Method and device for intelligent allocation of resources of virtual desktop based on multi-dimensional constraints

Abstract

The application relates to a resource intelligent allocation method and device of a virtual desktop based on multi-dimensional constraints, which comprises the following steps: identifying the CPU performance type of a blade server, dynamically adjusting the super-division coefficient based on the CPU performance type of the blade server; calculating the total core number of a virtual CPU, calculating the theoretical allocation number of a virtual machine based on the total core number of the virtual CPU, combining the virtual desktop specification demand of a user, the physical resource of the blade and the number of the allocated virtual machines of the blade; calculating the actual allocation number of the virtual machine based on the total core number of the virtual CPU and the upper limit of the physical resource specification of the blade; judging whether the allocation result meets the multi-dimensional constraints based on the theoretical allocation number of the virtual machine and the actual allocation number of the virtual machine; when the allocation result meets the multi-dimensional constraints, establishing an allocation priority rule, selecting an optimal target blade based on the allocation priority rule, executing the virtual machine resource allocation by using the optimal target blade, and updating the virtual machine resource state.

Description

Technical Field

[0001] This invention relates to the field of cloud computing and virtualization resource scheduling technology, and in particular to a method and apparatus for intelligent resource allocation of virtual desktops based on multidimensional constraints. Background Technology

[0002] With the popularization of cloud computing technology, enterprises' demand for virtual desktops is growing rapidly. The CPU, memory, and display performance requirements for virtual desktops vary significantly across different business scenarios. As the core hardware for virtual desktops, the rationality of resource allocation for blade servers directly determines the service quality and resource utilization efficiency of virtual desktops. Existing allocation technologies, lacking refined scheduling strategies, struggle to balance the conflict between resource utilization and service quality. Common shortcomings include: fixed over-division coefficients and poor resource adaptability: Most allocation schemes use fixed CPU over-division coefficients without dynamically adjusting based on the CPU performance type of the blade server. This results in underutilization of high-performance CPU resources, while low-power CPUs experience performance bottlenecks due to excessive over-division, failing to match the resource carrying capacity of different hardware configurations. Single allocation dimension and low resource utilization: Existing allocation algorithms often calculate the number of allocable virtual machines solely from the perspective of CPU or memory, without comprehensively considering the synergistic constraints between the two. This easily leads to imbalances such as idle CPUs but insufficient memory or idle memory but saturated CPUs, resulting in overall low resource utilization. The lack of consideration for display requirements and blade specifications leads to unreliable service quality: Existing technologies lack specific allocation mechanisms for different high-fidelity display needs and blade specifications. This can easily result in a single blade being allocated more virtual machines than its display processing capacity can handle, causing issues such as virtual desktop stuttering and latency. The allocation strategy lacks priority guidance, resulting in low scheduling efficiency: Resource allocation does not consider the blade's historical allocation records and randomly selects blades for resource allocation. This can lead to fragmented resource usage, increasing the complexity of subsequent resource maintenance and fault migration, while also reducing the overall efficiency of resource scheduling. Summary of the Invention

[0003] This invention provides a method and apparatus for intelligent resource allocation of virtual desktops based on multidimensional constraints, aiming to solve at least one of the technical problems existing in the prior art.

[0004] The technical solution of this invention is a resource intelligent allocation method for virtual desktops based on multi-dimensional constraints, which includes:

[0005] Identify the CPU performance type of the blade server and dynamically adjust the super-resolution coefficient based on the CPU performance type of the blade server;

[0006] Calculate the total number of virtual CPU cores. Based on the total number of virtual CPU cores, combined with the user's virtual desktop specification requirements, blade physical resources, and the number of virtual machines already allocated to the blade, calculate the theoretical number of virtual machines to be allocated. Calculate the actual number of virtual machines to be allocated based on the total number of virtual CPU cores and the upper limit of blade physical resource specifications.

[0007] Based on the theoretical number of virtual machines allocated and the actual number of virtual machines allocated, determine whether the allocation result meets the multi-dimensional constraints.

[0008] When the allocation result satisfies multi-dimensional constraints, an allocation priority rule is established, and the optimal target blade is selected based on the allocation priority rule.

[0009] The optimal target blade is used to perform virtual machine resource allocation and update the virtual machine resource status.

[0010] According to some embodiments of the present invention, the identification of the CPU performance type of the blade server includes:

[0011] The CPU performance type of the blade server is identified by matching the CPU model with a preset performance library. The CPU performance type of the blade server includes the high-performance library of the Intel Xeon Platinum series and the low-power library of the ARM architecture.

[0012] The corresponding performance correction factor is automatically determined based on the CPU performance type of the blade server.

[0013] According to some embodiments of the present invention, the dynamic adjustment of the super-resolution coefficient based on the CPU performance type of the blade server includes:

[0014] Obtain the base super-resolution coefficient and the performance correction coefficient corresponding to the CPU performance type of the blade server;

[0015] The actual super-resolution coefficient is obtained based on the base super-resolution coefficient corresponding to the CPU performance type of the blade server and the performance correction coefficient.

[0016] The formula for calculating the actual over-resolution coefficient is as follows:

[0017] K = K0 × Kp,

[0018] Where K is the actual super-resolution coefficient, K0 is the basic super-resolution coefficient, and Kp is the performance correction coefficient.

[0019] According to some embodiments of the present invention, establishing allocation priority rules and selecting the optimal target blade based on the allocation priority rules includes:

[0020] Establish blade priority sorting rules: blades with existing allocation records are set as first priority, blades with the number of allocated virtual machines greater than or equal to the preset upper limit are set as second priority, and blades with at least one allocated virtual machine are set as third priority; blades are filtered from high to low priority, and the first blade that meets the conditions is selected as the optimal allocation target blade; if there are multiple blades with the same priority, the blade with the most allocated virtual machines is selected as the optimal allocation target blade.

[0021] According to some embodiments of the present invention, calculating the total number of virtual CPU cores includes:

[0022] Obtain the physical CPU core count of the blade server and the actual super-resolution coefficient;

[0023] The total number of virtual CPU cores is calculated based on the physical CPU core count and the actual super-resolution coefficient. The formula for calculating the total number of virtual CPU cores is as follows:

[0024] ,

[0025] in, This represents the total number of virtual CPU cores. K represents the physical number of CPU cores in the blade server, and K represents the actual super-resolution coefficient.

