Cloud host migration method, computer storage medium and program product

Abstract

The invention relates to the technical field of cloud computing, and discloses a cloud host migration method, a computer storage medium and a program product.The cloud host migration method comprises the steps that the current resource use state of a host machine and the value degree of a cloud host on the host machine are obtained, and the higher the importance degree of an upper-layer application of the cloud host is, the higher the value degree is; determining a target host machine and a non-target host machine in the host machines based on the current resource use state and the value degree of the cloud host; calculating the business busy degree of the cloud host on the non-target host machine based on the load data of the cloud host on the non-target host machine; determining a to-be-migrated cloud host in the cloud hosts on the non-target host machine based on the current resource use state of the non-target host machine and the value degree and the business busy degree of the cloud hosts on the non-target host machine; and migrating the to-be-migrated cloud host to the target host machine. According to the method, the selection problem of the cloud host is solved, and the influence on service operation can be reduced while load balancing is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of cloud computing, and in particular to a cloud host migration method, a computer storage medium and a program product. Background Art

[0002] Cloud computing is a model that provides basic platforms, infrastructure, and software services. The cloud platform runs multiple cloud hosts based on multiple physical hosts. However, the specifications and number of cloud hosts on each host are often not uniform, which can easily lead to uneven distribution of host resources and tasks, overloading of some hosts, and unbalanced load on the hosts.

[0003] Among them, the virtual machine can perform virtual machine migration from the source host to the target host without shutting down the service, integrating the cloud hosts on different hosts into a few hosts, improving the resource utilization of the host, and when the task distribution is uneven, load balancing can be performed through real-time migration.

[0004] In actual use, especially in scenarios involving important applications, if the host machine is overloaded, selecting a cloud host for migration may affect the business operation of the upper-level application. Therefore, in the face of uneven distribution of host machine resources, how to select a cloud host for migration is the main problem currently faced. Summary of the invention

[0005] In view of this, the present invention provides a cloud host migration method, a computer storage medium and a program product to solve the problem of cloud host selection, which can reduce the impact on business operations while achieving load balancing.

[0006] In a first aspect, the present invention provides a cloud host migration method, which includes: obtaining the current resource usage status of the host machine and the value of the cloud host on the host machine, the higher the importance of the upper-layer application of the cloud host, the higher the value; based on the current resource usage status and the value of the cloud host, determining the target host machine and the non-target host machine in the host machine; based on the load data of the cloud host on the non-target host machine, calculating the business busyness of the cloud host on the non-target host machine; based on the current resource usage status of the non-target host machine, the value and business busyness of the cloud host on the non-target host machine, determining the cloud host to be migrated in the cloud host on the non-target host machine; migrating the cloud host to be migrated to the target host machine.

[0007] In this implementation, according to the importance of the upper-layer applications of the cloud host on the host, the value measurement index of the cloud host is added, and the target host and non-target host are divided. The method of selecting the target host and non-target host provided can reduce the impact on the cloud host business when the cloud host is migrated. At the same time, the business busyness measurement index of the cloud host is added, and the cloud host is screened and migrated in combination with the resource status of the non-target host, so as to avoid problems such as migration failure caused by selecting a cloud host with busy business or a long time for business to reset to zero during the migration process. The method of selecting a cloud host provided by this application can reduce the impact on business operations while achieving load balancing.

[0008] In an optional embodiment, based on the current resource usage status and the value of the cloud host, determining the target host and the non-target host in the host includes: classifying the host based on the current resource usage status to obtain high-load host, medium-load host and low-load host; when the host is a high-load host, determining the host as a non-target host; when the host is a medium-load host, determining the host as a target host; when the host is a low-load host, judging whether there is at least one cloud host on the host whose value is greater than a value threshold; when there is at least one cloud host on the host whose value is greater than the value threshold, determining the host as a target host; when there is no cloud host on the host whose value is greater than the value threshold, determining the host as a non-target host.

[0009] In this implementation, a high-load host is used as a non-target host that needs to be migrated, and a medium-load host is used as a target host that needs to receive migration. Load balancing of high-load hosts can be achieved. For low-load hosts, the value of the cloud host on it is further determined. When the value is high, the host is used as a target host that needs to receive migration. When the value is low, the host is used as a non-target host that needs to be migrated. When the value of the cloud host on the low-load host is low, it can be migrated to save energy. When the value of the cloud host on the low-load host is high, it can be used as a target host to share the load of the high-load host, thereby achieving load balancing and energy saving.

