Virtual machine scheduling processing method, system and storage medium with priority on remaining cloud storage capacity

Abstract

The present invention discloses a virtual machine scheduling processing method, system and storage medium with cloud storage remaining capacity priority. According to the running data volume and expected data storage volume of the virtual machine, combined with the building instructions received in real time, a first screening is first performed to obtain multiple candidate physical machines with more suitable memory remaining capacity for scheduling and building the virtual machine; the remaining capacity of the physical machine of the existing virtual machine can be used as much as possible by using the scheduling strategy with remaining capacity priority; at the same time, in the subsequent screening operation, a second screening is performed according to the actual situation of the target virtual machine and the candidate physical machine to obtain the target physical machine; thus, a comprehensive analysis is performed with the remaining memory capacity of the physical machine being prioritized, supplemented by the performance and bandwidth of the physical machine, so as to quickly and flexibly select the target physical machine of the target virtual machine.

Description

Technical Field

[0001] The present invention relates to the field of virtual machine scheduling, and in particular to a virtual machine scheduling processing method, system and storage medium with priority given to remaining capacity of cloud storage. Background Art

[0002] Virtualization technology can abstract physical resources into virtual resources, thereby improving the utilization of computing resources and the flexibility of resource management. Virtual machine scheduling and allocation processing is a crucial part of cloud computing. Its core goal is to effectively allocate virtual machines to physical machines to meet the needs of applications and maximize the utilization of physical machines. However, in practical applications, virtual machine scheduling and allocation processing faces many challenges and difficulties.

[0003] First, virtual machine scheduling and allocation often adopts random scheduling and allocation, which will cause some physical machines to have too much data, resulting in poor performance of the physical machines that are allocated. At the same time, some physical machines will not have virtual machines allocated, resulting in waste of resources of the physical machines that are running idle.

[0004] Secondly, the virtual machine scheduling and allocation process needs to quickly and accurately select the most suitable physical machine and schedule the virtual machine to the physical machine. Due to the large number and attribute information of physical machines, how to quickly select the optimal physical machine is also a difficult point.

[0005] In addition, as the scale of cloud computing platforms grows, management costs and operation and maintenance costs become increasingly high, how to achieve automated and intelligent virtual machine scheduling and allocation processing is also an urgent issue. It is necessary to study and design more advanced and efficient virtual machine scheduling and allocation processing algorithms and technologies to meet the ever-changing business needs and application scenarios. Summary of the invention

[0006] The purpose of the present invention is to provide a virtual machine scheduling processing method, system and storage medium with priority on remaining capacity of cloud storage, which solves the above-mentioned technical problems pointed out in the prior art.

[0007] The present invention proposes a virtual machine scheduling processing method with priority on remaining capacity of cloud storage, comprising the following steps:

[0008] Receive a build instruction of a target virtual machine; the build instruction includes the running data volume of the target virtual machine and the estimated data storage volume corresponding to the running data volume;

[0009] According to the building instruction, a plurality of candidate physical machines that meet the basic data storage capacity requirement are screened and obtained from the physical machine group;

[0010] Acquire the attribute information of each of the candidate physical machines; perform comprehensive analysis and screening to acquire the target physical machine according to the running data volume of the target virtual machine, the estimated data storage volume corresponding to the running data volume, and the attribute information of the candidate physical machines;

[0011] The target virtual machine is scheduled to the target physical machine, and a virtual machine running operation is performed.

[0012] Preferably, the attribute information of the physical machine to be selected includes the serial number information of the physical machine to be selected, the data processing rate of the physical machine to be selected, and the network bandwidth value of the physical machine to be selected;

[0013] Preferably, the step of performing comprehensive analysis and screening to obtain the target physical machine according to the running data volume of the target virtual machine, the estimated data storage volume corresponding to the running data volume, and the attribute information of the candidate physical machine comprises the following steps:

[0014] Normalizing the attribute information of each of the candidate physical machines to obtain normalized attribute information of the candidate physical machines, and establishing a normalized attribute information set of the candidate physical machines according to the normalized attribute information of all the candidate physical machines;

[0015] Acquire attribute information of multiple evaluation experts and initial evaluation scores f of each of the evaluation experts for each of the candidate physical machines; analyze and screen the target physical machine based on the initial evaluation scores and the normalized attribute information set of the candidate physical machines;

[0016] Preferably, the operation of normalizing the attribute information of the physical machine to be selected to obtain the normalized attribute information of the physical machine to be selected is specifically:

[0017] Acquire the data processing rate Y of each of the candidate physical machines; normalize the data processing rate to obtain the normalized data processing rate Y' of each of the candidate physical machines;

[0018] The normalized data processing rate Y' is calculated as follows:

[0019] Where, Y is the data processing rate of the physical machine to be selected; n is the number of physical machines to be selected;

[0020] Acquire a network bandwidth value X of each of the candidate physical machines; normalize the network bandwidth value X to obtain a normalized network bandwidth value X' of each of the candidate physical machines;

[0021] The normalized network bandwidth value X' is calculated as follows:

[0022] In the formula, X is the network bandwidth value of the physical machine to be selected; n is the number of physical machines to be selected;

[0023] According to the normalized data processing rate Y' and the normalized network bandwidth value X', normalized attribute information di (Y', X') of the candidate physical machine is obtained; wherein di is the i-th candidate physical machine;

[0024] Preferably, the analyzing and screening the target physical machine according to the initial evaluation score and the normalized attribute information set of the candidate physical machine comprises the following steps:

[0025] A performance weight wy is preset for the normalized data processing rate Y'; and a bandwidth weight wx is preset for the normalized network bandwidth value X';

