting or adding the live migration methods in a preset white list under the instruction of a user, and determining each live migration method contained in the white list as a live migration method contained in a label value.
In the embodiment of the application, a white list of live migration methods may be set, the live migration method in the white list is a method which can be currently used for implementing the live migration of the virtual machine, and the live migration methods outside the white list are methods which are currently not allowed to be used for implementing the live migration of the virtual machine; in order to enable the intelligent model to meet the current knowledge (whether safe or appropriate) of the user about each live migration method, the live migration methods used in the setting and training of the intelligent model in the embodiment of the present application only include each live migration method in the white list. In another implementation manner, a corresponding option may be set for each live migration method in the white list, and if the option is in the selected state, it indicates that the live migration method corresponding to the option can be used for implementing the virtual machine live migration, otherwise, it indicates that the live migration method corresponding to the option cannot be used for implementing the virtual machine live migration.
In addition, if a user needs to add or remove a live migration method in the white list, a corresponding instruction may be sent to the virtual machine live migration apparatus at any time, and after receiving the instruction, the virtual machine live migration apparatus adds the live migration method corresponding to the instruction to the white list or removes the live migration method from the white list.
In the virtual machine live migration method provided in the embodiment of the present invention, it is determined that the corresponding live migration method and parameter output by the intelligent model are the target live migration method and the target parameter, respectively, and the method may include:
if the intelligent model outputs a thermal migration method and a thermal migration parameter, determining the thermal migration method and the thermal migration parameter output by the intelligent model as a target thermal migration method and a target parameter respectively; and if the intelligent model outputs a plurality of thermal migration methods and parameters, determining the thermal migration method with the highest priority in the plurality of thermal migration methods and parameters and the corresponding parameter as a target thermal migration method and a target parameter respectively.
The intelligent model learns the samples to be input, and possibly can learn two or more thermal migration methods and parameters, so that the priority can be set for each thermal migration method in advance by a user, and then when the intelligent model outputs two or more thermal migration methods and parameters, the thermal migration method and parameter with the highest priority can be directly selected from the thermal migration methods and parameters, and the thermal migration of the virtual machine can meet the requirements of the user.
The virtual machine live migration method provided by the embodiment of the invention can further comprise the following steps:
and receiving a priority updating instruction input by an external user, and updating the priority of each live migration method and the corresponding parameter based on the priority updating instruction.
It should be noted that, if a user needs to modify the priority of the live migration method, the virtual machine live migration apparatus may update the priority of each live migration method by inputting a corresponding priority update instruction to the virtual machine live migration apparatus, so that the priority of each live migration method may meet the real-time requirement of the user.
In the virtual machine live migration method provided by the embodiment of the invention, external factors can include a service scene, a load type, a bandwidth and configuration of a host machine; the value of the service scenario may include an industry, the value of the load type may include CPU intensive, I/O intensive, memory intensive, network intensive, and hybrid, the value of the bandwidth may include a bandwidth transmission speed and a bandwidth stability, and the configured value of the host may include whether the host in which the corresponding virtual machine is located supports the network card and whether the host supports the hardware compression.
The external factors in the embodiment of the present application may include different service scenarios, different load types (the load type may specifically be a typical load type), different bandwidths, different configurations of hosts (hosts where virtual machines are located, and specifically hosts where virtual machines that need to implement live migration are located), values of the service scenarios may include industries to which the hosts belong (values of which, such as financial industry, new energy industry, etc., the financial industry requires using a lossless compression method and absolute security of a live migration process), different typical load types (values of which, such as CPU intensive type, I/O intensive type, memory intensive type, network intensive type, and hybrid type, etc.), different bandwidths (values of which, such as transmission speed of bandwidth and stability of bandwidth, etc.), and different configurations of the hosts (whether a network card is supported and whether hardware compression is supported, and the hardware compression is data compression and decompression completed by means of hardware resources), the training of the intelligent model and the determination of the heat transfer method are realized through all the external factors, and obviously, the obtained result is more accurate and effective.