[0010] In an optional embodiment, the current resource usage status includes a resource usage rate, and the host machine is classified based on the current resource usage status to obtain a high-load host machine, a medium-load host machine, and a low-load host machine, including: when the resource usage rate of the host machine is greater than a first resource usage rate threshold, the host machine is classified as a high-load host machine; when the resource usage rate of the host machine is less than or equal to the first resource usage rate threshold, and greater than or equal to the second resource usage rate threshold, the host machine is classified as a medium-load host machine; when the resource usage rate of the host machine is less than the second resource usage rate threshold, the host machine is classified as a low-load host machine.

[0011] In this implementation, the host machines are divided into different types according to their resource utilization rates, so that different types of host machines are divided into different functions. The host machines that need load balancing and energy-saving shutdown can be divided, thereby achieving load balancing and energy saving at the same time.

[0012] In an optional embodiment, the business busyness includes the current business busyness. Based on the load data of the cloud host on the non-target host, calculating the business busyness of the cloud host on the non-target host includes: determining multiple loads of the cloud host on the non-target host, and obtaining the current load utilization and load coefficient of each load; using the load coefficient as a weight to weightedly sum the load utilization of multiple loads to obtain the current business busyness.

[0013] In this implementation, the current service busyness of the cloud host is used as the busyness measurement standard. By combining multiple load states to calculate the current service busyness, multiple factors affecting the cloud host can be comprehensively considered to improve the evaluation accuracy of the cloud host.

[0014] In an optional implementation, the business busyness also includes a business busyness trend. Based on the load data of the cloud host on the non-target host, calculating the business busyness of the cloud host on the non-target host also includes: obtaining the historical business busyness of the cloud host, training a recurrent neural network based on the historical business busyness; predicting the business busyness trend of the cloud host based on the trained recurrent neural network, the business busyness trend including an upward trend, a downward trend, a low peak and a peak.

[0015] In this implementation, the cloud host service busyness trend is used as the busyness measurement standard. Using the neural network prediction model to predict the resource usage trend of the cloud host can improve the trend prediction accuracy and further improve the evaluation accuracy of the cloud host.

[0016] In an optional embodiment, based on the current resource usage status of the non-target host, the value and business busyness of the cloud hosts on the non-target host, determining the cloud hosts to be migrated among the cloud hosts on the non-target host includes: when the non-target host is a high-load host, sorting the cloud hosts based on the value, and selecting a preset number of cloud hosts with the lowest value; judging whether the business busyness of the preset number of cloud hosts is lower than the busyness threshold, and whether the business busyness trend is a downward trend or a low peak; determining the cloud hosts whose business busyness is lower than the busyness threshold and whose business busyness trend is a downward trend or a low peak as cloud hosts to be migrated; when the non-target host is a low-load host, confirming all cloud hosts as cloud hosts to be migrated.

[0017] In this implementation, cloud hosts with low value are given priority for migration to avoid selecting cloud hosts with high value and affecting the operation of high-value businesses. Cloud hosts with idle businesses are then selected for migration to avoid selecting cloud hosts with busy businesses that may cause migration failure or a long time for business to be reset to zero during the migration process.

[0018] In an optional embodiment, migrating the cloud host to be migrated to the target host includes: sorting the target host machines from large to small based on the resource idle rate; for a cloud host to be migrated, judging whether the current target host machine meets the migration conditions according to the arrangement order; if the current target host machine meets the migration conditions, migrating the cloud host to be migrated to the current target host machine; if the current target host machine does not meet the migration conditions, judging whether the next target host machine meets the migration conditions, until the migration of the cloud host to be migrated is completed or the target host machine traversal is completed, then continuing to migrate the next migration cloud host.

[0019] In an optional embodiment, for a cloud host to be migrated, determining whether the current target host meets the migration conditions includes: obtaining the amount of resources to be migrated of the cloud host to be migrated; based on the amount of resources to be migrated, determining whether the post-migration resource utilization of the current target host meets or exceeds a first resource utilization threshold; wherein the post-migration resource utilization is the resource utilization of the current target host after the cloud host to be migrated is migrated to the current target host.

[0020] In this implementation, the resource usage of the target host is calculated after the simulated cloud host migration to evaluate whether the target host is suitable for receiving the cloud host to be migrated, so as to avoid excessive load on the target host after the migration is completed, resulting in frequent migration of the cloud host.

[0021] In a second aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the cloud host migration method of the first aspect or any corresponding embodiment thereof.