[0026] Obtaining the evaluation weight wz of each evaluation expert; calculating and obtaining the adoptable evaluation score fz of the initial evaluation score according to the evaluation weight;

[0027] Calculate and obtain a second evaluation score f' according to the adoptable evaluation score fz, the performance weight wy, and the bandwidth weight wx;

[0028] Set iteration parameters; the iteration parameters include fitness calculation parameters λ; traverse each of the candidate physical machines, and calculate the fitness F of the candidate physical machine according to the second evaluation score and the weighted value of the basic attribute evaluation of the candidate physical machine and the iteration parameters; determine whether the fitness F is greater than or equal to the preset minimum fitness threshold; if so, record the fitness F to obtain a fitness list; if not, clear the fitness list, and adjust the iteration parameters to obtain new iteration parameters, and repeat the above operation using the new iteration parameters until the fitness of all the candidate physical machines is greater than or equal to the preset minimum fitness threshold, and a fitness list is obtained;

[0029] The candidate physical machine with the highest fitness in the fitness list is selected as the target physical machine.

[0030] Preferably, the calculation method of the acceptable evaluation score fz is:

[0031] fz=f×wz;

[0032] The second evaluation score f' is calculated as follows:

[0033] f'=fz×((Y'×wy)+(X'×wx));

[0034] The fitness F is calculated as follows:

[0035]

[0036] Where, ((Y'×wy)+(X'×wx)) is the weighted value of the basic attribute evaluation of the candidate physical machine; p is the standard deviation of the weighted values ​​of the basic attribute evaluation of all candidate physical machines; f i ' is the second evaluation score of the i-th candidate physical machine;

[0037] Preferably, the iterative parameter is adjusted to obtain a new iterative parameter by subtracting a preset iterative parameter adjustment threshold β from the fitness calculation parameter λ to obtain a new fitness calculation parameter;

[0038] The calculation formula of the new fitness calculation parameter is: λ-β;

[0039] Preferably, the step of obtaining the evaluation weight wz of each evaluation expert comprises the following steps:

[0040] Acquire multiple analysis dimensions K of the evaluation weights of the evaluation experts; the analysis dimensions can be expressed as K={k1, k2, k3, ...km}; where km is the mth analysis dimension;

[0041] Obtain the evaluation score Lz of each evaluation expert in each analysis dimension; the evaluation score can be expressed as L={l1, l2, l3, ...ln}; wherein ln represents the evaluation score of the nth evaluation expert;

[0042] Establish an expert evaluation matrix R according to the analysis dimension K and the evaluation score L;

[0043] The expert evaluation matrix R can be expressed as:

[0044]

[0045] Among them, kmln represents the evaluation score of the nth evaluation expert in the mth evaluation dimension;

[0046] Preset the evaluation weight wu of each evaluation dimension;

[0047] The evaluation weight wz of each evaluation expert is calculated based on the evaluation weight wu and the expert evaluation matrix R;

[0048] The evaluation weight wz is calculated as follows:

[0049]

[0050] Accordingly, the present invention also provides a virtual machine scheduling processing system with a priority on remaining capacity, comprising a receiving module, a first screening module, a second screening module and a calling-in operation module;

[0051] The receiving module is used to receive a construction instruction of a target virtual machine; the construction instruction includes the running data volume of the target virtual machine and the estimated data storage volume corresponding to the running data volume;

[0052] The first screening module is used to screen and obtain multiple candidate physical machines that meet the basic data storage capacity requirement from the physical machine group according to the building instruction;

[0053] The second screening module is used to obtain the attribute information of each of the candidate physical machines; and to perform comprehensive analysis and screening to obtain the target physical machine according to the running data volume of the target virtual machine, the estimated data storage volume corresponding to the running data volume, and the attribute information of the candidate physical machines;

[0054] The calling-in operation module is used to schedule the target virtual machine to the target physical machine and execute the virtual machine operation.

[0055] The present invention also provides a storage medium, which includes a stored program, wherein when the program is running, the device where the storage medium is located is controlled to execute the above-mentioned virtual machine scheduling processing method with priority on remaining capacity of cloud storage.

[0056] Compared with the prior art, the embodiments of the present invention have at least the following technical advantages:

[0057] By analyzing the above-mentioned method, system and storage medium for virtual machine scheduling with priority to remaining capacity of cloud storage provided by the present invention, it can be known that in specific applications, by reasonably utilizing the remaining capacity in the physical unit, the utilization rate of system resources can be maximized, resource waste can be avoided, and the utilization efficiency of the entire computing resources can be improved;

[0058] According to the running data volume of the virtual machine and the expected data storage volume, combined with the real-time received setup instructions, the system can first perform a first screening to obtain multiple candidate physical machines with more suitable remaining memory capacity for scheduling and building virtual machines, reducing manual intervention and management costs and improving the automation level of the system; using the remaining capacity priority scheduling strategy, the remaining capacity of the physical machines of the existing virtual machines can be used as much as possible to reduce the deployment cost of new virtual machines, while also reducing the cost of system maintenance and management; the basic data storage capacity requirement is set according to actual conditions to meet the needs of different business scenarios, and in subsequent screening operations, a second screening is performed according to the specific needs of the target virtual machine to obtain the target physical machine, improving the flexibility and scalability of the system; thus, a comprehensive analysis is performed with the remaining memory capacity of the physical machine as a priority, supplemented by the performance and bandwidth of the physical machine, so as to quickly and flexibly select the target physical machine for the target virtual machine. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] 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.