[0026] According to some embodiments of the present invention, the calculation of the theoretical number of virtual machines to be allocated based on the total number of virtual CPU cores, combined with the user's virtual desktop specification requirements, blade physical resources, and the number of virtual machines already allocated to the blade, includes:

[0027] Analyze user virtual desktop specification requirements to determine whether the display requirement is single-screen or dual-screen display, and then determine the single / dual-screen coefficient;

[0028] Calculate the difference between the total number of virtual CPU cores and the number of allocated virtual CPUs;

[0029] Obtain the virtual CPU quantity requirement requested by the user, and calculate the product of the virtual CPU quantity requirement requested by the user and the single / dual screen coefficient;

[0030] The theoretical number of virtual machines allocated in the CPU dimension is calculated by multiplying the difference between the total number of virtual CPU cores and the number of allocated virtual CPUs, the user's requested virtual CPU quantity requirement, and the single / dual-screen coefficient. The formula for calculating the theoretical number of virtual machines allocated in the CPU dimension is as follows:

[0031] ,

[0032] in, Theoretically, how many virtual machines should be allocated based on CPU capacity? This represents the total number of virtual CPU cores. The number of virtual CPUs already allocated. For the number of virtual CPUs requested by the user, For single / dual screen coefficients.

[0033] According to some embodiments of the present invention, it further includes:

[0034] Get the total physical memory and allocated physical memory of the blade server;

[0035] Analyze user memory requirements for virtual desktop specifications;

[0036] Calculate the difference between the total physical memory of the blade server and the allocated physical memory and reserved memory;

[0037] The theoretical number of virtual machines to be allocated in terms of memory is calculated based on the difference between the total physical memory of the blade server and the allocated physical memory and reserved memory, and the user's memory requirement. The formula for calculating the theoretical number of virtual machines to be allocated in terms of memory is as follows:

[0038] ,

[0039] in, The theoretical number of virtual machines to allocate in terms of memory dimension. This refers to the total physical memory of the blade server. This is the allocated physical memory; 8GB is reserved memory. User memory requirements;

[0040] The theoretical number of virtual machines is determined based on the theoretical number of virtual machines allocated in terms of CPU and memory. The formula for the theoretical number of virtual machines allocated is as follows:

[0041] ,

[0042] in, Theoretically allocate a number of virtual machines. Theoretically, how many virtual machines should be allocated based on CPU capacity? The theoretical number of virtual machines to be allocated in terms of memory.

[0043] According to some embodiments of the present invention, the calculation of the actual number of virtual machines allocated based on the total number of virtual CPU cores and the upper limit of blade physical resource specifications includes:

[0044] Obtain the upper limit of blade physical resource specifications based on the blade model of the blade server;

[0045] The remaining number of allocable virtual machines is determined based on the upper limit of the blade physical resource specifications and the number of virtual machines already allocated. The formula for calculating the remaining number of allocable virtual machines is as follows:

[0046] ,

[0047] in, The remaining number of allocable virtual machines. This represents the upper limit of the physical resource specifications for the blade. The number of virtual machines allocated to the blade;

[0048] The actual number of virtual machines allocated is determined based on the remaining allocable virtual machine count and the theoretical allocation count. The formula for calculating the actual allocation count is as follows:

[0049] ,

[0050] in, The actual number of virtual machines allocated. Theoretically allocate a number of virtual machines. The remaining number of allocable virtual machines. This indicates rounding to the nearest integer.

[0051] According to some embodiments of the present invention, determining whether the allocation result satisfies multi-dimensional constraints based on the theoretical number of virtual machines allocated and the actual number of virtual machines allocated includes:

[0052] Determine whether the theoretical number of virtual machines to be allocated is less than or equal to the number of remaining allocable virtual machines;

[0053] If the theoretical number of virtual machines to be allocated is less than or equal to the number of remaining allocable virtual machines, and the allocation result satisfies the multi-dimensional constraints, then the actual number of virtual machines allocated is determined to be the theoretical number of virtual machines to be allocated.

[0054] If the theoretical number of virtual machines to be allocated is not less than or equal to the number of remaining allocable virtual machines, and the allocation result does not meet the multi-dimensional constraints, then the actual number of virtual machines to be allocated is determined to be the number of remaining allocable virtual machines.

[0055] The present invention also relates to a computer device, including a memory and a processor, wherein the processor performs the above-described method when executing a computer program stored in the memory.

[0056] The present invention also relates to a computer-readable storage medium storing computer program instructions thereon, which, when executed by a processor, implement the above-described method.

[0057] The intelligent resource allocation method and apparatus for virtual desktops based on multi-dimensional constraints provided in this invention embodiment has at least one of the following advantages or beneficial effects: It automatically identifies the CPU performance type of the blade server, understands the CPU performance type of the blade server, and dynamically adjusts the super-division coefficient in the future. By dynamically adjusting the super-division coefficient, the super-division strategy is matched with hardware performance, avoiding performance waste or overload and optimizing resource utilization. The total number of virtual CPU cores is calculated to determine the total number of available virtual CPU cores on the blade server. Based on the total number of virtual CPU cores, combined with the user's virtual desktop specification requirements, blade physical resources, and the number of virtual machines already allocated to the blade, the theoretical number of virtual machines to be allocated is calculated from the two core dimensions of CPU and memory. By introducing a load balancing algorithm, the fairness and efficiency of resource allocation are ensured. Based on the total number of virtual CPU cores, and combined with the upper limit of the blade physical resource specifications, the actual number of virtual machines allocated is calculated, accurately calculating the actual number of virtual machines allocated, avoiding resource allocation imbalance. By considering the limitations of other resources such as memory and storage of the blade server, the utilization rate of physical resources is regularly evaluated, and the resource allocation strategy is dynamically adjusted to ensure that the allocation result meets multi-dimensional constraints.

[0058] Subsequently, based on the theoretical and actual number of virtual machines allocated, it is determined whether the allocation result meets multi-dimensional constraints. This ensures precise matching between blades of different specifications and corresponding high-fidelity display requirements, guaranteeing that the number of allocated virtual machines does not exceed the blade's capacity and improving the virtual desktop user experience. When the allocation result meets the multi-dimensional constraints, an allocation priority rule is established. Based on this rule, blades with existing allocation records are selected as the optimal target blades for new resource allocation. This achieves centralized resource management, improves resource scheduling efficiency and maintenance convenience, reduces resource maintenance costs, and ensures the quality of virtual desktop services. Virtual machine resource allocation is performed using the optimal target blade, and the virtual machine resource status is updated, achieving efficient resource scheduling.