[0022] In a third aspect, the present invention provides a computer program product, comprising computer instructions, where the computer instructions are used to enable a computer to execute the cloud host migration method of the first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0024] Figure 1 is a flow chart of a cloud host migration method according to an embodiment of the present invention;

[0025] Figure 2 is a flow chart of another cloud host migration method according to an embodiment of the present invention;

[0026] Figure 3 is a flowchart of determining a target host and a non-target host according to an embodiment of the present invention;

[0027] Figure 4 is a flow chart of predicting business busyness trend by using a GRU model according to an embodiment of the present invention;

[0028] Figure 5 is a schematic diagram of a process for determining a cloud host to be migrated according to an embodiment of the present invention;

[0029] Figure 6 is a schematic diagram of a process of selecting a target host machine for migration according to an embodiment of the present invention;

[0030] Figure 7 is a structural block diagram of a cloud host migration device according to an embodiment of the present invention;

[0031] Figure 8 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0033] A cloud platform is a platform that provides computing, network and storage capabilities based on hardware and software resources. Generally, a cloud platform can have multiple cloud hosts running on multiple hosts. A host can be a hardware device that supports the operation of a cloud host in a cloud platform, and a cloud host can be an independent host virtualized on the basis of a host. A cloud platform can provide users with cloud hosts with different CPU performance and memory performance according to user needs.

[0034] Among them, cloud hosts on the same type of host can be adapted to achieve mutual migration and integration, while cloud hosts on different types of host cannot be adapted. Therefore, the cloud host migration method of the present application is to achieve the migration of cloud hosts on the same type of host.

[0035] Before executing the method of the present application, the host machines are classified according to the host machine types in the cloud platform system, and then the cloud hosts on the host machines are classified. The cloud host migration method of the present application is executed for each type of host machine and the cloud host on the host machine.

[0036] Specifically, according to an embodiment of the present invention, a cloud host migration method embodiment is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0037] In this embodiment, a cloud host migration method is provided. Figure 1 is a flow chart of a cloud host migration method according to an embodiment of the present invention. It should be noted that if there are substantially the same results, this embodiment does not use Figure 1 The process sequence shown is limited. Figure 1 As shown, the process includes the following steps:

[0038] Step S101, obtaining the current resource usage status of the host machine and the value of the cloud host on the host machine.

[0039] The current resource usage status includes the total resource amount, the reserved resource amount and the used resource amount.

[0040] In order to ensure the safety of the host operation, not all resources on the host can be used for the operation of the cloud host business, and some resources are extracted as reserved resources.

[0041] The used resources are the amount of resources used by the current business operation of the cloud host on the host.

[0042] For hosts of the same type, obtain the total resources, reserved resources, and used resources of the host.

[0043] Different cloud hosts run different businesses, and the importance of the business will affect the value of the cloud host. The more important the upper-layer application of the cloud host is, the higher its value is.

[0044] In one implementation, the importance of the upper-layer application of the cloud host is quantified. The higher the importance, the higher the importance value. The importance data is used as the value of the cloud host.

[0045] Step S102, based on the current resource usage status and the value of the cloud host, determine the target host and non-target host in the host.

[0046] The target host is the host that receives the migrated cloud host, and the non-target host is the host that sends the migrated cloud host.

[0047] First, a non-target host is determined according to the current resource usage state. Specifically, when the host is determined to be overloaded according to the current resource usage state of the host, the host is used as a non-target host.

[0048] Secondly, the target host and non-target host are further determined based on the current resource usage status and the value of the cloud host. Specifically, when the host is judged to be underloaded based on the current resource usage status of the host, and the value of the cloud host on the host is high, the migration impact is large, then the host is used as the target host; when the host is judged to be underloaded based on the current resource usage status of the host, and the value of the cloud host on the host is low, the migration impact is small, then the host is used as a non-target host.

[0049] Finally, other hosts are used as target hosts.

[0050] Step S103, based on the load data of the cloud host on the non-target host, the service busyness of the cloud host on the non-target host is calculated.

[0051] The load data includes CPU (Central Processing Unit) data, memory data, disk data, network data, etc.

[0052] Load data characterizes the resource usage status of the cloud host when running business tasks. Determine the load data used by the cloud host, quantify the load data, and calculate the business busyness of the cloud host by combining multiple load data.

[0053] The higher the business busyness, the greater the impact of cloud host migration on the current non-target host.

[0054] Step S104 , based on the current resource usage status of the non-target host machine, the value and business busyness of the cloud host on the non-target host machine, determine the cloud host to be migrated from the cloud hosts on the non-target host machine.

[0055] When the non-target host is overloaded, the higher the value of the cloud host, the greater the impact of cloud host migration; the higher the business busyness of the cloud host, the greater the impact of cloud host migration. Select some cloud hosts with low value and low business busyness as cloud hosts to be migrated.

[0056] When the non-target host is underloaded, according to the above step S104, the value of the cloud host is low, and all cloud hosts are used as cloud hosts to be migrated.

[0057] Step S105, migrating the cloud host to be migrated to the target host machine.