[0060] Figure 1 A schematic diagram of the overall operation steps of a virtual machine scheduling processing method with priority on remaining capacity of cloud storage provided in the first embodiment of the present invention;

[0061] Figure 2 A logical diagram of the steps of a virtual machine scheduling method with priority on remaining capacity of cloud storage provided in the first embodiment of the present invention;

[0062] Figure 3 A schematic diagram of the operation steps of secondary screening to a target physical machine in a virtual machine scheduling processing method with priority on remaining capacity of cloud storage provided in the first embodiment of the present invention;

[0063] Figure 4 A schematic diagram of the operation steps of obtaining a target physical machine by analyzing and screening a virtual machine scheduling method with priority on remaining capacity of cloud storage provided in the first embodiment of the present invention;

[0064] Figure 5 A schematic diagram of the operation steps of obtaining the evaluation weight wz of each evaluation expert in a virtual machine scheduling processing method with priority on remaining capacity of cloud storage provided in the first embodiment of the present invention;

[0065] Figure 6 A schematic diagram of the overall architecture of a virtual machine scheduling processing system with remaining capacity priority provided in the first embodiment of the present invention.

[0066] Reference numerals: receiving module 10 , first screening module 20 , second screening module 30 , calling-in and running module 40 . DETAILED DESCRIPTION

[0067] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0068] The present invention is further described in detail below through specific embodiments in conjunction with the accompanying drawings.

[0069] Embodiment 1

[0070] like Figure 1 or Figure 2 As shown, the present invention proposes a virtual machine scheduling processing method with priority on remaining capacity of cloud storage, comprising the following steps:

[0071] Step S10: receiving a construction instruction of a target virtual machine; the construction instruction includes the operating data volume of the target virtual machine and the estimated data storage volume corresponding to the operating data volume;

[0072] Explanation: During operation, the embodiment of the present application receives the building instructions of the target virtual machine in real time, selects a suitable physical machine from the physical machine group as the target physical machine according to the building instructions of the target virtual machine (the specific operation is shown in the following steps S20-S40), and builds the target virtual machine on the target physical machine, thereby achieving the normal operation of the target virtual machine; the operating data volume of the target virtual machine corresponding to the above-mentioned building instructions and the estimated data storage volume corresponding to the said operating data volume are used to perform a second screening in the subsequent steps S30-S40 to obtain the target physical machine corresponding to the target virtual machine.

[0073] Step S20: Filter and obtain multiple candidate physical machines that meet the basic data storage capacity requirement from the physical machine group according to the building instruction;

[0074] Explanation: The above embodiment of the present application selects multiple physical machines to be selected from the physical machine group of the current server according to the preset basic data storage capacity requirement after receiving the instruction to build the target virtual machine;

[0075] The target virtual machine may be a virtual machine to be scheduled, and the basic data storage capacity requirement is preset based on empirical values, which is the minimum remaining capacity threshold required for basic operation and data cache of the virtual machine when the virtual machine is usually built;

[0076] In this embodiment, after receiving the instruction to build the target virtual machine, firstly, the remaining memory capacity of each physical machine in the physical machine group is obtained; according to the remaining memory capacity of each physical machine, multiple candidate physical machines that meet the basic data storage capacity requirement are screened out from the physical machine group, that is, the candidate physical machines whose remaining memory capacity is greater than or equal to the basic data storage capacity requirement are screened out, and the multiple candidate physical machines can be respectively recorded as {J1, J2, ...Jm}, where m is the total number of the multiple candidate physical machines, and m is greater than or equal to 1;

[0077] It should be noted that step S20 is the first screening operation to screen out multiple physical machines to be selected;

[0078] The above-mentioned embodiment of the present application obtains multiple candidate physical machines through the first screening, and uses any one of the multiple candidate physical machines as the target physical machine to call into the target virtual machine in subsequent operations, thereby avoiding multiple virtual machines from being called into a fully loaded physical machine at the same time, causing the physical machine to be overloaded, and avoiding the waste of resources caused by multiple physical machines running at no load or half load.

[0079] Step S30: Acquire the attribute information of each of the candidate physical machines; perform comprehensive analysis and screening to acquire the target physical machine according to the running data volume of the target virtual machine, the estimated data storage volume corresponding to the running data volume, and the attribute information of the candidate physical machines;

[0080] The attribute information of the physical machine to be selected includes the serial number information of the physical machine to be selected, the data processing rate of the physical machine to be selected, and the network bandwidth value of the physical machine to be selected;

[0081] Step S40: scheduling the target virtual machine to the target physical machine, and executing the virtual machine running operation.

[0082] The above-mentioned target virtual machine is scheduled (or deployed) to the target physical machine, which specifically refers to deploying the target virtual machine instance to the target physical machine; that is, it is determined that the basic requirements of the virtual instance created by the target virtual machine on the target physical machine through virtualization technology are met. The virtual machine uses virtualization software to divide the resources of the physical machine (such as processor, memory, hard disk) and provide them to the virtual machine for use, thereby achieving the effect of running multiple virtual machines on the same physical machine at the same time. Deploying the target virtual machine instance to the target physical machine refers to the process of installing and running the virtual machine on the physical machine through virtualization software after it is created and configured. In this way, the hardware resources of the physical machine can be used to support the operation of the virtual machine, so that multiple independent virtual machine instances can be run on the physical machine at the same time.