[0059] Furthermore, additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0060] Figure 1 This is a flowchart illustrating the overall process of the intelligent resource allocation method for virtual desktops based on multidimensional constraints provided in this embodiment of the invention.

[0061] Figure 2 This is a detailed flowchart of step S100 in the intelligent resource allocation method for virtual desktops based on multidimensional constraints provided in this embodiment of the invention.

[0062] Figure 3 This is a detailed flowchart of step S400 in the intelligent resource allocation method for virtual desktops based on multidimensional constraints provided in this embodiment of the invention.

[0063] Figure 4 This is a detailed flowchart of step S200 in the intelligent resource allocation method for virtual desktops based on multidimensional constraints provided in this embodiment of the invention.

[0064] Figure 5 This is a detailed flowchart of step S300 in the intelligent resource allocation method for virtual desktops based on multidimensional constraints provided in this embodiment of the invention. Detailed Implementation

[0065] The following will provide a clear and complete description of the concept, specific structure, and technical effects of the present invention in conjunction with the embodiments and accompanying drawings, so as to fully understand the purpose, solution, and effects of the present invention.

[0066] It should be noted that, unless otherwise specified, when a feature is referred to as "fixed" or "connected" to another feature, it can be directly fixed or connected to the other feature, or indirectly fixed or connected to the other feature. The singular forms "a," "described," and "the" used herein are also intended to include the plural forms, unless the context clearly indicates otherwise. Furthermore, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this specification is for the purpose of describing particular embodiments only and not for limiting the invention. The term "and / or" as used herein includes any combination of one or more of the associated listed items.

[0067] It should be understood that although the terms first, second, third, etc., may be used to describe various elements in this disclosure, these elements should not be limited to these terms. These terms are used only to distinguish elements of the same type from one another. For example, a first element may also be referred to as a second element without departing from the scope of this disclosure, and similarly, a second element may also be referred to as a first element. Any and all instances or exemplary language (“e.g.,” “such as,” etc.) provided herein are intended only to better illustrate embodiments of the invention and, unless otherwise required, do not impose a limitation on the scope of the invention.

[0068] This invention provides a method and apparatus for intelligent resource allocation of virtual desktops based on multidimensional constraints. Through dynamic super-division, multidimensional calculation and priority scheduling, it achieves accurate resource allocation, providing efficient and stable technical support for enterprise cloud desktop deployment, and has important practical application value.

[0069] Please refer to the following. Figures 1 to 5 The resource intelligent allocation method and apparatus for virtual desktops based on multidimensional constraints provided in the embodiments of the present invention will be further described.

[0070] Reference Figure 1 As shown, Figure 1 This is a flowchart illustrating the overall process of a resource intelligent allocation method for virtual desktops based on multidimensional constraints, as provided in this embodiment of the invention. The resource intelligent allocation method for virtual desktops based on multidimensional constraints includes, but is not limited to, steps S100 to S500. Specifically,

[0071] S100: Identifies the CPU performance type of the blade server and dynamically adjusts the super-resolution coefficient based on the CPU performance type of the blade server;

[0072] S200: Calculates the total number of virtual CPU cores. Based on the total number of virtual CPU cores, combined with the user's virtual desktop specification requirements, blade physical resources, and the number of virtual machines already allocated to the blade, calculates the theoretical number of virtual machines to be allocated. Based on the total number of virtual CPU cores and the upper limit of blade physical resource specifications, calculates the actual number of virtual machines to be allocated.

[0073] S300: Based on the theoretical number of virtual machines allocated and the actual number of virtual machines allocated, determine whether the allocation result meets the multi-dimensional constraints;

[0074] S400: When the allocation result satisfies multi-dimensional constraints, establish an allocation priority rule and select the optimal target blade based on the allocation priority rule;

[0075] S500: Perform virtual machine resource allocation using the optimal target blade and update the virtual machine resource status.

[0076] In some embodiments of the present invention, the intelligent resource allocation method for virtual desktops based on multi-dimensional constraints includes: automatically identifying the CPU performance type of the blade server to understand its performance type and dynamically adjust the super-resolution coefficient. By dynamically adjusting the super-resolution coefficient, the super-resolution strategy is matched with hardware performance, avoiding performance waste or overload and optimizing resource utilization. The total number of virtual CPU cores is calculated to determine the total number of available virtual CPU cores on the blade server. Based on the total number of virtual CPU cores, combined with the user's virtual desktop specification requirements (vCPU core count, memory capacity, display resolution, single / dual-screen display requirements) as input, and combined with the blade's physical resources (CPU core count, memory capacity) and the number of virtual machines already allocated to the blade, the theoretical number of virtual machines to be allocated is calculated from the two core dimensions of CPU and memory. By introducing a load balancing algorithm, the fairness and efficiency of resource allocation are ensured. Based on the total number of virtual CPU cores, and combined with the upper limit of the blade's physical resource specifications, the actual number of virtual machines allocated is calculated. This accurate calculation of the actual number of virtual machines avoids resource allocation imbalance. By considering the limitations of other resources such as memory and storage on the blade server, the utilization rate of physical resources is periodically evaluated, and the resource allocation strategy is dynamically adjusted to ensure that the allocation result meets multi-dimensional constraints.

[0077] Subsequently, based on the theoretical and actual number of virtual machines allocated, it is determined whether the allocation result meets multi-dimensional constraints. This ensures precise matching between blades of different specifications and corresponding high-fidelity display requirements, guaranteeing that the number of allocated virtual machines does not exceed the blade's capacity and improving the virtual desktop user experience. When the allocation result meets the multi-dimensional constraints, an allocation priority rule is established. Based on this rule, blades with existing allocation records are selected as the optimal target blades for new resource allocation. This achieves centralized resource management, improves resource scheduling efficiency and maintenance convenience, reduces resource maintenance costs, and ensures the quality of virtual desktop services. Virtual machine resource allocation is performed using the optimal target blade, and the virtual machine resource status is updated.