[0058] For the cloud host to be migrated, select the appropriate target host for migration.

[0059] The cloud host migration method provided in this embodiment increases the value measurement index of the cloud host according to the importance of the upper-layer applications of the cloud host on the host, divides the target host and the non-target host, and provides a method for selecting the target host and the non-target host, which can reduce the impact on the cloud host business when performing cloud host migration. At the same time, the business busyness measurement index of the cloud host is increased, and the cloud host is screened and migrated in combination with the resource status of the non-target host, so as to avoid problems such as migration failure caused by selecting a cloud host with busy business or a long time for business to reset to zero during the migration process. The method for selecting a cloud host provided in this application can reduce the impact on business operations while achieving load balancing.

[0060] In this embodiment, a cloud host migration method is provided, which can be used for the above-mentioned mobile terminals, such as mobile phones, tablet computers, etc. Figure 2 is a flowchart of another cloud host migration method according to an embodiment of the present invention. It should be noted that if there are substantially the same results, this embodiment does not use Figure 2 The process sequence shown is limited. Figure 2 As shown, the process includes the following steps:

[0061] Step S201, obtaining the current resource usage status of the host machine and the value of the cloud host on the host machine.

[0062] The current resource usage status of the host includes the total resource amount, the reserved resource amount, the used resource amount and the resource usage rate. The resource usage rate is calculated using the total resource amount, the reserved resource amount and the used resource amount.

[0063] In one implementation, the resource usage rate calculation formula is:

[0064]

[0065] Among them, R util is the resource utilization rate, R total is the total resources, R reserve is the reserved resource amount, R used is the amount of resources used. Resource utilization rate R util ∈[0,1].

[0066] The value of the cloud host is set according to the importance of the upper-layer applications of the cloud host. The higher the importance of the upper-layer applications of the cloud host, the higher the value.

[0067] Understandably, the higher the value of the cloud host, the greater the impact on the operation of upper-layer applications during migration.

[0068] Step S202 , based on the current resource usage status and the value of the cloud host, determine the target host and non-target host in the host.

[0069] Specifically, the above step S202 includes:

[0070] Step S2021, classifying the host machines based on the current resource usage status to obtain high-load host machines, medium-load host machines and low-load host machines.

[0071] Sort the hosts from small to large according to resource usage, and we get:

[0072]

[0073] In one implementation, a first resource usage threshold and a second resource usage threshold are set, wherein the first resource usage threshold is greater than the second resource usage threshold, and the host machine is divided into a high-load host machine, a medium-load host machine, and a low-load host machine according to the first resource usage threshold and the second resource usage threshold.

[0074] Generally, the first resource usage threshold is 80%-90%, and the second resource usage threshold is 20%-30%. In actual applications, the first resource usage threshold and the second resource usage threshold are adjusted according to the operating status of the host machine. If the current scenario has high operating requirements for the host machine, the first resource usage threshold is appropriately lowered.

[0075] Specifically, a host machine whose resource usage rate is greater than a first resource usage rate threshold is classified as a high-load host machine; a host machine whose resource usage rate is less than or equal to the first resource usage rate threshold and greater than or equal to a second resource usage rate threshold is classified as a medium-load host machine; and a host machine whose resource usage rate is less than the second resource usage rate threshold is classified as a low-load host machine.

[0076] For example, the first resource usage threshold is 80%, and the second resource usage threshold is 30%. A host machine with a resource usage greater than 80% is classified as a high-load host machine; a host machine with a resource usage less than or equal to 80% and greater than or equal to 30% is classified as a medium-load host machine; a host machine with a resource usage less than 30% is classified as a low-load host machine.

[0077] Please refer to Figure 3 , Figure 3 The flowchart of determining the target host machine and the non-target host machine according to an embodiment of the present invention is as follows: According to the resource usage status of the host machine, the host machine is divided into a high-load host machine, a medium-load host machine and a low-load host machine.

[0078] In another implementation, the host machines are divided into high-load host machines, medium-load host machines, and low-load host machines according to a ratio.

[0079] Specifically, the host machines in the first proportion that are ranked higher are classified as low-load host machines, the host machines in the second proportion that are ranked lower are classified as high-load host machines, and the other host machines are classified as medium-load host machines.

[0080] Exemplarily, the first 20% of host machines are classified as low-load host machines, the last 15% of host machines are classified as high-load host machines, and the other host machines are classified as medium-load host machines.

[0081] Step S2022, based on the value of the cloud host, determine the target host and non-target host from the high-load host, the medium-load host, and the low-load host.

[0082] Please continue to combine Figure 3 , respectively determine the target host and non-target host among the high-load host, the medium-load host and the low-load host.