[0083] It should be noted that the above-mentioned embodiment of the present application can maximize the utilization rate of system resources and avoid resource waste by reasonably utilizing the remaining capacity in the physical unit, thereby improving the utilization efficiency of the entire computing resources; according to the running data volume and the expected data storage volume of the virtual machine, combined with the real-time received building instructions, the system can first perform a first screening to obtain multiple candidate physical machines with more suitable memory remaining capacity for scheduling and building virtual machines, thereby reducing manual intervention and management costs and improving the automation level of the system;

[0084] By using the scheduling strategy of prioritizing remaining capacity, the remaining capacity of the physical machines of existing virtual machines can be utilized as much as possible, reducing the deployment cost of new virtual machines, while also reducing the cost of system maintenance and management. The basic data storage capacity requirement is set according to actual conditions to meet the needs of different business scenarios. In subsequent screening operations, a second screening is performed according to the specific needs of the target virtual machine to obtain the target physical machine, thereby improving the flexibility and scalability of the system. This results in a comprehensive analysis that prioritizes the remaining memory capacity of the physical machine, supplemented by the performance and bandwidth of the physical machine, so as to quickly and flexibly select the target physical machine for the target virtual machine.

[0085] Specifically, Figure 3 As shown, in step S30, a comprehensive analysis and screening is performed to obtain a target physical machine according to the running data volume of the target virtual machine, the estimated data storage volume corresponding to the running data volume, and the attribute information of the candidate physical machine, including the following operation steps:

[0086] Step S31: normalizing the attribute information of each of the candidate physical machines to obtain the normalized attribute information of the candidate physical machines, and establishing a normalized attribute information set of the candidate physical machines according to the normalized attribute information of all the candidate physical machines;

[0087] Step S32: obtaining attribute information of multiple evaluation experts and initial evaluation scores f of each of the evaluation experts for each of the candidate physical machines; analyzing and screening the target physical machine according to the initial evaluation scores and the normalized attribute information set of the candidate physical machines;

[0088] It should be noted that, in the specific implementation of the above embodiment of the present application, the main purpose of step S30 is to perform a second screening operation by comprehensively analyzing the running data volume of the target virtual machine, the estimated data storage volume, and the attribute information of the candidate physical machine to obtain the target physical machine;

[0089] The above step S31 is to normalize the attribute information of the candidate physical machine, that is, to normalize the data processing rate of each candidate physical machine and the network bandwidth value of the candidate physical machine respectively to obtain the normalized attribute information of the candidate physical machine; then, based on these normalized attribute information, a normalized attribute information set of the candidate physical machine is constructed; this set will help us better understand the advantages and disadvantages of each physical machine;

[0090] The above-mentioned operation of normalizing the attribute information of the physical machine to be selected to obtain the normalized attribute information of the physical machine to be selected is specifically:

[0091] Acquire the data processing rate Y of each of the candidate physical machines; normalize the data processing rate to obtain the normalized data processing rate Y' of each of the candidate physical machines;

[0092] The normalized data processing rate Y' is calculated as follows:

[0093] Where, Y is the data processing rate of the physical machine to be selected; n is the number of physical machines to be selected;

[0094] Acquire a network bandwidth value X of each of the candidate physical machines; normalize the network bandwidth value X to obtain a normalized network bandwidth value X' of each of the candidate physical machines;

[0095] The normalized network bandwidth value X' is calculated as follows:

[0096] In the formula, X is the network bandwidth value of the physical machine to be selected; n is the number of physical machines to be selected;

[0097] According to the normalized data processing rate Y' and the normalized network bandwidth value X', normalized attribute information di (Y', X') of the candidate physical machine is obtained; wherein di is the i-th candidate physical machine;

[0098] The above step S32 is mainly to obtain the opinions of multiple evaluation experts; these experts will evaluate the various characteristics of the physical machine to be selected and give their respective initial evaluation scores; these scores will serve as the basis for the second screening to obtain the target physical machine;

[0099] Through the above steps, we can more scientifically and reasonably select the appropriate physical machine to run the virtual machine, thereby improving the operating efficiency of the entire system; at the same time, this method also fully considers the requirements of virtual machine operation and the characteristics of the selected physical machine, making the virtual machine scheduling process more refined and intelligent;

[0100] In summary, in step S30, we conduct a second screening for each candidate physical machine by comprehensively analyzing various factors to obtain the target physical machine; we can more accurately select the appropriate physical machine to run the virtual machine and improve the performance of the entire system; this process fully reflects the characteristics of intelligence and refinement, and provides strong support for virtual machine scheduling.

[0101] Specifically, Figure 4 As shown, in step S32, the target physical machine is obtained by analyzing and screening according to the initial evaluation score and the normalized attribute information set of the candidate physical machine, including the following operation steps:

[0102] Step S321: preset a performance weight wy for the normalized data processing rate Y'; and preset a bandwidth weight wx for the normalized network bandwidth value X';

[0103] Step S322: obtaining the evaluation weight wz of each evaluation expert; and calculating and obtaining the adoptable evaluation score fz of the initial evaluation score according to the evaluation weight;

[0104] The calculation method of the admissible evaluation score fz is:

[0105] fz=f×wz;

[0106] It should be noted that, in the above-mentioned embodiment of the present application, since the evaluation capabilities of each evaluation expert on the data processing rate and bandwidth value of the physical machine to be selected fluctuate, it is necessary to preset the evaluation weight of each evaluation expert, and then calculate the adoptable evaluation score that can be actually adopted in subsequent calculations from the initial evaluation score of each evaluation expert based on the evaluation weight, so as to calculate a more accurate virtual machine call-in score of the physical machine to be selected based on the adoptable evaluation score; the above-mentioned method of pre-setting the evaluation weight is to score the evaluation capability of each evaluation expert and then normalize them to obtain the evaluation weight of each evaluation expert.