[0078] It is understood that in this embodiment of the invention, by dynamically adjusting the super-resolution coefficient, the super-resolution coefficient of the high-performance CPU blade can be increased to 3.0, while the super-resolution coefficient of the low-power CPU blade is maintained at 2.0, maximizing resource utilization while ensuring performance. Simultaneously, the dual-dimensional computing model of CPU and memory avoids resource idleness caused by single-dimensional limitations. Establishing allocation priority rules, by prioritizing blades with existing allocation records as the optimal target blades for new resource allocation, helps to consolidate server load, reduce the number of active physical nodes, thereby reducing overall system energy consumption and simplifying operation and maintenance management.

[0079] Therefore, this invention constructs a three-level resource allocation system of "hardware performance awareness - multi-dimensional constraint calculation - priority scheduling", dynamically adjusts the super-resolution coefficient to match hardware performance, and constructs constraints based on CPU, memory, display requirements and blade specifications, and finally achieves efficient virtual machine resource scheduling based on historical allocation records.

[0080] Reference Figure 2 As shown, Figure 2 This is a detailed flowchart of step S100 in the intelligent resource allocation method for virtual desktops based on multi-dimensional constraints provided in this embodiment of the invention. Step S100 includes, but is not limited to, steps S110 to S120. Specifically,

[0081] S110: Identifies the CPU performance type of the blade server by matching the CPU model with the preset performance library. The CPU performance type of the blade server includes the high-performance library of the Intel Xeon Platinum series and the low-power library of the ARM architecture.

[0082] S120: Automatically determines the corresponding performance correction factor based on the CPU performance type of the blade server.

[0083] In some embodiments of the present invention, identifying the CPU performance type of a blade server includes: using a hardware detection tool (such as the `lscpu` command) to obtain the CPU model of the blade server, and matching the obtained CPU model with a preset performance library. If a model in the high-performance library (such as the Intel Xeon Platinum series) is matched, it is identified as a high-performance type; if a model in the low-power library (such as the ARM architecture) is matched, it is identified as a low-power type. Based on the identified CPU performance type, a corresponding performance correction factor is automatically determined. For example, if the CPU model belongs to the high-performance library, the correction factor in the high-performance library is used. The correction factor in the high-performance library is usually high (e.g., a correction factor of 3.0), indicating that more aggressive super-scaling can be performed. If the CPU model belongs to the low-power library, the correction factor in the low-power library is used (e.g., a correction factor of 2.0). The correction factor in the low-power library is usually low, indicating that more conservative super-scaling is needed to avoid performance bottlenecks.

[0084] By establishing a preset performance library and matching it with the CPU model, the CPU performance type of the blade server can be accurately identified, and the corresponding performance correction factor can be automatically determined. This method can provide a more accurate basis for subsequent resource allocation, thereby optimizing the performance and resource utilization of virtual machines.

[0085] In some embodiments of the present invention, dynamically adjusting the super-resolution coefficient based on the CPU performance type of the blade server includes:

[0086] Obtain the base super-resolution coefficient and performance correction coefficient corresponding to the CPU performance type of the blade server;

[0087] The actual super-resolution coefficient is obtained by using the base super-resolution coefficient and performance correction coefficient corresponding to the CPU performance type of the blade server.

[0088] The formula for calculating the actual over-scoring coefficient is:

[0089] K = K0 × Kp,

[0090] Where K is the actual super-resolution coefficient, K0 is the basic super-resolution coefficient, and Kp is the performance correction coefficient.

[0091] In some embodiments of the present invention, the basic super-resolution coefficient is set to K0 (default value 2.0, configurable), and the performance correction coefficient is set to Kp (1.5 for high-performance CPUs and 1.0 for low-power CPUs). Then, the actual super-resolution coefficient K = K0 × Kp. After that, the Kp value is automatically determined by matching the CPU model with the preset performance library (such as the high-performance library containing the Intel Xeon Platinum series and the low-power library containing the ARM architecture), and finally achieving: high-performance CPU (K=3.0) and low-power CPU (K=2.0).

[0092] The method obtains a base super-resolution coefficient and a performance correction coefficient based on the CPU performance type of the blade server, and then calculates the actual super-resolution coefficient. This method can more accurately reflect the actual performance and resource allocation capabilities of the CPU, thereby optimizing the resource allocation and performance of virtual machines. Through this adjustment mechanism, the resource potential of high-performance CPUs is fully released, while low-power CPUs avoid performance inadequacy due to excessive super-resolution, providing an accurate resource base for subsequent allocation calculations.

[0093] By automatically matching and correcting the coefficients using a preset CPU performance library, the super-scoring coefficient is dynamically adjusted according to the CPU type (3.0 / 2.0), solving the problem of resource waste or insufficient performance caused by the existing fixed super-scoring. The core of this method lies in the quantitative definition and matching logic of the performance correction coefficient.

[0094] Reference Figure 3 As shown, Figure 3 This is a detailed flowchart of step S400 in the intelligent resource allocation method for virtual desktops based on multidimensional constraints provided in this embodiment of the invention. Step S400 includes, but is not limited to, steps S410 to S430. Specifically,

[0095] S410: Establish blade priority sorting rules, pre-set blades with existing allocation records as first priority, blades with the number of allocated virtual machines greater than or equal to the preset upper limit as second priority, and blades allocated at least one virtual machine as third priority; S420: Filter blades from high to low priority, and select the first blade that meets the conditions as the optimal allocation target blade; S430: If there are multiple blades with the same priority, select the blade with the most allocated virtual machines as the optimal allocation target blade.

[0096] In this embodiment of the invention, to improve resource scheduling efficiency and maintenance convenience, a "nearest allocation" priority rule is adopted, and the specific decision-making logic is as follows:

[0097] Establish blade priority sorting rules: First priority (blades with existing user allocation records) > Second priority (blades with allocated virtual machines ≥ 50% of the specification limit) > Third priority (other blades that meet the requirement of allocated virtual machine count > 0);

[0098] Select blades from high to low priority and choose the first blade that meets the criteria as the optimal target blade for allocation;

[0099] If blade servers exist with the same priority, select the blade server that allocates the most virtual machines to ensure optimal resource utilization.

[0100] The optimal target blade is selected efficiently based on preset priority rules, thereby optimizing the resource allocation and management of virtual machines. This strategy enables centralized resource management, reduces the complexity of cross-blade resource scheduling, and lowers the resource reorganization cost during fault migration.