[0083] When the host is a high-load host, it is determined that the host needs to be load balanced and is added to the list of hosts that need load balancing, that is, the host is determined to be a non-target host and is added to the non-target host list.

[0084] When the host machine is a medium-load host machine, the host machine is determined to be a target host machine and is added to the target host machine list.

[0085] When the host is a low-load host, it is determined that the host needs to be shut down for energy conservation and added to the list of hosts that need to be shut down for energy conservation. Further, it is determined whether there is at least one cloud host on the host whose value is greater than the value threshold. When there is at least one cloud host on the host whose value is greater than the value threshold, the host is determined to be a target host and added to the list of target hosts; when there is no cloud host on the host whose value is greater than the value threshold, the host is determined to be a non-target host and added to the list of non-target hosts.

[0086] Traverse all hosts. When the non-target host list is empty, it means there is no host that needs resource adjustment. When the target host list is empty, it means there is no host that needs auxiliary resource adjustment.

[0087] When the non-target host list and the target host are not empty, for the host in the non-target host list, the method of step S204 is used to select a cloud host from the non-target host to migrate to the target host.

[0088] Among them, the host machines are divided into different types of host machines according to their resource utilization rates, so as to perform different functional divisions on different types of host machines, and can divide the host machines that need load balancing and those that need to be shut down to save energy, thereby achieving load balancing and energy saving at the same time.

[0089] At the same time, the high-load host is regarded as a non-target host that needs to be migrated, and the medium-load host is regarded as a target host that needs to receive the migration. Load balancing can be achieved for the high-load host. For the low-load host, the value of the cloud host on it is further judged. When the value is high, the host is regarded as the target host that needs to receive the migration. When the value is low, the host is regarded as a non-target host that needs to be migrated. When the value of the cloud host on the low-load host is low, it can be migrated to save energy; when the value of the cloud host on the low-load host is high, it can be used as a target host to share the load of the high-load host, and load balancing and energy saving can be achieved at the same time.

[0090] Step S203, based on the load data of the cloud host on the non-target host, the service busyness of the cloud host on the non-target host is calculated.

[0091] The business busyness includes the current business busyness and the business busyness trend.

[0092] The current business busyness is used to characterize the current business status of the cloud host, and the business busyness area is used to characterize the future business status trend of the cloud host.

[0093] Specifically, the above step S203 includes:

[0094] Step S2031, based on the load data of the cloud host on the non-target host, calculate the current service busyness of the cloud host on the non-target host.

[0095] Determine multiple loads of the cloud host on the non-target host. The loads include CPU, memory, disk, network, etc. Obtain the load data of each load, where the load data includes load utilization and load factor. Calculate the current business busyness based on the current load utilization and load factor of each load.

[0096] Specifically, the current load utilization and load coefficient of each load are obtained, and the load coefficient is used as a weight to perform weighted summation on the load utilizations of multiple loads to obtain the current service busyness.

[0097] Exemplarily, CPU utilization, memory utilization, disk utilization, and network utilization are obtained, and the correlation coefficient of CPU utilization, memory utilization, disk utilization, and network utilization are obtained, and the current service busyness is calculated as:

[0098] U vm =α*U cpu +β*U mem +γ*U disk +δ*U net .

[0099] Among them, U vm is the current business busyness of the cloud host VM, U cpu is the CPU utilization, U mem is the memory utilization, U disk is the disk utilization, U net is the network utilization, α, β, γ, δ are the correlation coefficients of CPU utilization, memory utilization, disk utilization, and network utilization, respectively. Among them, α≥0; β≥0; γ≥0; δ≥0.

[0100] Among them, α, β, γ, and δ can be set based on experience or according to the importance of CPU, memory, disk, and network to the current business operation. For example, if the current business operation has high requirements for memory, the value of β can be appropriately increased. In order to make the current business busyness of all cloud hosts a unified dimension, set α+β+γ+δ=1.

[0101] In this implementation, the current service busyness of the cloud host is used as the busyness measurement standard. By combining multiple load states to calculate the current service busyness, multiple factors affecting the cloud host can be comprehensively considered to improve the evaluation accuracy of the cloud host.

[0102] Step S2032: Calculate the service busyness trend of the cloud host on the non-target host based on the load data of the cloud host on the non-target host.

[0103] Generally, the load status of a cloud host shows regularity, and the business busyness of a cloud host is related to time, so the future busyness can be predicted based on the historical busyness.

[0104] For example, if the cloud host is in a relatively busy period from 07:00 to 24:00 and in a relatively low-load period from 00:00 to 07:00, a startup storm is likely to occur in the period from 08:00 to 09:00.