The network card can be an RDMA network card, and the application can be set to solidify a zstd compression algorithm (the zstd compression algorithm is a lossless data compression algorithm which is open by Facebook, and zstd has a compression ratio which is almost the same as that of a DEFLAT (zip, gzip) algorithm in design but has higher compression and decompression speed) in hardware, and the hardware supports compression acceleration to release CPU resources.
The virtual machine live migration method provided by the embodiment of the invention obtains an intelligent model by training a training sample and a label value, and comprises the following steps:
and training a pre-established decision tree by using the training samples and the label values to obtain an intelligent model.
Specifically, the intelligent model can be realized by using a Decision Tree, specifically, the Decision Tree, namely, a Decision Tree, is a machine learning supervised classification algorithm; the generation algorithm of the decision tree is ID3, C4.5, CART and the like, and is a tree structure, wherein each internal node represents judgment on an attribute, each branch represents output of a judgment result, and finally each leaf node represents a classification result; however, supervised learning is to give enough training samples, each of which has a set of attributes and a classification result, and then a decision tree model (i.e., an intelligent model) is obtained by learning the training samples, so as to obtain the capability of correctly classifying new data. The migration method can be effectively determined through the decision tree.
The virtual machine live migration method provided in the embodiment of the present invention, after implementing corresponding live migration on a target virtual machine according to the target live migration method based on a target parameter, may further include:
and sending the information of completing the hot migration of the target virtual machine to a corresponding management terminal.
In order to facilitate that a management terminal corresponds to a manager or a user to know that the target virtual machine realizes the live migration, in the embodiment of the present application, after the live migration of the target virtual machine is completed, information of completing the live migration of the target virtual machine is also sent to the corresponding management terminal.
In a specific scenario, a virtual machine live migration method provided in an embodiment of the present invention may include the following steps:
step one, establishing a virtual machine live migration method white list:
before the live migration of the virtual machine is carried out, the cloud platform displays a series of currently supported live migration methods (a multi-thread compression migration method, a post-copy migration method, an XBZRLE migration method, an automatic convergence migration method, an RDMA migration method and the like), and a user can select to add or remove the corresponding live migration method, wherein the removal means that the method cannot be selected when the live migration is realized subsequently; each live migration method may also be prioritized, i.e., which live migration method is scheduled preferentially per virtual machine live migration. The step is only needed to be set once, and can be changed at any time if a user needs the step;
secondly, establishing an intelligent model of the thermal migration method on the virtualization platform:
the method comprises the following steps that (1) total time consumption of different live migration methods and different parameters of the live migration methods in a live migration method white list is tested based on external factors such as different service scenes (for example, the financial industry requires to use a lossless compression mode and absolute safety of a live migration process), different typical load types, different bandwidths (including transmission speed of the bandwidths and stability of the bandwidths), different configurations of hosts (whether a network card supporting RDMA exists or not and whether hardware compression is supported) and the like, and then the live migration method with the shortest total time consumption in a corresponding scene is selected as an optimal method in the scene, so that a training sample is established; specifically, each scene (different values of extrinsic factors) is used as a characteristic attribute of a training sample, and the thermal migration method and the parameter with the shortest total time consumption are label values; establishing a certain amount of training samples and then sending the training samples into a decision tree to obtain an optimal heat transfer intelligent model;
the intelligent model has certain universality, so that the intelligent model can be established once and can be fixed as a heat transfer intelligent module; when the intelligent model is used for carrying out the thermal migration each time, corresponding characteristic data can be recorded for further training, the accuracy of the thermal migration intelligent model is enhanced, and the intelligent model is further trained by utilizing the values of all external factors, the thermal migration method and the parameters when the thermal migration is realized at the current time;
thirdly, determining external factors of the virtual machine:
the bandwidth detection module is used for obtaining real-time network bandwidth (namely bandwidth transmission speed) which can be utilized by virtual machine migration, and the stability of the bandwidth is determined based on the previous running condition (whether network disconnection occurs in a period of time before the current moment, and the like); determining a service scene of the cloud platform operation when the cloud platform is installed and deployed, and subsequently changing the service scene; determining a real-time load type of the virtual machine by using a load detection module; detecting whether a host has available hardware compression resources (a zstd compression algorithm or other compression algorithms are solidified in hardware) and determining whether a network card of the host is a network card supporting RDMA (remote direct memory access);
fourthly, performing thermal migration by using a cloud platform:
and sending the external factors determined based on the steps into an intelligent model, determining a virtual machine live migration method and parameters, and executing the virtual machine live migration. The specific flow is shown in fig. 2, wherein the decision tree model is also an intelligent model.