[0107] Step S323: Calculate and obtain a second evaluation score f' according to the adoptable evaluation score fz, the performance weight wy, and the bandwidth weight wx;

[0108] The second evaluation score f' is calculated as follows:

[0109] f'=fz×((Y'×wy)+(X'×wx));

[0110] It should be noted that the above-mentioned embodiment of the present application obtains the basic attribute evaluation weighted value of each candidate physical machine (i.e. (Y'×wy)+(X'×wx)) by multiplying the performance weight wy of each candidate physical machine with the normalized data processing rate Y', and comprehensively multiplying the normalized network bandwidth value of each candidate physical machine with the bandwidth weight wx, and then multiplies the basic attribute evaluation weighted value and the adoptable evaluation score to obtain the second evaluation score f' of the candidate physical machine; the second evaluation score f' can be analyzed in further operations to obtain the call-in score of the candidate physical machine.

[0111] Step S324: setting iteration parameters; the iteration parameters include fitness calculation parameter λ;

[0112] Traverse each of the candidate physical machines, and calculate the fitness F of the candidate physical machine according to the second evaluation score, the weighted value of the basic attribute evaluation of the candidate physical machine and the iteration parameter; determine whether the fitness F is greater than or equal to the preset minimum fitness threshold; if so, record the fitness F to obtain a fitness list; if not, clear the fitness list, and adjust the iteration parameter to obtain a new iteration parameter, and repeat the above operation using the new iteration parameter until the fitness of all the candidate physical machines is greater than or equal to the preset minimum fitness threshold, and a fitness list is obtained;

[0113] The fitness F is calculated as follows:

[0114]

[0115] Where, ((Y'×wy)+(X'×wx)) is the weighted value of the basic attribute evaluation of the candidate physical machine; p is the standard deviation of the weighted values ​​of the basic attribute evaluation of all candidate physical machines; f i ' is the second evaluation score of the i-th candidate physical machine;

[0116] Specifically, the above-mentioned adjustment of the iteration parameter to obtain a new iteration parameter is to subtract a preset iteration parameter adjustment threshold β from the fitness calculation parameter λ to obtain a new fitness calculation parameter;

[0117] The calculation formula of the new fitness calculation parameter is: λ-β;

[0118] Explanation: Under the basic condition of meeting the basic data storage capacity requirements, it is considered that the current candidate physical machines can be used as target physical machines. However, if any of the candidate physical machines has a fitness F less than the preset minimum fitness threshold, it proves that the iteration parameter obtained by the current fitness calculation formula is unreasonable, so it is necessary to readjust the iteration parameter λ; the new fitness calculation parameter, i.e. λ-β; continue to iterate until the fitness of all current candidate physical machines is greater than or equal to the preset minimum fitness threshold, then the current iteration parameter λ is considered stable. The above step S324 provides a technical basis for the subsequent selection of the candidate physical machine with the highest fitness as the target physical machine.

[0119] Step S325: Select the candidate physical machine with the highest fitness in the fitness list as the target physical machine.

[0120] It should be noted that the above-mentioned embodiment of the present application can perform weighted processing on the data processing rate and the network bandwidth according to different requirements and expert evaluation capabilities by presetting weights, so as to more accurately evaluate the performance of the candidate physical machine, which helps to improve the accuracy of the evaluation and make the evaluation score more in line with actual needs; by obtaining the weights of the evaluation experts and calculating the adoptable evaluation scores, the evaluation capabilities and preferences of different evaluation experts can be taken into account, the objectivity of the evaluation results can be improved, and it is helpful to integrate the opinions and weights of various evaluation experts to obtain more comprehensive and accurate evaluation results;

[0121] By comprehensively considering the adoptable evaluation score, performance weight and bandwidth weight, the second evaluation score of the candidate physical machine is obtained, which more accurately reflects the impact of its comprehensive attributes, helps to comprehensively evaluate the candidate physical machine, and provides a basis for subsequent fitness calculation; by setting iteration parameters and calculating fitness, the candidate physical machines can be screened and sorted to improve the efficiency and accuracy of the evaluation, which helps to determine the fitness threshold and screen out the candidate physical machines with higher fitness, providing a basis for selecting the target physical machine; by selecting the candidate physical machine with the highest fitness as the target physical machine, the physical machine with the best performance can be selected according to the evaluation results, improving system performance and efficiency; by comprehensively considering various evaluation indicators and weights, the optimal physical machine is selected as the target, achieving more accurate virtual machine transfer score;

[0122] In summary, the embodiment of the present application comprehensively and accurately evaluates and screens the candidate physical machines by comprehensively considering various evaluation indicators and weights, and using iterative parameters and fitness calculation methods, so as to select the target physical machine with the best performance, which helps to improve the accuracy and efficiency of the evaluation and provide a scientific basis for system optimization and decision-making.

[0123] Specifically, Figure 5 As shown, in step S322, obtaining the evaluation weight wz of each evaluation expert includes the following steps:

[0124] Step S3221: obtaining multiple analysis dimensions K of the evaluation weights of the evaluation experts; the analysis dimensions can be expressed as K={k1, k2, k3, ...km}; wherein km is the mth analysis dimension;

[0125] It should be noted that in order to ensure the accuracy of the expert scores for each candidate physical machine, and thus ensure the accuracy of the final target physical machine, it is necessary to evaluate the evaluation weights of each evaluation expert on the candidate physical machine in many aspects. For example, the number of times the evaluation experts evaluate the evaluation scores of the candidate physical machines, the accuracy, and the professional ability of the evaluation experts to evaluate the candidate physical machines are analyzed and evaluated in multiple dimensions.