[0105] Specifically, a prediction is performed for each cloud host, the historical service busyness of the cloud host is obtained, a recurrent neural network is trained based on the historical service busyness, and the service busyness trend of the cloud host is predicted based on the trained recurrent neural network.

[0106] The business busyness trend includes an upward trend, a downward trend, a low peak, and a peak.

[0107] In one implementation, the recurrent neural network uses an optimized version of the long short-term memory neural network for prediction, that is, a GRU (Gated Recurrent Unit networks) model is used for prediction.

[0108] See also Figure 4 , Figure 4 FIG. 1 is a flow chart of using a GRU model to predict a business busyness trend according to an embodiment of the present invention. Figure 4 As shown, using the GRU model to predict the business busyness trend includes two parts: data processing and data prediction.

[0109] Get historical data. Historical data includes historical resource usage data and load data of cloud hosts.

[0110] The historical data is processed in time series. The load status of the cloud host shows regularity, and the historical data is arranged in time series to facilitate subsequent model training and prediction.

[0111] Data cleaning. Detect and remove erroneous data, duplicate data, and abnormal data in historical data, and normalize the removed data to obtain a data set. Perform model training and prediction based on the data set.

[0112] Initialize the GRU model. Split the data set into a training data set and a prediction data set. The training data set is used to train the initial GRU model, and the prediction data set is used to predict the business busyness trend. Determine the initial network structure, weights, and threshold parameters of the GRU model. Obtain the trained GRU model.

[0113] GRU model training. Use the initial GRU model to make forward predictions on the training data set to obtain the prediction results, and calculate the error between the prediction results and the actual results to determine whether the error meets the accuracy requirements. If the error does not meet the accuracy requirements, adjust the relevant parameters of the initial GRU model and make predictions again. If the accuracy requirements are met, the trained GRU model is obtained.

[0114] GRU model prediction: Use the trained GRU model to make forward predictions on the prediction data set to obtain the business busyness trend.

[0115] In this implementation, the cloud host service busyness trend is used as the busyness measurement standard. Using the neural network prediction model to predict the resource usage trend of the cloud host can improve the trend prediction accuracy and further improve the evaluation accuracy of the cloud host.

[0116] Step S204: determining a cloud host to be migrated from the cloud hosts on the non-target host machines based on the value and service busyness of the cloud hosts on the non-target host machines.

[0117] See also Figure 5 , Figure 5 The flowchart of determining a cloud host to be migrated according to an embodiment of the present invention is as follows: a non-target host is obtained from a non-target host list, and it is determined whether the non-target host is a high-load host or a low-load host.

[0118] When the non-target host is a high-load host, the cloud hosts are sorted according to their values ​​on the non-target host, and the N cloud hosts with the lowest values ​​are selected to determine whether the business busyness of the N cloud hosts is lower than the busyness threshold and whether the business busyness trend is a downward trend or a low peak. The cloud hosts whose business busyness is lower than the busyness threshold and whose business busyness trend is a downward trend or a low peak are determined as cloud hosts to be migrated and added to the list of cloud hosts to be migrated.

[0119] When the non-target host is a low-load host, all cloud hosts are confirmed as cloud hosts to be migrated and added to the list of cloud hosts to be migrated.

[0120] This application selects cloud hosts with idle services for migration, avoiding selecting cloud hosts with busy services that may cause migration failure or long service reset time during migration, etc. Select cloud hosts with low value for migration, avoiding selecting cloud hosts with high value that may affect the operation of high-value services.

[0121] Step S205: Migrate the cloud host to be migrated to the target host machine.

[0122] For the cloud hosts to be migrated in the list of cloud hosts to be migrated, select a suitable target host for migration.

[0123] Specifically, see Figure 6 , Figure 6 The figure is a schematic diagram of a process of selecting a target host machine for migration according to an embodiment of the present invention.

[0124] Sort by the current resource usage status of the target host, and give priority to idle target hosts for migration.

[0125] The target host machines are sorted from large to small based on resource idle rate or from small to large based on resource utilization rate, and the target host machines are filtered in order of arrangement.

[0126] For each cloud host to be migrated, simulate the resource usage status of the target host after the cloud host to be migrated is migrated to the target host in the order of arrangement. Specifically, determine whether the current target host meets the migration conditions of the cloud host to be migrated in the order of arrangement. If the current target host meets the migration conditions, migrate the cloud host to the current target host; if the current target host does not meet the migration conditions, determine whether the next target host meets the migration conditions. Until the migration of the cloud host to be migrated is completed or the target host traversal is completed, if there is no target host that meets the migration conditions, continue to migrate the next migration cloud host.