By the method and the device, the cloud platform can flexibly configure various live migration methods when the live migration service of the virtual machine is carried out, and the total time consumed by live migration is reduced; the generation of the intelligent thermal migration model does not need a user to judge which thermal migration method and parameters are selected according to a specific scene, so that the threshold of a cloud platform user is reduced while thermal migration acceleration is realized, and the influence of thermal migration of the virtual machine on user services is further reduced; the intelligent thermal migration model can select the most appropriate and safe thermal migration method according to the service scene while accelerating the thermal migration process, so that the safety of the thermal migration process is greatly improved. In addition, the specific implementation steps of the method can follow the industrial technical standard or the general technical specification, so that the heat transfer efficiency can be obviously improved, and the heat transfer time can be shortened.
An embodiment of the present invention further provides a virtual machine live migration apparatus, as shown in fig. 3, which may include:
an obtaining module 11, configured to: determining a virtual machine needing to realize the live migration as a target virtual machine, and acquiring values of various external factors influencing the live migration in a scene where the target virtual machine is located as a sample to be input;
a determining module 12 for: inputting a sample to be input into an intelligent model, and determining a corresponding thermal migration method and parameters output by the intelligent model as a target thermal migration method and target parameters respectively; the intelligent model is obtained by utilizing a training sample in advance for training, and the training sample comprises values of various external factors in different scenes and a heat transfer method and parameters with the shortest total time consumption in the corresponding scenes;
a migration module 13 for: and based on the target parameters, implementing corresponding live migration on the target virtual machine according to the target live migration method.
The virtual machine live migration apparatus provided in the embodiment of the present invention may further include:
an operation module to: before the intelligent model is obtained by training with the training samples and the label values, deleting or adding the thermal migration methods in a preset white list under the instruction of a user, and determining each thermal migration method contained in the white list as the thermal migration method contained in the label values.
In an embodiment of the present invention, a determining module of a virtual machine live migration apparatus may include:
a determination unit configured to: if the intelligent model outputs a thermal migration method and a thermal migration parameter, determining the thermal migration method and the thermal migration parameter output by the intelligent model as a target thermal migration method and a target parameter respectively; and if the intelligent model outputs a plurality of thermal migration methods and parameters, determining the thermal migration method with the highest priority in the plurality of thermal migration methods and parameters and the corresponding parameter as a target thermal migration method and a target parameter respectively.
The virtual machine live migration apparatus provided in the embodiment of the present invention may further include:
an update module to: and receiving a priority updating instruction input by an external user, and updating the priority of each live migration method and the corresponding parameter based on the priority updating instruction.
In the virtual machine live migration apparatus provided in the embodiment of the present invention, the external factors may include a service scenario, a load type, a bandwidth, and a configuration of a host; the values of the service scenes comprise industries, the values of the load types comprise CPU intensive type, I/O intensive type, memory intensive type, network intensive type and mixed type, the values of the bandwidth comprise bandwidth transmission speed and bandwidth stability, and the configured values of the host machine comprise whether the host machine where the corresponding virtual machine is located supports a network card and whether hardware compression is supported.
The virtual machine live migration apparatus provided in the embodiment of the present invention may further include:
a sending module configured to: and based on the target parameters, after the target virtual machine is subjected to corresponding thermal migration according to the target thermal migration method, sending information of completing the thermal migration of the target virtual machine to a corresponding management terminal.
The virtual machine live migration apparatus provided in the embodiment of the present invention may further include:
a training module to: and training a pre-established decision tree by using the training samples and the label values to obtain an intelligent model.
An embodiment of the present invention further provides a virtual machine live migration apparatus, which may include:
a memory for storing a computer program;
a processor for implementing the steps of the virtual machine live migration method as described above when executing the computer program.