[0101] Establish a two-dimensional priority rule based on "user relevance + resource occupancy" (P1>P2>P3) to prioritize resource allocation from blades with existing allocation records, thereby improving scheduling efficiency and maintenance convenience, reducing fault migration costs, and solving the resource dispersion problem caused by random allocation in existing technologies.

[0102] In some embodiments of the present invention, calculating the total number of virtual CPU cores includes:

[0103] Obtain the physical CPU core count and actual super-resolution coefficient of the blade server;

[0104] The total number of virtual CPU cores is calculated based on the physical CPU core count and the actual super-resolution coefficient. The formula for calculating the total number of virtual CPU cores is as follows:

[0105] ,

[0106] in, This represents the total number of virtual CPU cores. K represents the physical number of CPU cores in the blade server, and K represents the actual super-resolution coefficient.

[0107] It is understandable that the super-resolution coefficient is used to convert the number of physical CPU cores of the blade into the total number of allocable vCPU cores, and its value is dynamically adjusted according to CPU performance. In this embodiment of the invention, the total number of virtual CPU cores is calculated. This process involves obtaining the physical CPU core count Cphy and the actual super-resolution coefficient K, then calculating the product of Cphy and K. The resulting product is the total number of virtual CPU cores. The total number of virtual CPU cores is calculated, thus providing a basis for resource allocation in virtual machines.

[0108] Reference Figure 4 As shown, Figure 4 This is a detailed flowchart of step S200 in the intelligent resource allocation method for virtual desktops based on multi-dimensional constraints provided in this embodiment of the invention. Step S200 includes, but is not limited to, steps S210 to S240. Specifically,

[0109] S210: Analyze the user's virtual desktop specification requirements to determine whether the display requirement is single-screen display or dual-screen display, and then determine the single / dual-screen coefficient;

[0110] S220: Calculate the difference between the total number of virtual CPU cores and the number of allocated virtual CPUs;

[0111] S230: Obtain the virtual CPU quantity requirement requested by the user, and calculate the product of the virtual CPU quantity requirement requested by the user and the single / dual screen coefficient;

[0112] S240: The theoretical number of virtual machines to be allocated in the CPU dimension is calculated by multiplying the difference between the total number of virtual CPU cores and the number of allocated virtual CPUs, the user's requested number of virtual CPUs, and the single / dual screen coefficient.

[0113] In some embodiments of the present invention, the calculation of the theoretical number of virtual machines to be allocated, in combination with the user's virtual desktop specification requirements, blade physical resources, and the number of virtual machines already allocated to the blade, includes: First, determining whether the user's display requirement is single-screen display or dual-screen display, so as to determine the setting of the single / dual-screen coefficient. For example, when the user's display requirement is single-screen display, the single / dual-screen coefficient is set to 1, and when the user's display requirement is dual-screen display, the single / dual-screen coefficient is set to 2.

[0114] The process begins by calculating the difference between the total number of virtual CPU cores and the number of allocated virtual CPUs, providing the data basis for calculating the remaining available CPU cores. Next, the user's requested virtual CPU core requirement is obtained, representing the number of CPU cores the user wishes to allocate to their virtual machines. The product of the user's requested virtual CPU core requirement and the single / dual-screen coefficient is then calculated; this step adjusts the requested CPU core count to accommodate single-screen or dual-screen display needs. Finally, the theoretical number of virtual machines that can be allocated in terms of CPU is calculated, based on the remaining CPU cores and the adjusted user-requested CPU core count.

[0115] Calculating the theoretical number of virtual machines allocated based on CPU usage using the methods described above can help system administrators or automation tools allocate resources rationally, ensuring that user needs are met while avoiding resource waste.

[0116] In some embodiments of the present invention, the resource intelligent allocation method for virtual desktops based on multidimensional constraints further includes:

[0117] Get the total physical memory and allocated physical memory of the blade server;

[0118] Analyze user memory requirements for virtual desktop specifications;

[0119] Calculate the difference between the total physical memory of the blade server and the allocated physical memory and reserved memory;

[0120] The theoretical number of virtual machines to be allocated based on memory is calculated by taking the difference between the total physical memory of the blade server, the allocated physical memory, the reserved physical memory, and the user's memory requirements. The formula for calculating the theoretical number of virtual machines to be allocated based on memory is as follows:

[0121]

[0122] in, The theoretical number of virtual machines to allocate in terms of memory dimension. This refers to the total physical memory of the blade server. This is the allocated physical memory; 8GB is reserved memory. User memory requirements;

[0123] The theoretical number of virtual machines is determined based on the theoretical number of virtual machines allocated at the CPU level and the theoretical number of virtual machines allocated at the memory level. The formula for the theoretical number of virtual machines allocated is as follows:

[0124] ,

[0125] in, Theoretically allocate a number of virtual machines. Theoretically, how many virtual machines should be allocated based on CPU capacity? The theoretical number of virtual machines to be allocated in terms of memory.

[0126] In some embodiments of the present invention, the total physical memory of the blade server is collected in real time. With allocated amount ; parse the user's virtual desktop specifications to obtain the memory required for a single virtual machine. ; Calculate the theoretical number of virtual machines that can be redeployed in terms of memory using the above formula; simultaneously obtain the theoretical number of virtual machines allocated in terms of CPU using the formula for CPU; take the min( , As the final The result is rounded down and then output.

[0127] Exposing both memory and CPU as independent performance limits simultaneously avoids over-focusing on CPU usage leading to memory exhaustion or vice versa. 8 GB of memory is reserved as a "safety cushion" for the host kernel, KVM, OVS, monitoring agents, etc., significantly reducing the probability of OutOfMemoryError (OOM) and improving host stability. By using single / dual-screen linkage, if the user switches to dual screens... Synchronous amplification with vCPU Automatic shrinking ensures no degradation in experience.

[0128] when When the value is 0, the scheduler can immediately mark the blade as no longer to be allocated, triggering a hibernation or power-down strategy, reducing the overall power consumption of the battery by 8%–15%.

[0129] Long-term records The distribution, analysis, and determination of whether "memory runs out first" or "CPU runs out first" guide the next purchase of memory modules or CPUs, avoiding blind expansion.