[0127] In one implementation, a cloud host to be migrated is randomly selected from the list of cloud hosts to be migrated, the amount of resources to be migrated of the cloud host to be migrated is obtained, and the post-migration resource utilization rate of the current target host after the cloud host to be migrated is calculated according to the arrangement order. The post-migration resource utilization rate is:

[0128]

[0129] Among them, R util_new is the resource utilization rate after migration, R vm The amount of resources to be migrated of the cloud host VM to be migrated.

[0130] Determine whether the resource usage rate of the current target host after migration is less than or equal to the first resource usage rate threshold. When the resource usage rate after migration is less than or equal to the first resource usage rate threshold, it means that the impact of this migration on the resource occupancy of the current target host is controllable, and the current target host meets the migration conditions, and the cloud host to be migrated is migrated to the current target host; when the resource usage rate after migration is greater than the first resource usage rate threshold, it means that after this migration, the target host becomes a high-load host, and the current target host does not meet the migration conditions, and the next target host is selected.

[0131] By using the above method, migration judgment and migration operation are performed on the cloud host to be migrated in a loop until the migration of the cloud host to be migrated is completed or the traversal of the target host is completed.

[0132] In another implementation, the amount of resources to be migrated of the cloud hosts to be migrated is obtained, the cloud hosts to be migrated are sorted according to the amount of resources to be migrated, and the migration operation is performed preferentially on the cloud hosts to be migrated with large amounts of resources to be migrated.

[0133] Furthermore, after all the cloud hosts to be migrated in the non-target host machines that are low-load hosts are migrated, the host machines are shut down.

[0134] The cloud host migration method provided in this embodiment calculates the resource usage of the target host machine after simulating the cloud host migration, and evaluates whether the target host machine is suitable for receiving the cloud host to be migrated, so as to avoid excessive load on the target host machine after the migration is completed, resulting in frequent migration of the cloud host.

[0135] This application selects cloud hosts for real-time migration based on the value, business busyness and busyness trend of cloud hosts to achieve load balancing and energy saving of hosts, preventing some hosts from being continuously overloaded and affecting cloud host services, or some hosts from being continuously underloaded and wasting energy. This solves the problem of selecting cloud hosts and target hosts, and can reduce the impact on business operations while achieving load balancing and energy saving.

[0136] In this embodiment, a cloud host migration device is also provided, which is used to implement the above embodiments and preferred implementations, and will not be repeated here. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0137] This embodiment provides a cloud host migration device, such as Figure 7 As shown, Figure 7 is a structural block diagram of a cloud host migration device according to an embodiment of the present invention, the cloud host migration device comprising:

[0138] The acquisition module 701 is used to acquire the current resource usage status of the host machine and the value of the cloud host on the host machine. The higher the importance of the upper layer application of the cloud host, the higher the value.

[0139] The first determination module 702 is used to determine a target host machine and a non-target host machine among the host machines based on the current resource usage status and the value of the cloud host.

[0140] The calculation module 703 is used to calculate the service busyness of the cloud host on the non-target host based on the load data of the cloud host on the non-target host.

[0141] The second determination module 704 is used to determine the cloud host to be migrated from the cloud hosts on the non-target host based on the current resource usage status of the non-target host, the value and business busyness of the cloud host on the non-target host.

[0142] The migration module 705 is used to migrate the cloud host to be migrated to the target host.

[0143] The further functional description of each of the above modules is the same as that of the above corresponding embodiments and will not be repeated here.

[0144] The cloud host migration device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0145] The embodiment of the present invention also provides a computer device having the above Figure 7 The cloud host migration device is shown.

[0146] See also Figure 8 , Figure 8 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Figure 8 As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 8 A processor 10 is taken as an example.

[0147] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0148] The memory 20 stores instructions executable by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0149] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0150] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0151] The computer device also includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 may be connected via a bus or other means. Figure 8 The example of connecting through bus is taken in the following.

[0152] The input device 30 can receive input digital or character information, and generate key signal input related to the user settings and function control of the computer device, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator bar, one or more mouse buttons, a trackball, a joystick, etc. The output device 40 may include a display device, an auxiliary lighting device (e.g., an LED) and a tactile feedback device (e.g., a vibration motor), etc. The above-mentioned display device includes but is not limited to a liquid crystal display, a light emitting diode, a display and a plasma display. In some optional embodiments, the display device can be a touch screen.

[0153] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0154] A part of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the existence of the computer program instruction in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc., and accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium accessible to the computer.

[0155] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A cloud host migration method, characterized in that: The method comprises: Obtaining the current resource usage status of the host machine and the value of the cloud host on the host machine, wherein the higher the importance of the upper layer application of the cloud host, the higher the value; Determining a target host and a non-target host among the host machines based on the current resource usage status and the value of the cloud host; Calculating the service busyness of the cloud host on the non-target host based on the load data of the cloud host on the non-target host; Determine a cloud host to be migrated from the cloud hosts on the non-target host based on the current resource usage status of the non-target host, the value of the cloud hosts on the non-target host, and the business busyness; Migrate the cloud host to be migrated to the target host machine.

2. The cloud host migration method according to claim 1, characterized in that: The determining of the target host and the non-target host in the host based on the current resource usage status and the value of the cloud host includes: Classifying the host machines based on the current resource usage status to obtain high-load host machines, medium-load host machines, and low-load host machines; When the host machine is the high-load host machine, determining that the host machine is a non-target host machine; When the host machine is the medium-load host machine, determining the host machine as the target host machine; When the host machine is the low-load host machine, determining whether there is at least one cloud host on the host machine whose value is greater than a value threshold; When there is at least one cloud host on the host machine whose value is greater than the value threshold, determining the host machine as a target host machine; When the value of the cloud host that does not exist on the host machine is greater than the value threshold, the host machine is determined to be a non-target host machine.

3. The cloud host migration method according to claim 2, characterized in that: The current resource usage status includes a resource usage rate, and the classifying the host machine based on the current resource usage status to obtain a high-load host machine, a medium-load host machine, and a low-load host machine includes: When the resource usage rate of the host machine is greater than a first resource usage rate threshold, classifying the host machine as the high-load host machine; When the resource usage rate of the host machine is less than or equal to the first resource usage rate threshold and greater than or equal to the second resource usage rate threshold, classifying the host machine as the medium-load host machine; When the resource usage rate of the host machine is less than the second resource usage rate threshold, the host machine is classified as the low-load host machine.

4. The cloud host migration method according to claim 2, characterized in that: The business busyness includes the current business busyness, and the calculating the business busyness of the cloud host on the non-target host based on the load data of the cloud host on the non-target host includes: Determine multiple loads of the cloud host on the non-target host, and obtain the current load utilization and load factor of each load; The load coefficient is used as a weight to perform a weighted summation on the load utilization rates of multiple loads to obtain the current service busyness.

5. The cloud host migration method according to claim 4, characterized in that: The business busyness also includes a business busyness trend. The calculating the business busyness of the cloud host on the non-target host based on the load data of the cloud host on the non-target host also includes: Acquire the historical business busyness of the cloud host, and train a recurrent neural network based on the historical business busyness; The business busyness trend of the cloud host is predicted based on the trained recurrent neural network, and the business busyness trend includes an upward trend, a downward trend, a low peak and a peak.

6. The cloud host migration method according to claim 5, characterized in that: The determining of a cloud host to be migrated from the cloud hosts on the non-target host based on the current resource usage status of the non-target host, the value of the cloud hosts on the non-target host, and the business busyness includes: When the non-target host is the high-load host, sorting the cloud hosts based on the value, and selecting a preset number of cloud hosts with the lowest value; Determine whether the service busyness of a preset number of the cloud hosts is lower than a busyness threshold, and whether the service busyness trend is a downward trend or a low peak; Determine the cloud host whose service busyness is lower than the busyness threshold and whose service busyness trend is a downward trend or a low peak as the cloud host to be migrated; When the non-target host machine is the low-load host machine, all the cloud hosts are confirmed as the cloud hosts to be migrated.

7. The cloud host migration method according to claim 3, characterized in that: The migrating the cloud host to be migrated to the target host includes: Sort the target host machines from large to small based on resource idle rates; For a cloud host to be migrated, determine whether the current target host meets the migration conditions according to the arrangement order; If the current target host machine meets the migration condition, migrating the cloud host to be migrated to the current target host machine; If the current target host does not meet the migration conditions, it is determined whether the next target host meets the migration conditions, until the migration of the cloud host to be migrated is completed or the traversal of the target host is completed, and then the migration of the next migration cloud host is continued.

8. The cloud host migration method according to claim 7, characterized in that: For a cloud host to be migrated, determining whether the current target host meets the migration condition includes: Obtaining the amount of resources to be migrated of the cloud host to be migrated; Determining whether the post-migration resource usage rate of the current target host machine satisfies or is less than or equal to the first resource usage rate threshold based on the amount of resources to be migrated; The post-migration resource usage rate is the resource usage rate of the current target host machine after the cloud host to be migrated is migrated to the current target host machine.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the cloud host migration method according to any one of claims 1 to 8.

10. A computer program product, characterized in that The method comprises computer instructions, wherein the computer instructions are used to cause a computer to execute the cloud host migration method according to any one of claims 1 to 8.

Images