[0126] Step S3222: Obtain the evaluation score Lz of each evaluation expert in each analysis dimension; the evaluation score can be expressed as L={l1, l2, l3, ...ln}; wherein ln represents the evaluation score of the nth evaluation expert;

[0127] Step S3223: establishing an expert evaluation matrix R according to the analysis dimension K and the evaluation score L;

[0128] The expert evaluation matrix R can be expressed as:

[0129]

[0130] Among them, kmln represents the evaluation score of the nth evaluation expert in the mth evaluation dimension;

[0131] Step S3224: presetting the evaluation weight wu of each evaluation dimension;

[0132] It should be noted that each evaluation dimension is extremely important for the calculation of the evaluation weight of the evaluation expert. For example, the professional ability of the evaluation expert in evaluating the evaluation score of each candidate physical machine is very important for evaluating the initial evaluation score of the candidate physical machine and even the fitness value obtained by subsequent calculation. Therefore, the professional ability of the evaluation expert has the highest weight. For another example, the number of times the evaluation expert evaluates the candidate physical machine is not particularly important. Therefore, the weight of the number of times the evaluation expert evaluates the candidate physical machine should be relatively low.

[0133] Step S3225: Calculate and obtain the evaluation weight wz of each evaluation expert according to the evaluation weight wu and the expert evaluation matrix R;

[0134] The evaluation weight wz is calculated as follows:

[0135]

[0136] It should be noted that the above-mentioned embodiment of the present application can more comprehensively consider the evaluation ability and preference of the evaluation experts for the evaluation weights of the selected physical machines through multi-dimensional analysis of the evaluation experts, improve the objectivity and accuracy of the evaluation results, and help to establish a complete basis for the evaluation weights of the evaluation experts, and provide detailed information for the subsequent calculation of the evaluation weights; by obtaining the evaluation scores of the evaluation experts in each analysis dimension, the evaluation ability of the evaluation experts in different aspects can be quantified, and data support can be provided for the subsequent calculation of the evaluation weights, which is helpful for a comprehensive evaluation of the evaluation experts and lays the foundation for the establishment of an expert evaluation matrix; by establishing an expert evaluation matrix, the evaluation scores of the evaluation experts can be combined with the multi-dimensional analysis weights to form a comprehensive evaluation data matrix, which provides a basis for the subsequent calculation of the evaluation weights, and helps to combine the evaluation scores of the evaluation experts with the analysis dimension weights, and lays the foundation for calculating the evaluation weights;

[0137] At the same time, when executing the above steps, the evaluation weights of the evaluation dimensions are also preset, which can determine the importance of each dimension in the evaluation in advance, provide guidance for the subsequent evaluation weight calculation, help to clarify the weight relationship of each dimension, and provide guidance and basis for the evaluation weight calculation of the evaluation experts; by calculating the evaluation weight according to the evaluation weight and the expert evaluation matrix, the evaluation factors of multiple dimensions can be comprehensively considered to obtain the final evaluation weight, improve the accuracy of the evaluation results, help to determine the weight of each evaluation expert in the evaluation, and provide an objective basis for the subsequent evaluation of the candidate physical machine;

[0138] In summary, the embodiment of the present application uses multi-dimensional analysis and weight calculation, comprehensively considers the evaluation capabilities and preferences of evaluation experts in different aspects, and derives the final evaluation weights, thereby ensuring the accuracy and precision of the target physical machine finally selected, helping to establish a complete evaluation system and improving the scientificity and objectivity of the evaluation results.

[0139] Embodiment 2

[0140] like Figure 6 As shown, the present invention also proposes a virtual machine scheduling processing system with remaining capacity priority, including a receiving module 10, a first screening module 20, a second screening module 30 and a call-in operation module 40;

[0141] The receiving module 10 is used to receive a construction instruction of a target virtual machine; the construction instruction includes the running data volume of the target virtual machine and the estimated data storage volume corresponding to the running data volume;

[0142] The first screening module 20 is used to screen and obtain multiple candidate physical machines that meet the basic data storage capacity requirement from the physical machine group according to the building instruction;

[0143] The second screening module 30 is used to obtain the attribute information of each of the candidate physical machines; perform comprehensive analysis and screening to obtain the target physical machine according to the running data volume of the target virtual machine, the estimated data storage volume corresponding to the running data volume, and the attribute information of the candidate physical machines;

[0144] The attribute information of the physical machine to be selected includes the serial number information of the physical machine to be selected, the data processing rate of the physical machine to be selected, and the network bandwidth value of the physical machine to be selected;

[0145] The calling-in running module 40 is used to schedule the target virtual machine to the target physical machine and execute the virtual machine running operation.

[0146] Preferably, the second screening module 30 is specifically used to normalize the attribute information of each of the candidate physical machines to obtain the normalized attribute information of the candidate physical machines, and to establish a normalized attribute information set of the candidate physical machines according to the normalized attribute information of all the candidate physical machines;

[0147] Acquire attribute information of multiple evaluation experts and initial evaluation scores f of each of the evaluation experts for each of the candidate physical machines; analyze and screen the target physical machine based on the initial evaluation scores and the normalized attribute information set of the candidate physical machines;

[0148] Preferably, the second screening module 30 is further configured to preset a performance weight wy for the normalized data processing rate Y'; and preset a bandwidth weight wx for the normalized network bandwidth value X';

[0149] Obtaining the evaluation weight wz of each evaluation expert; calculating and obtaining the adoptable evaluation score fz of the initial evaluation score according to the evaluation weight;

[0150] The calculation method of the admissible evaluation score fz is:

[0151] fz=f×wz;

[0152] Calculate and obtain a second evaluation score f' according to the adoptable evaluation score fz, the performance weight wy, and the bandwidth weight wx;

[0153] The second evaluation score f' is calculated as follows:

[0154] f'=fz×((Y'×wy)+(X'×wx));

[0155] Set iteration parameters; the iteration parameters include fitness calculation parameters λ; traverse each of the candidate physical machines, and calculate the fitness F of the candidate physical machine according to the second evaluation score and the weighted value of the basic attribute evaluation of the candidate physical machine and the iteration parameters; determine whether the fitness F is greater than or equal to the preset minimum fitness threshold; if so, record the fitness F to obtain a fitness list; if not, clear the fitness list, and adjust the iteration parameters to obtain new iteration parameters, and repeat the above operation using the new iteration parameters until the fitness of all the candidate physical machines is greater than or equal to the preset minimum fitness threshold, and a fitness list is obtained;

[0156] The fitness F is calculated as follows:

[0157]

[0158] Where, ((Y'×wy)+(X'×wx)) is the weighted value of the basic attribute evaluation of the candidate physical machine; p is the standard deviation of the weighted values ​​of the basic attribute evaluation of all candidate physical machines; f i ' is the second evaluation score of the i-th candidate physical machine;

[0159] Specifically, the above-mentioned adjustment of the iteration parameter to obtain a new iteration parameter is to subtract a preset iteration parameter adjustment threshold β from the fitness calculation parameter λ to obtain a new fitness calculation parameter;

[0160] The calculation formula of the new fitness calculation parameter is: λ-β;

[0161] The candidate physical machine with the highest fitness in the fitness list is selected as the target physical machine.

[0162] Preferably, the second screening module 30, in a specific implementation process, is used to obtain a plurality of analysis dimensions K of the evaluation weights of the evaluation experts; the analysis dimensions can be expressed as K={k1, k2, k3, ...km}; wherein km is the mth analysis dimension;

[0163] Obtain the evaluation score Lz of each evaluation expert in each analysis dimension; the evaluation score can be expressed as L={l1, l2, l3, ...ln}; wherein ln represents the evaluation score of the nth evaluation expert;

[0164] Establish an expert evaluation matrix R according to the analysis dimension K and the evaluation score L;

[0165] The expert evaluation matrix R can be expressed as:

[0166]

[0167] Among them, kmln represents the evaluation score of the nth evaluation expert in the mth evaluation dimension;

[0168] Preset the evaluation weight wu of each evaluation dimension;

[0169] The evaluation weight wz of each evaluation expert is calculated based on the evaluation weight wu and the expert evaluation matrix R;

[0170] The evaluation weight wz is calculated as follows:

[0171]

[0172] In summary, the method, system and storage medium for virtual machine scheduling with cloud storage remaining capacity priority proposed in the examples of the present invention can maximize the utilization of system resources and avoid resource waste by rationally utilizing the remaining capacity in the physical unit, thereby improving the utilization efficiency of the entire computing resources.

[0173] According to the running data volume and expected data storage volume of the virtual machine, combined with the real-time received building instructions, the system can first perform the first screening to obtain multiple candidate physical machines with more suitable remaining memory capacity for scheduling and building virtual machines, reducing manual intervention and management costs and improving the automation level of the system; using the remaining capacity priority scheduling strategy, the remaining capacity of the physical machine of the existing virtual machine can be used as much as possible to reduce the deployment cost of new virtual machines, while also reducing the cost of system maintenance and management; the basic data storage capacity requirement is set according to the actual situation to meet the needs of different business scenarios, and at the same time, in the subsequent screening operation, the target physical machine is obtained by a second screening based on the specific needs of the target virtual machine, thereby improving the flexibility and scalability of the system; thus, a comprehensive analysis is performed with the remaining memory capacity of the physical machine as the priority, supplemented by the performance and bandwidth of the physical machine, so as to quickly and flexibly select the target physical machine for the target virtual machine;

[0174] In the specific secondary screening process, by comprehensively analyzing the attribute information of each candidate physical machine and the evaluation scores of the evaluation experts, and further by comprehensively considering various evaluation indicators and weights, and using iterative parameters and fitness calculation methods, the candidate physical machines are comprehensively and accurately evaluated and screened, so as to select the target physical machine with the best performance, which helps to improve the accuracy and efficiency of the evaluation and provide a scientific basis for system optimization and decision-making; a second screening is performed for each candidate physical machine to obtain the target physical machine; the appropriate physical machine can be selected more accurately to run the virtual machine, thereby improving the performance of the entire system; this process fully reflects the characteristics of intelligence and refinement, and provides strong support for virtual machine scheduling;

[0175] In addition, through multi-dimensional analysis and weight calculation, the evaluation experts' evaluation capabilities and preferences in different aspects are comprehensively considered to obtain the final evaluation weights, which ensures the accuracy and precision of the target physical machine finally selected, helps to establish a complete evaluation system, and improves the scientificity and objectivity of the evaluation results.

[0176] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. A person skilled in the art may modify the technical solutions described in the above embodiments, or replace part or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A virtual machine scheduling method with priority on remaining cloud storage capacity, characterized in that: The steps are as follows: Receive a build instruction of a target virtual machine; the build instruction includes the running data volume of the target virtual machine and the estimated data storage volume corresponding to the running data volume; According to the building instruction, a plurality of candidate physical machines that meet the basic data storage capacity requirement are screened and obtained from the physical machine group; Obtaining attribute information of each of the to-be-selected physical machines; Comprehensively analyze and screen the target physical machine based on the attribute information of the candidate physical machine; Dispatching the target virtual machine to the target physical machine and executing the virtual machine running operation; The attribute information of the physical machine to be selected includes the serial number information of the physical machine to be selected, the data processing rate of the physical machine to be selected, and the network bandwidth value of the physical machine to be selected; The step of selecting and obtaining a plurality of candidate physical machines that meet the basic data storage capacity requirement from the physical machine group according to the building instruction includes the following steps: After receiving the target virtual machine construction instruction, first obtain the remaining memory capacity of each physical machine in the physical machine group; according to the remaining memory capacity of each physical machine, select multiple candidate physical machines that meet the basic data storage capacity requirement from the physical machine group, that is, select the candidate physical machines whose remaining memory capacity is greater than or equal to the basic data storage capacity requirement, and the multiple candidate physical machines can be respectively recorded as {J1, J2, ...Jm}, where m is the total number of the multiple candidate physical machines, and m is greater than or equal to 1; The step of performing comprehensive analysis and screening to obtain the target physical machine according to the attribute information of the candidate physical machine includes the following steps: Normalizing the attribute information of each of the candidate physical machines to obtain normalized attribute information of the candidate physical machines, and establishing a normalized attribute information set of the candidate physical machines according to the normalized attribute information of all the candidate physical machines; Acquire attribute information of multiple evaluation experts and initial evaluation scores f of each of the evaluation experts for each of the candidate physical machines; analyze and screen the target physical machine based on the initial evaluation scores and the normalized attribute information set of the candidate physical machines; The step of normalizing the attribute information of each of the candidate physical machines to obtain the normalized attribute information of the candidate physical machines includes the following steps: Acquire the data processing rate Y of each of the candidate physical machines; normalize the data processing rate to obtain the normalized data processing rate Y' of each of the candidate physical machines; The normalized data processing rate Y' is calculated as follows: Where, Y is the data processing rate of the physical machine to be selected; n is the number of physical machines to be selected; Acquire a network bandwidth value X of each of the candidate physical machines; normalize the network bandwidth value X to obtain a normalized network bandwidth value X' of each of the candidate physical machines; The normalized network bandwidth value X' is calculated as follows: In the formula, X is the network bandwidth value of the physical machine to be selected; n is the number of physical machines to be selected; According to the normalized data processing rate Y' and the normalized network bandwidth value X', normalized attribute information di (Y', X') of the candidate physical machine is obtained; wherein di is the i-th candidate physical machine; The step of analyzing and selecting the target physical machine according to the initial evaluation score and the normalized attribute information set of the candidate physical machine includes the following steps: A performance weight wy is preset for the normalized data processing rate Y'; and a bandwidth weight wx is preset for the normalized network bandwidth value X'; Obtaining the evaluation weight wz of each evaluation expert; calculating and obtaining the adoptable evaluation score fz of the initial evaluation score according to the evaluation weight; Calculate and obtain a second evaluation score f' according to the adoptable evaluation score fz, the performance weight wy, and the bandwidth weight wx; Setting iteration parameters; the iteration parameters include fitness calculation parameter λ; Traverse each of the candidate physical machines, and calculate the fitness F of the candidate physical machine according to the second evaluation score, the weighted value of the basic attribute evaluation of the candidate physical machine and the iteration parameter; determine whether the fitness F is greater than or equal to the preset minimum fitness threshold; if so, record the fitness F to obtain a fitness list; if not, clear the fitness list, and adjust the iteration parameter to obtain a new iteration parameter, and repeat the above operation using the new iteration parameter until the fitness of all the candidate physical machines is greater than or equal to the preset minimum fitness threshold, and a fitness list is obtained; Selecting the candidate physical machine with the highest fitness in the fitness list as the target physical machine; The calculation method of the admissible evaluation score fz is: fz=f×wz; The second evaluation score f' is calculated as follows: f'=fz×((Y'×wy)+(X'×wx)); The fitness F is calculated as follows: Where, ((Y'×wy)+(X'×wx)) is the weighted value of the basic attribute evaluation of the candidate physical machine; p is the standard deviation of the weighted values ​​of the basic attribute evaluation of all candidate physical machines; f i ' is the second evaluation score of the i-th candidate physical machine; The step of obtaining the evaluation weight wz of each evaluation expert comprises the following steps: Acquire multiple analysis dimensions K of the evaluation weights of the evaluation experts; the analysis dimensions can be expressed as K={k1, k2, k3, ...km}; where km is the mth analysis dimension; Obtain the evaluation score Lz of each evaluation expert in each analysis dimension; the evaluation score can be expressed as L={l1, l2, l3, ...ln}; wherein ln represents the evaluation score of the nth evaluation expert; Establish an expert evaluation matrix R according to the analysis dimension K and the evaluation score L; The expert evaluation matrix R can be expressed as: Among them, kmln represents the evaluation score of the nth evaluation expert in the mth evaluation dimension; Preset the evaluation weight wu of each evaluation dimension; The evaluation weight wz of each evaluation expert is calculated based on the evaluation weight wu and the expert evaluation matrix R; The evaluation weight wz is calculated as follows:

2. The method for virtual machine scheduling with priority given to remaining capacity of cloud storage according to claim 1, characterized in that: The step of adjusting the iteration parameter to obtain a new iteration parameter is to subtract a preset iteration parameter adjustment threshold β from the fitness calculation parameter λ to obtain a new fitness calculation parameter; The calculation formula of the new fitness calculation parameter is: λ-β.

3. A storage medium, characterized in that: The storage medium includes a stored program, wherein when the program is running, the device where the storage medium is located is controlled to execute a virtual machine scheduling processing method with priority on remaining capacity of cloud storage as described in any one of claims 1-2 above.

Images