[0130] By querying server hardware information or using system commands, the total physical memory and allocated physical memory of the blade server can be obtained. Analyzing the user's virtual desktop specification requirements determines the amount of memory needed for each virtual machine requested by the user. Calculating the difference between the blade server's total physical memory and the allocated physical memory and reserved memory helps to understand the remaining amount of memory available for allocation. Based on the remaining memory and the user's memory requirements, the theoretical number of virtual machines to be allocated in terms of memory is calculated. Finally, based on the theoretical allocation numbers in terms of CPU and memory, the final theoretical number of virtual machines to be allocated is determined.

[0131] This calculation method ensures that the limitations of both CPU and memory are taken into account when allocating resources, thereby avoiding resource waste and performance bottlenecks.

[0132] In some embodiments of the present invention, calculating the actual number of virtual machines allocated based on the total number of virtual CPU cores and the upper limit of blade physical resource specifications includes:

[0133] Obtain the upper limit of blade physical resource specifications based on the blade model of the blade server;

[0134] The remaining number of allocable virtual machines is determined based on the upper limit of blade physical resource specifications and the number of virtual machines already allocated. The formula for calculating the remaining number of allocable virtual machines is as follows:

[0135] ,

[0136] in, The remaining number of allocable virtual machines. This represents the upper limit of the physical resource specifications for the blade. The number of virtual machines allocated to the blade;

[0137] The actual number of virtual machines allocated is determined based on the remaining allocable virtual machines and the theoretical number of virtual machines to be allocated. The formula for calculating the actual number of virtual machines allocated is as follows:

[0138] ,

[0139] in, The actual number of virtual machines allocated. Theoretically allocate a number of virtual machines. The remaining number of allocable virtual machines. This indicates rounding to the nearest integer.

[0140] In some embodiments of the present invention, the preset blade model - physical resource specification upper limit is used. The mapping table retrieves the upper limit of the blade's physical resource specifications. This upper limit, as we understand it, represents the maximum number of virtual machines that can be allocated to the blade. Then, the number of virtual machines already allocated to the blade is read in real-time. The remaining number of allocable virtual machines is calculated using the formula above. The theoretical number of virtual machines to be allocated is then calculated. With the remaining number of allocable virtual machines To make a comparison, the actual number of virtual machines allocated... The output is rounded to the nearest integer, avoiding partial allocation or fragmented virtual machines.

[0141] The initial capacity is determined by parallel evaluation of virtual computing resources and physical memory resources. This initial capacity is then forcibly verified and integrated with the remaining quota limited by the inherent capacity limit of the physical server, ensuring that the allocation result sequentially satisfies constraints across all independent dimensions. The core of this method lies in integrating multi-dimensional constraints into a complete and precise mathematical decision-making process through logical concatenation and layer-by-layer filtering.

[0142] In some embodiments of the present invention

[0143] Virtual machine allocation strategy,

[0144] Based on blade specifications and virtual desktop display requirements, a differentiated allocation strategy was developed, clarifying standardized rules and capacity limits for resource allocation, as follows:

[0145] Blade specifications and display load correspondence rules:

[0146] 1) Blade specification BC-B: Maximum of 8 virtual machines can be allocated, supporting 8-channel 1080P high-fidelity display output;

[0147] 2) Blade specification BC-S: Maximum of 4 virtual machines can be allocated, supporting 4-channel 2K high-fidelity display output;

[0148] 3) Blade specification BC-P: can allocate up to 2 virtual machines and supports 2-way 4K high-fidelity display output (including main and secondary screen scenarios).

[0149] vCPU and memory allocation standardization rules: 1) vCPU allocation follows fixed specifications: 4 cores, 8 cores, 16 cores, 24 cores, to avoid resource fragmentation caused by non-standard allocation;

[0150] 2) Memory allocation follows fixed specifications: 8GB, 16GB, 32GB, and the total allocation amount must not exceed the blade physical memory. At the same time, 8GB of memory must be reserved for the virtual machine base to ensure the basic operation of the virtual machine.

[0151] Multidimensional assessment and intelligent decision-making

[0152] Based on dynamic hardware performance adaptation, centered on multi-dimensional constraint calculation, and optimized by priority scheduling, a closed-loop intelligent allocation logic is formed, supporting real-time resource scheduling for large-scale blade clusters. The core algorithm consists of dynamic super-resolution coefficient adjustment, multi-dimensional allocatable quantity calculation, and priority allocation collaboratively to achieve intelligent resource allocation.

[0153] Reference Figure 5 As shown, Figure 5 This is a detailed flowchart of step S300 in the intelligent resource allocation method for virtual desktops based on multidimensional constraints provided in this embodiment of the invention. Step S300 includes, but is not limited to, steps S310 to S330. Specifically,

[0154] S310: Determine whether the theoretical number of virtual machines to be allocated is less than or equal to the number of remaining allocable virtual machines;

[0155] S320: If the theoretical number of virtual machines to be allocated is less than or equal to the number of remaining allocable virtual machines, and the allocation result satisfies the multi-dimensional constraints, then the actual number of virtual machines allocated is determined to be the theoretical number of virtual machines to be allocated.

[0156] S330: If the theoretical number of virtual machines to be allocated is not less than or equal to the number of remaining allocable virtual machines, the allocation result is determined not to meet the multi-dimensional constraints, and the actual number of virtual machines allocated is determined to be the number of remaining allocable virtual machines.

[0157] In some embodiments of the present invention, determining whether the allocation result satisfies multi-dimensional constraints based on the theoretical number of virtual machines allocated and the actual number of virtual machines allocated includes: determining whether the theoretical number of virtual machines allocated is less than or equal to the remaining allocable number of virtual machines, and setting the theoretical number of virtual machines allocated... Less than or equal to the number of remaining allocatable virtual machines As the sole gatekeeper, it ensures that no scheduling will exceed the verified physical resource limit of the blade, completely eliminating blue screens, OOM, or scheduling rollbacks caused by over-provisioning.

[0158] When the theoretical number of virtual machines to be allocated is less than or equal to the number of remaining allocable virtual machines, and the allocation result satisfies the multi-dimensional constraints, the actual number of virtual machines allocated is directly taken as the theoretical number of virtual machines to be allocated. The small size means that the scheduling plan does not require secondary modification and can be successfully issued immediately, significantly reducing the number of round-trip communications between the orchestration system and the underlying virtualization, and shortening the virtual machine creation latency by more than 20% on average.

[0159] If the theoretical number of virtual machines to be allocated is not less than or equal to the remaining number of allocable virtual machines, and if the theoretical number of virtual machines to be allocated exceeds the remaining number of allocable virtual machines, the allocation result is determined to not meet the multi-dimensional constraints, and the allocation is automatically downgraded to determining the actual number of virtual machines to be allocated as the remaining number of allocable virtual machines. This ensures at least partial delivery and prevents user requests from completely failing; it also triggers capacity alarms to provide quantitative basis for capacity expansion for operations and maintenance.

[0160] In some embodiments of the present invention, the resource intelligent allocation method for virtual desktops based on multidimensional constraints further includes:

[0161] If the user's virtual desktop specification requirement is less than the actual number of virtual machines allocated, determine whether to continue trying to allocate;

[0162] If a continue allocation instruction is received, the difference between the user's virtual desktop specification requirements and the actual number of virtual machines allocated is calculated to obtain the difference. The difference is used as the difference in the number of virtual desktops allocated. According to the allocation priority rules, the optimal number of target blades is selected based on the difference in the number of virtual desktops allocated.

[0163] If a termination allocation instruction is received, the user-requested virtual desktop resource is created, and the resource status is updated.

[0164] Understandably, when the user's virtual desktop requirement is less than the actual number of virtual machines available for allocation, the system does not blindly allocate all virtual desktops at once. Instead, it enters a secondary decision-making process to avoid resource idleness and subsequent fragmentation caused by over-allocation. By subtracting the user's virtual desktop requirement from the actual number of virtual machines allocated, the system obtains the difference in the number of virtual desktops to allocate. The remaining quota is quantified into the number of virtual desktops to allocate in this way, making the action of further allocation clear, measurable, and auditable.

[0165] The same set of priority rules is reused during the differential phase. Blades with existing allocation records are given the first priority, blades that meet the requirement of having more than or equal to the preset maximum number of allocated virtual machines are given the second priority, and blades that have been allocated at least one virtual machine are given the third priority. This consistent strategy reduces maintenance costs.

[0166] It provides two command entry points: "Continue" and "End". If a "Continue" command is received, only the difference in allocation is replenished, and the optimal number of target blades is quickly selected from multiple blades to meet the needs of elastic scaling scenarios. If a "Continue" command is received, the acquired resources are created immediately, shortening the delivery time and making it suitable for one-time delivery or demo scenarios. The actual number of virtual machines allocated is immediately released back to the resource pool, and the scheduler can allocate them to other tenants, improving overall machine utilization by 3%–8%. If "Continue" is selected, the difference in allocation is still subject to the N_remain hard upper limit constraint, which will not cause overload.

[0167] It should be understood that the method steps in the embodiments of the present invention can be implemented or carried out by computer hardware, a combination of hardware and software, or by computer instructions stored in a non-transitory computer-readable storage medium. The method can use standard programming techniques. Each program can be implemented in a high-level procedural or object-oriented programming language to communicate with the computer system. However, if necessary, the program can be implemented in assembly or machine language. In any case, the language can be a compiled or interpreted language. Furthermore, for this purpose, the program can run on a programmed application-specific integrated circuit (ASIC).

[0168] Furthermore, the procedures described herein may be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by the context. The procedures described herein (or variations and / or combinations thereof) may be executed under the control of one or more computer systems configured with executable instructions, and may be implemented by hardware or a combination thereof as code (e.g., executable instructions, one or more computer programs, or one or more applications) that commonly executes on one or more processors. The computer program comprises a plurality of instructions executable by one or more processors.

[0169] Furthermore, the method can be implemented in any suitable type of computing platform, including but not limited to personal computers, minicomputers, mainframes, workstations, networked or distributed computing environments, standalone or integrated computer platforms, or in communication with charged particle tools or other imaging devices, etc. Aspects of the invention can be implemented as machine-readable code stored on a non-transitory storage medium or device, whether removable or integrated into a computing platform, such as a hard disk, optical read and / or write storage medium, RAM, ROM, etc., such that it is readable by a programmable computer, and when the storage medium or device is read by the computer, it can be used to configure and operate the computer to perform the processes described herein. Furthermore, the machine-readable code, or portions thereof, can be transmitted via wired or wireless networks. The invention described herein includes these and other different types of non-transitory computer-readable storage media when such media comprises instructions or programs that implement the steps described above in conjunction with a microprocessor or other data processor. When programmed according to the methods and techniques described in the invention, the invention may also include the computer itself.

[0170] A computer program can be applied to input data to perform the functions described herein, thereby transforming the input data to generate output data stored in non-volatile memory. The output information can also be applied to one or more output devices, such as a display. In a preferred embodiment of the invention, the transformed data represents physical and tangible objects, including specific visual depictions of physical and tangible objects generated on the display.

[0171] The above description is merely a preferred embodiment of the present invention. The present invention is not limited to the above-described embodiments. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention, as long as they achieve the technical effects of the present invention by the same means, should be included within the scope of protection of the present invention. Within the scope of protection of the present invention, the technical solutions and / or implementation methods can have various modifications and variations.

Claims

1. A resource intelligent allocation method for virtual desktops based on multidimensional constraints, characterized in that, include: Identify the CPU performance type of the blade server and dynamically adjust the super-resolution coefficient based on the CPU performance type of the blade server; Calculate the total number of virtual CPU cores, and based on the total number of virtual CPU cores, combined with the user's virtual desktop specification requirements, blade physical resources, and the number of virtual machines already allocated to the blade, calculate the theoretical number of virtual machines to be allocated. The actual number of virtual machines allocated is calculated based on the total number of virtual CPU cores and the upper limit of blade physical resource specifications. Based on the theoretical number of virtual machines allocated and the actual number of virtual machines allocated, determine whether the allocation result meets the multi-dimensional constraints. When the allocation result satisfies multi-dimensional constraints, an allocation priority rule is established, and the optimal target blade is selected based on the allocation priority rule. The optimal target blade is used to perform virtual machine resource allocation and update the virtual machine resource status.

2. The resource intelligent allocation method for virtual desktops based on multidimensional constraints according to claim 1, characterized in that, The CPU performance types of the identified blade servers include: The CPU performance type of the blade server is identified by matching the CPU model with a preset performance library. The CPU performance type of the blade server includes the high-performance library of the Intel Xeon Platinum series and the low-power library of the ARM architecture. The corresponding performance correction factor is automatically determined based on the CPU performance type of the blade server.

3. The resource intelligent allocation method for virtual desktops based on multidimensional constraints according to claim 2, characterized in that, The dynamic adjustment of the super-resolution coefficient based on the CPU performance type of the blade server includes: Obtain the base super-resolution coefficient and the performance correction coefficient corresponding to the CPU performance type of the blade server; The actual super-resolution coefficient is obtained based on the base super-resolution coefficient corresponding to the CPU performance type of the blade server and the performance correction coefficient. The formula for calculating the actual over-resolution coefficient is as follows: K = K0 × Kp, Where K is the actual super-resolution coefficient, K0 is the basic super-resolution coefficient, and Kp is the performance correction coefficient.

4. The resource intelligent allocation method for virtual desktops based on multidimensional constraints according to claim 1, characterized in that, The step of establishing allocation priority rules and selecting the optimal target blade based on the allocation priority rules includes: Establish blade priority sorting rules: blades with existing allocation records are set as first priority, blades with the number of allocated virtual machines greater than or equal to the preset upper limit are set as second priority, and blades with at least one allocated virtual machine are set as third priority; blades are filtered from high to low priority, and the first blade that meets the conditions is selected as the optimal allocation target blade; if there are multiple blades with the same priority, the blade with the most allocated virtual machines is selected as the optimal allocation target blade.

5. The resource intelligent allocation method for virtual desktops based on multidimensional constraints according to claim 3, characterized in that, The calculation of the total number of virtual CPU cores includes: Obtain the physical CPU core count of the blade server and the actual super-resolution coefficient; , The total number of virtual CPU cores is calculated based on the physical CPU core count and the actual super-resolution coefficient. The formula for calculating the total number of virtual CPU cores is as follows: in, This represents the total number of virtual CPU cores. K represents the physical number of CPU cores in the blade server, and K represents the actual super-resolution coefficient.

6. The resource intelligent allocation method for virtual desktops based on multidimensional constraints according to claim 5, characterized in that, The calculation of the theoretical number of virtual machines to be allocated based on the total number of virtual CPU cores, combined with the user's virtual desktop specification requirements, blade physical resources, and the number of virtual machines already allocated to the blade, includes: Analyze user virtual desktop specification requirements to determine whether the display requirement is single-screen or dual-screen display, and then determine the single / dual-screen coefficient; Calculate the difference between the total number of virtual CPU cores and the number of allocated virtual CPUs; Obtain the virtual CPU quantity requirement requested by the user, and calculate the product of the virtual CPU quantity requirement requested by the user and the single / dual screen coefficient; The theoretical number of virtual machines allocated in the CPU dimension is calculated by multiplying the difference between the total number of virtual CPU cores and the number of allocated virtual CPUs, the user's requested virtual CPU quantity requirement, and the single / dual-screen coefficient. The formula for calculating the theoretical number of virtual machines allocated in the CPU dimension is as follows: , in, Theoretically, how many virtual machines should be allocated based on CPU capacity? This represents the total number of virtual CPU cores. The number of virtual CPUs already allocated. For the number of virtual CPUs requested by the user, For single / dual screen coefficients.

7. The resource intelligent allocation method for virtual desktops based on multidimensional constraints according to claim 6, characterized in that, Also includes: Get the total physical memory and allocated physical memory of the blade server; Analyze user memory requirements for virtual desktop specifications; Calculate the difference between the total physical memory of the blade server and the allocated physical memory and reserved memory; The theoretical number of virtual machines to be allocated based on memory is calculated using the difference between the total physical memory of the blade server, the allocated physical memory, and the reserved memory, along with the user's memory requirement. The formula for calculating the quantity is: , in, The theoretical number of virtual machines to allocate in terms of memory dimension. This refers to the total physical memory of the blade server. This is the allocated physical memory; 8GB is reserved memory. User memory requirements; The theoretical number of virtual machines is determined based on the theoretical number of virtual machines allocated in terms of CPU and memory. The formula for the theoretical number of virtual machines allocated is as follows: , in, Theoretically allocate a number of virtual machines. Theoretically, how many virtual machines should be allocated based on CPU capacity? The theoretical number of virtual machines to be allocated in terms of memory.

8. The resource intelligent allocation method for virtual desktops based on multidimensional constraints according to claim 7, characterized in that, The calculation of the actual number of virtual machines allocated based on the total number of virtual CPU cores and the upper limit of blade physical resource specifications includes: Obtain the upper limit of blade physical resource specifications based on the blade model of the blade server; The remaining number of allocable virtual machines is determined based on the upper limit of the blade physical resource specifications and the number of virtual machines already allocated. The formula for calculating the remaining number of allocable virtual machines is as follows: , in, The remaining number of allocable virtual machines. This represents the upper limit of the physical resource specifications for the blade. The number of virtual machines allocated to the blade; The actual number of virtual machines allocated is determined based on the remaining allocable virtual machine count and the theoretical allocation count. The formula for calculating the actual allocation count is as follows: , in, The actual number of virtual machines allocated. Theoretically allocate a number of virtual machines. The remaining number of allocable virtual machines. This indicates rounding to the nearest integer.

9. The intelligent resource allocation method for virtual desktops based on multidimensional constraints according to claim 8, wherein determining whether the allocation result satisfies the multidimensional constraints based on the theoretical allocation number of the virtual machines and the actual allocation number of the virtual machines includes: Determine whether the theoretical number of virtual machines to be allocated is less than or equal to the number of remaining allocable virtual machines; If the theoretical number of virtual machines to be allocated is less than or equal to the number of remaining allocable virtual machines, and the allocation result satisfies the multi-dimensional constraints, then the actual number of virtual machines allocated is determined to be the theoretical number of virtual machines to be allocated. If the theoretical number of virtual machines to be allocated is not less than or equal to the number of remaining allocable virtual machines, and the allocation result does not meet the multi-dimensional constraints, then the actual number of virtual machines to be allocated is determined to be the number of remaining allocable virtual machines.

10. A computer device comprising a memory and a processor, characterized in that, When the processor executes a computer program stored in the memory, it performs the method as described in any one of claims 1 to 9.

Images