The embodiment of the invention also provides a computer-readable storage medium, on which a computer program is stored, and when being executed by a processor, the computer program implements the steps of the virtual machine live migration method as described above.
It should be noted that for descriptions of relevant parts in a virtual machine live migration apparatus, a device, and a storage medium provided in the embodiments of the present invention, reference is made to detailed descriptions of corresponding parts in a virtual machine live migration method provided in the embodiments of the present invention, and details are not described herein again. In addition, parts of the technical solutions provided in the embodiments of the present invention that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail, so as to avoid redundant description.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. A method for virtual machine live migration, comprising:
determining a virtual machine needing to realize the live migration as a target virtual machine, and acquiring values of various external factors influencing the live migration in a scene where the target virtual machine is located as a sample to be input;
inputting the sample to be input into an intelligent model, and determining corresponding thermal migration methods and parameters output by the intelligent model as a target thermal migration method and a target parameter respectively; the intelligent model is obtained by utilizing a training sample in advance, and the training sample comprises values of various external factors in different scenes and the heat transfer method and parameters with the shortest total time consumption in the corresponding scene;
and based on the target parameters, implementing corresponding live migration on the target virtual machine according to the target live migration method.
2. The method of claim 1, further comprising, prior to training the intelligent model using training samples and label values:
deleting or adding a live migration method in a preset white list under the instruction of a user, and determining each live migration method contained in the white list as the live migration method contained in the label value.
3. The method of claim 2, wherein determining the respective live migration method and parameter of the intelligent model output as a target live migration method and target parameter, respectively, comprises:
if the intelligent model outputs a thermal migration method and a thermal migration parameter, determining that the thermal migration method and the thermal migration parameter output by the intelligent model are a target thermal migration method and a target parameter respectively; and if the intelligent model outputs a plurality of thermal migration methods and parameters, determining the thermal migration method with the highest priority in the plurality of thermal migration methods and parameters and the corresponding parameter as a target thermal migration method and a target parameter respectively.
4. The method of claim 3, further comprising:
and receiving a priority updating instruction input by an external user, and updating the priority of each live migration method and the corresponding parameter based on the priority updating instruction.
5. The method of claim 4, wherein the extrinsic factors include traffic scenario, load type, bandwidth, and configuration of hosts; the values of the service scenes comprise industries, the values of the load types comprise CPU intensive type, I/O intensive type, memory intensive type, network intensive type and mixed type, the values of the bandwidths comprise bandwidth transmission speed and bandwidth stability, and the configured values of the host machine comprise whether the host machine where the corresponding virtual machine is located supports a network card and whether hardware compression is supported.
6. The method of claim 5, wherein training with training samples and label values yields the intelligent model, comprising:
and training a pre-established decision tree by using the training samples and the label values to obtain the intelligent model.
7. The method of claim 6, wherein after implementing the corresponding live migration for the target virtual machine according to the target live migration method based on the target parameter, further comprising:
and sending the information of completing the hot migration of the target virtual machine to a corresponding management terminal.
8. A virtual machine live migration apparatus, comprising:
an acquisition module to: determining a virtual machine needing to realize the live migration as a target virtual machine, and acquiring values of various external factors influencing the live migration in a scene where the target virtual machine is located as a sample to be input;
a determination module to: inputting the sample to be input into an intelligent model, and determining corresponding thermal migration methods and parameters output by the intelligent model as a target thermal migration method and a target parameter respectively; the intelligent model is obtained by utilizing a training sample in advance, and the training sample comprises values of various external factors in different scenes and the heat transfer method and parameters with the shortest total time consumption in the corresponding scene;
a migration module to: and based on the target parameters, implementing corresponding live migration on the target virtual machine according to the target live migration method.
9. A virtual machine live migration apparatus, comprising:
a memory for storing a computer program;
a processor for implementing the steps of the virtual machine live migration method according to any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, having stored thereon a computer program which, when executed by a processor, carries out the steps of the virtual machine live migration method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011035520.8A CN112181594A (en) | 2020-09-27 | 2020-09-27 | Virtual machine live migration method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011035520.8A CN112181594A (en) | 2020-09-27 | 2020-09-27 | Virtual machine live migration method, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112181594A true CN112181594A (en) | 2021-01-05 |
Family
ID=73944279
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011035520.8A Pending CN112181594A (en) | 2020-09-27 | 2020-09-27 | Virtual machine live migration method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112181594A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112817689A (en) * | 2021-01-08 | 2021-05-18 | 北京金山云网络技术有限公司 | Method and device for sorting virtual machines and electronic equipment |
| CN116560802A (en) * | 2023-07-05 | 2023-08-08 | 麒麟软件有限公司 | Virtual machine load-based virtual machine self-adaptive thermal migration method and system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108595250A (en) * | 2018-05-02 | 2018-09-28 | 南京大学 | A kind of scheduling of resource efficiency optimization method and system towards IaaS cloud platform |
| CN110399253A (en) * | 2019-07-25 | 2019-11-01 | 北京百度网讯科技有限公司 | Delay machine treating method and apparatus |
-
2020
- 2020-09-27 CN CN202011035520.8A patent/CN112181594A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108595250A (en) * | 2018-05-02 | 2018-09-28 | 南京大学 | A kind of scheduling of resource efficiency optimization method and system towards IaaS cloud platform |
| CN110399253A (en) * | 2019-07-25 | 2019-11-01 | 北京百度网讯科技有限公司 | Delay machine treating method and apparatus |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112817689A (en) * | 2021-01-08 | 2021-05-18 | 北京金山云网络技术有限公司 | Method and device for sorting virtual machines and electronic equipment |
| CN116560802A (en) * | 2023-07-05 | 2023-08-08 | 麒麟软件有限公司 | Virtual machine load-based virtual machine self-adaptive thermal migration method and system |
| CN116560802B (en) * | 2023-07-05 | 2023-09-26 | 麒麟软件有限公司 | Virtual machine load-based virtual machine self-adaptive thermal migration method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US12038976B2 (en) | Model pushing method and device, model requesting method and device, storage medium and electronic device | |
| CN112055422B (en) | Method and device for associating 5G signaling with user plane data | |
| CN113704046B (en) | Fault alarm processing method and device, equipment and storage medium | |
| CN116360954B (en) | Industrial Internet of things management and control method and system based on cloud edge cooperative technology | |
| WO2022142927A1 (en) | Rule-based fault analysis method and apparatus, device, and storage medium | |
| CN112988485A (en) | Simulation test method and device for power Internet of things equipment | |
| CN112181594A (en) | Virtual machine live migration method, device, equipment and storage medium | |
| KR102221052B1 (en) | Fault Management System for SDN Network Equipment that supports open flow protocol | |
| US20230097087A1 (en) | Image detection method based on neural network model, electronic device, and storage medium | |
| CN106406980A (en) | Deployment method and device for virtual machines | |
| CN113791792A (en) | Application calling information acquisition method and device and storage medium | |
| CN112787949A (en) | Flow acquisition and transportation management method, control device and storage medium | |
| CN117076937A (en) | Node training method and device of distributed system and distributed system | |
| CN110263251A (en) | A kind of O&M knowledge method for pushing and device based on context model | |
| CN115865682A (en) | SDN link detection processing method, controller, system and medium | |
| CN104516883A (en) | Webpage sharing method, device and system | |
| CN115599494A (en) | Virtual machine migration method and device, upgrading method and server | |
| US10897401B2 (en) | Determining the importance of network devices based on discovered topology, managed endpoints, and activity | |
| CN114629917A (en) | Data processing method and device for cross-system communication and electronic equipment | |
| CN117857457B (en) | Internet of things flow optimization system based on edge calculation | |
| CN110377396A (en) | A kind of virtual machine Autonomic Migration Framework method, system and electronic equipment | |
| CN114253747B (en) | Distributed message management system and method | |
| US20240340225A1 (en) | Machine learning-based analysis of network devices utilization and forecasting | |
| CN115766615B (en) | Cloud-edge-based equipment twin data management method and system | |
| CN110324354B (en) | Method, device and system for network tracking long chain attack |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |