Partition management method and device of hyper-converged architecture, electronic equipment and storage medium

Abstract

Embodiments of the present application provide a partition management method and device of an ultra-converged architecture, electronic equipment and a storage medium. The method comprises: dividing a CPU system into a support area and a business area according to business nature; collecting node load data of an ultra-converged node in the support area in multiple dimensions; dividing the support area into a shared support area and a private support area according to the node load data; and planning service resources of the support area and the business area according to total amounts of resources corresponding to the shared support area and the private support area respectively. Embodiments of the present application can achieve logical isolation of services of the support area and the business area, guarantee service quality, and at the same time, refine resource management and control of the support area, and improve service quality of support services.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a partition management method, apparatus, electronic device and storage medium for a hyperconverged architecture. Background Technology

[0002] Against the backdrop of the rapid development of cloud computing, various forms of cloud computing products are growing rapidly, ranging from large-scale full-stack cloud platforms to small- to medium-sized converged cloud platforms. Cloud platforms can generally be divided into five categories: management, computing, storage, networking, and cloud management. Management, storage, networking, and cloud management do not directly provide computing power to the business; these are the support systems. Computing directly provides computing power to the business; these are referred to as the business systems. In some small- to medium-sized cloud platforms, these five categories of components are often deployed in a converged manner to reduce resource consumption; this architecture is called hyperconverged cloud architecture. In hyperconverged cloud architecture, resource contention frequently occurs between the various components, leading to a decline in cloud service quality and instability.

[0003] Currently, in hyperconverged cloud architectures, to reduce the pressure on business systems from the support area (including management, network, storage, and cloud management), cgroups are commonly used to limit CPU (Central Processing Unit) resource usage. However, the actual CPU regions used for execution are not bound (e.g., CN110489232A only limits memory size, not regions) or are statically limited. This still leads to resource contention between the support system and business systems, resulting in incomplete resource isolation. As system pressure increases, resource contention occurs, causing a decline in the performance of business services and affecting the quality of service delivery. Methods for limiting resources in individual modules typically employ the following two approaches:

[0004] I. Global shared converged mode, which specifies management (4 cores), cloud management (4 cores), network (4 cores), storage (4 cores), computing (16 cores), and all resources are shared within the hyperconverged node;

[0005] II. Static partition fusion mode: The management service uses 4 CPUs, numbered 0-3; the cloud management service uses 4 CPUs, numbered 4-7; the network service uses 4 CPUs, numbered 8-12; the storage service uses 4 CPUs, numbered 12-15; and the computing service uses 16 CPUs, numbered 16-31. Each service has its own CPU resources, which are independent of each other and do not interfere with each other.

[0006] The aforementioned global shared and integrated model limits the total amount of resources used by each service. While global allocation maximizes resource utilization, it cannot eliminate interference between services, especially under heavy load, making it difficult to guarantee the quality of critical services. Furthermore, because the total resource usage of each service is limited, it cannot adjust resource requirements based on actual service downtime, resulting in poor flexibility.

[0007] The static partitioning fusion mode described above limits the total resources used by each service and allows it to run on its corresponding CPU, eliminating interference between components but failing to maximize resource utilization. Furthermore, the static limitation on the total resource usage of each service prevents adjustments to resource requirements based on unforeseen circumstances, resulting in poor flexibility. Summary of the Invention

[0008] The technical problem to be solved by the embodiments of this application is to provide a partition management method, device, electronic device and storage medium for hyperconverged architecture, so as to realize the logical isolation between the services of the support area and the services of the business area, ensure the quality of service, and at the same time, refine the resource control of the support area and improve the service quality of the support services.

[0009] In a first aspect, embodiments of this application provide a partition management method for a hyperconverged architecture, the method comprising:

[0010] The CPU system is divided into a support area and a service area based on the nature of the business.

[0011] Collect node load data of hyperconverged nodes in the support area across multiple dimensions;

[0012] Based on the node load data, the support area is divided into a shared support area and a dedicated support area;

[0013] Based on the total resources corresponding to the shared support area and the dedicated support area, the service resources of the support area and the business area are planned.

[0014] Optionally, the collection of node load data of hyperconverged nodes within the support area across multiple dimensions includes:

[0015] According to the set data collection cycle, the node load data of the hyperconverged node is collected in multiple dimensions.

[0016] Optionally, the step of collecting node load data of the hyperconverged node in multiple dimensions according to a set data collection cycle includes:

[0017] Using a probe-based acquisition method, node load data of the hyperconverged node is collected in multiple dimensions according to the set data acquisition cycle; or

[0018] Using a pre-defined script, node load data of the hyperconverged node is collected in multiple dimensions according to the pre-defined data collection cycle.

[0019] Optionally, dividing the support area into a shared support area and a dedicated support area based on the node load data includes:

[0020] Obtain the weights corresponding to multiple dimensions;

[0021] Calculate the super-resolution function corresponding to the hyperconverged node based on the weights and the corresponding node load data;

[0022] Based on the super-resolution function, the support region is divided into the shared support region and the exclusive support region.

[0023] Optionally, the step of planning the service resources of the support area and the business area based on the total resources corresponding to the shared support area and the dedicated support area respectively includes:

[0024] Based on the super-resolution function, determine the resource allocation for each support service within the support area;

[0025] Based on the resource allocation amount and super-division function of each support service, the service resources in the dedicated support area and the shared support area are merged to obtain the total shared resources of the shared support area and the total dedicated resources of the dedicated support area.

[0026] Based on the total amount of shared resources and the total amount of dedicated resources, plan the service resources for the support area and the business area.

[0027] Optionally, the step of planning the service resources of the support area and the business area based on the total amount of shared resources and the total amount of dedicated resources includes:

[0028] The total amount of resources corresponding to the support area is determined based on the total amount of shared resources and the total amount of dedicated resources;

[0029] Allocate the number of CPU cores within the CPU that matches the total resource amount to the support area;

[0030] Allocate the CPU cores other than the number of CPU cores corresponding to the total resource amount to the service area.

[0031] Optionally, the supporting services include: management services, storage services, network services, and operation and maintenance services.

[0032] Optionally, after planning the service resources of the support area and the business area based on the total resources corresponding to the shared support area and the dedicated support area respectively, the method further includes:

[0033] Based on the node load data, predict the node load data of the CPU system in the future.

[0034] Based on the predicted node load data, determine the resource adjustment function of the CPU system;

[0035] The service resources of the support area and the business area are adjusted according to the resource adjustment function.

[0036] Secondly, embodiments of this application provide a partition management device for a hyperconverged architecture, the device comprising:

[0037] The CPU system partitioning module is used to divide the CPU system into support areas and service areas according to the nature of the business.

[0038] The node load acquisition module is used to collect node load data of hyperconverged nodes in the support area in multiple dimensions;

[0039] The support area division module is used to divide the support area into a shared support area and a dedicated support area based on the node load data.

[0040] The service resource planning module is used to plan the service resources of the support area and the business area based on the total resources corresponding to the shared support area and the dedicated support area, respectively.

[0041] Optionally, the node load acquisition module includes:

[0042] The node load acquisition unit is used to collect node load data of the hyperconverged node in multiple dimensions according to a set data acquisition cycle.

[0043] Optionally, the node load acquisition unit includes:

[0044] The first node load acquisition subunit is used to acquire node load data of the hyperconverged node in multiple dimensions according to the set data acquisition cycle using a probe acquisition method.

[0045] The second node load acquisition subunit is used to collect node load data of the hyperconverged node in multiple dimensions according to the set data acquisition cycle using a set script.

[0046] Optionally, the support area division module includes:

[0047] The dimension weight acquisition unit is used to obtain the weights corresponding to multiple dimensions;

[0048] The super-resolution function calculation unit is used to calculate the super-resolution function corresponding to the hyper-converged node based on the weight and the corresponding node load data.

[0049] The support region division unit is used to divide the support region into the shared support region and the exclusive support region according to the super-resolution function.

[0050] Optionally, the service resource planning module includes:

[0051] The resource allocation determination unit is used to determine the resource allocation for each support service within the support area based on the super-resolution function.

[0052] The resource total acquisition unit is used to merge the service resources in the dedicated support area and the shared support area according to the resource allocation amount and the super-division function of each support service, so as to obtain the total shared resources of the shared support area and the total dedicated resources of the dedicated support area.

[0053] The service resource planning unit is used to plan the service resources of the support area and the business area based on the total amount of shared resources and the total amount of dedicated resources.

[0054] Optionally, the service resource planning unit includes:

[0055] The total resource quantity determination subunit is used to determine the total resource quantity corresponding to the support area based on the total shared resources and the total dedicated resources;

[0056] The first CPU core allocation subunit is used to allocate the number of CPU cores in the CPU that matches the total resource amount to the support area;

[0057] The second CPU core allocation subunit is used to allocate the number of CPU cores other than the number of CPU cores corresponding to the total resource amount to the service area.

[0058] Optionally, the supporting services include: management services, storage services, network services, and operation and maintenance services.

[0059] Optionally, the device further includes:

[0060] The load data prediction module is used to predict the node load data of the CPU system in the future based on the node load data.

[0061] The resource adjustment function determination module is used to determine the resource adjustment function of the CPU system based on the predicted node load data.

[0062] The service resource adjustment module is used to adjust the service resources of the support area and the business area according to the resource adjustment function.

[0063] Thirdly, embodiments of this application provide an electronic device, including:

[0064] A processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the partition management method of the hyperconverged architecture described above.

[0065] Fourthly, embodiments of this application provide a computer-readable storage medium that, when the instructions in the storage medium are executed by a processor of an electronic device, enables the electronic device to perform the partition management method of the hyperconverged architecture described in any of the preceding claims.

[0066] Compared with the prior art, the embodiments of this application have the following advantages:

[0067] In this embodiment, the CPU system is divided into a support zone and a service zone based on the nature of the business. Node load data of hyperconverged nodes within the support zone is collected across multiple dimensions. Based on this load data, the support zone is further divided into a shared support zone and a dedicated support zone. Service resources for the support zone and the service zone are planned according to the total resource volume corresponding to each zone. This embodiment effectively solves the resource contention problem in hyperconverged cloud architecture through partition binding technology, and this can be achieved with controllable engineering costs. It effectively addresses the service quality and resource utilization of hyperconverged clouds, improving product quality and service capabilities.

[0068] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0069] Figure 1 A flowchart illustrating the steps of a partition management method for a hyperconverged architecture provided in this application embodiment;

[0070] Figure 2 A flowchart illustrating the steps of a node load data acquisition method provided in this application embodiment;

[0071] Figure 3 A flowchart illustrating the steps of another node load data acquisition method provided in this application embodiment;

[0072] Figure 4 A flowchart illustrating the steps of a support region division method provided in this application embodiment;

[0073] Figure 5 A flowchart illustrating the steps of a service resource planning method provided in this application embodiment;

[0074] Figure 6 A flowchart illustrating the steps of a CPU core allocation method provided in this application embodiment;

[0075] Figure 7 A flowchart illustrating the steps of a service resource adjustment method provided in this application embodiment;

[0076] Figure 8 A schematic diagram of a hyperconverged cloud resource monitoring system architecture provided in an embodiment of this application;

[0077] Figure 9 This is a schematic diagram of the structure of a partition isolation system provided in an embodiment of this application;

[0078] Figure 10 This is a schematic diagram of the structure of a web page information extraction device provided in an embodiment of this application;

[0079] Figure 11 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0080] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0081] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. The singular forms “a,” “the,” and “the” used in the embodiments of this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0082] First, the terms mentioned in the embodiments of this application are explained as follows.

[0083] Hyperconverged cloud: It integrates and deploys resources such as management, computing, networking, storage and cloud management in the same unit device to achieve efficient use of resources. It is often used in small-scale computing scenarios such as edge computing, highways, hospitals, and communities.

[0084] Support services: Services that do not directly provide computing power to the business, including management services, storage services, network services, cloud management services, etc.

[0085] Business services: These refer to services that directly provide computing power to businesses, mainly computing services, such as virtual machines and containers.

[0086] To address the resource contention issue in hyperconverged clouds, this application proposes monitoring and analyzing the resource usage of various components. Components are categorized according to their resource consumption, and these categories are isolated into different availability zones. Different availability zones provide different quality assurances, and the size of the availability zones is dynamically adjusted based on the service load, thereby achieving quality assurance for the hyperconverged cloud.

[0087] The technical solution of this application will now be described in detail with reference to specific embodiments.

[0088] Reference Figure 1 This document illustrates a flowchart of the steps involved in a partition management method for a hyperconverged architecture provided in an embodiment of this application. Figure 1 As shown, the partition management method of this hyperconverged architecture may include steps 101, 102, 103 and 104.

[0089] Step 101: Divide the CPU system into support area and business area according to the nature of the business.

[0090] In this embodiment, the CPU system can be divided into a support zone and a service zone according to the nature of the business. Specifically, partitioning technology can be used to divide the CPU system into a support zone and a service zone according to the nature of the business. CPU binding technology can be used to isolate the support zone and the service zone. The services in the support zone and the services in the service zone are completely logically isolated to ensure service quality.

[0091] After dividing the CPU system into support and business areas according to the nature of the business, proceed to step 102.

[0092] Step 102: Collect node load data of the hyperconverged nodes in the support area in multiple dimensions.

[0093] In this embodiment, three systems are provided: a hyperconverged cloud resource monitoring system, a hyperconverged cloud partitioning and isolation system, and a hyperconverged cloud dynamic adjustment system. The three systems cooperate with each other to complete the automatic partitioning and isolation function.

[0094] After dividing the CPU system into support and business zones based on business nature, the hyperconverged cloud resource monitoring system can collect node load data of hyperconverged nodes in the support zone across multiple dimensions. In this embodiment, when collecting node load data, it is necessary to be precise down to each module in the hyperconverged architecture, namely management, cloud management, network, storage, and computing, to facilitate subsequent statistics and analysis.

[0095] In practical implementation, a collection cycle can be preset to periodically collect node load data. This implementation process can be combined with... Figure 2 The following is a detailed description.

[0096] Reference Figure 2 The flowchart illustrates the steps of a node load data acquisition method provided in an embodiment of this application, as follows: Figure 2 As shown, the node load data acquisition method may include: step 201.

[0097] Step 201: Collect node load data of the hyperconverged node in multiple dimensions according to the set data collection cycle.

[0098] In this embodiment, the set data acquisition cycle refers to the pre-set cycle for collecting load data of the hyperconverged node. In this example, the set data acquisition cycle can be 30s, 50s, 1min, etc. Specifically, the specific value of the set data acquisition cycle can be determined according to business needs, and this embodiment does not limit it.

[0099] Multiple dimensions can be CPU, memory, disk operations, network traffic, etc.

[0100] When collecting node load data for hyperconverged nodes in a hyperconverged architecture, a set data collection period can be obtained, and then node load data of the hyperconverged node in multiple dimensions can be collected according to the set data collection period.

[0101] In practical applications, node load data can be collected using probes or scripts. This implementation process can be combined with... Figure 3 The following is a detailed description.

[0102] Reference Figure 3 The diagram illustrates a flowchart of another node load data acquisition method provided in an embodiment of this application. Figure 3 As shown, the node load data acquisition method may include steps 301 and 302.

[0103] Step 301: Using a probe acquisition method, collect node load data of the hyperconverged node in multiple dimensions according to the set data acquisition cycle.

[0104] Step 302: Using a pre-defined script, collect node load data of the hyperconverged node in multiple dimensions according to the pre-defined data collection cycle.

[0105] In this embodiment, the node load data can be collected using either a probe-based collection method or a script-based collection method. When collecting node load data, a probe-based collection method can be used to collect node load data of the hyperconverged node across multiple dimensions according to a set data collection period. Alternatively, a script can be used to collect node load data of the hyperconverged node across multiple dimensions according to a set data collection period.

[0106] In practical implementation, SNMP / Shell scripts and other methods can be used to collect load data from hyperconverged nodes. The collector periodically collects data from the probes, defaulting to once per minute, but the collection cycle can be adjusted as needed. This includes data on CPU, memory, disk operations, network traffic, etc., and the overall system load is calculated through weighted averages. Data collection needs to be precise down to each module (management, cloud management, network, storage, computing) to facilitate subsequent statistics and analysis. The collected data is stored in a time-series database and accessed via a RESTful API. This process can be described as follows: Figure 8 As shown, node load data for CPU (Central Processing Unit), Memory, Network, and Disk can be collected using a probe-based acquisition method. The collected node load data is then stored in a time-series database, and a RESTful interface is provided to allow other systems to query the corresponding node load data.

[0107] After collecting node load data of hyperconverged nodes in the support area across multiple dimensions, step 103 is executed.

[0108] Step 103: Based on the node load data, divide the support area into a shared support area and a dedicated support area.

[0109] After collecting node load data of hyperconverged nodes within the support zone across multiple dimensions, the support zone can be divided into a shared support zone and a dedicated support zone based on the node load data of each hyperconverged node. In this embodiment, resources in the dedicated support zone cannot be shared and are only allowed for the user's own use, while resources in the shared support zone can be shared by nodes within the shared support zone.

[0110] In practical implementation, the super-resolution function of the hyperconverged node can be calculated based on the weights of each dimension and the corresponding node load data. Then, the support area is divided into shared support areas and dedicated support areas based on the super-resolution function. This implementation process can be combined with... Figure 4 The following is a detailed description.

[0111] Reference Figure 4 The flowchart illustrates the steps of a support area division method provided in an embodiment of this application, as follows: Figure 4 As shown, the support area division method may include steps 401, 402 and 403.

[0112] Step 401: Obtain the weights corresponding to multiple dimensions.

[0113] In this embodiment, when calculating the super-resolution function, weights corresponding to multiple dimensions can be obtained. Understandably, these multiple dimensions can be: CPU, memory, disk operations, network traffic, etc. Among them, the CPU dimension has a significant impact on the calculation of the super-resolution function. Therefore, a larger weight can be assigned to the CPU dimension, and smaller weights can be assigned to the other dimensions.

[0114] After obtaining the weights corresponding to multiple dimensions, proceed to step 402.

[0115] Step 402: Calculate the super-resolution function corresponding to the hyper-converged node based on the weights and the corresponding node load data.

[0116] After obtaining the weights corresponding to multiple dimensions, the super-resolution function of the hyperconverged node can be calculated based on the weights and the corresponding node load data.

[0117] In practice, the hyperconverged cloud partitioning and isolation system can logically isolate various services of the system based on data from the resource monitoring system, thereby ensuring the service quality of each service.

[0118] The hyperconverged cloud partition isolation system can read the load data of the resource monitoring system, including CPU, memory, network, disk, etc., and calculate the load of the service through weighted calculation. The load is represented by load, with a value of [0,1]. For different load conditions, the system adopts different super-scaling coefficients. The system with high load adopts low super-scaling or even no super-scaling, while the system with low load adopts high super-scaling. This can ensure the service quality of the system and improve the system resource utilization. In order to facilitate system implementation, the super-scaling function can be set to two levels: low load level and high load level. load > 0.5 is the high load level, and load ≤ 0.5 is the low load level. The super-scaling function is set according to the load data, as shown in the following formula (1):

[0119]

[0120] In the above formula (1), load is the system load (0≤load≤1), and k is a constant with a recommended value of 1.

[0121] After calculating the super-resolution function corresponding to the hyperconverged node based on the weight and the corresponding node load data, step 403 is executed.

[0122] Step 403: Based on the super-resolution function, divide the support region into the shared support region and the exclusive support region.

[0123] After calculating the super-resolution function corresponding to the hyperconverged node based on the weights and corresponding node load data, the support area can be divided into a shared support area and a dedicated support area according to the super-resolution function. The division method can be as follows:

[0124] (1) When load > 0.5, R = 1 / 1 = 1, that is, the over-division ratio is 1, which means it is not over-division. This is referred to as the exclusive support area.

[0125] (2) When load≤0.5, R=1 / 0.2=2, that is, the super-resolution ratio is 2, which means that 2 times super-resolution is enabled, and the area is called the shared support area.

[0126] After dividing the support area into shared support area and dedicated support area based on node load data, proceed to step 104.

[0127] Step 104: Based on the total resources corresponding to the shared support area and the dedicated support area, plan the service resources of the support area and the business area respectively.

[0128] After dividing the support area into shared and dedicated support areas based on node load data, service resources for the support area and business area can be planned according to the total resources corresponding to the shared and dedicated support areas, respectively. In specific implementations, the resource allocation for each support service within the support area can be calculated using a super-resolution function, and then the service resources for the support area and business area can be planned by combining the resource allocation and the super-resolution function. This implementation process can be further explained in conjunction with… Figure 5 The following is a detailed description.

[0129] Reference Figure 5 The flowchart illustrates the steps of a service resource planning method provided in an embodiment of this application, as shown below. Figure 5 As shown, the service resource planning method may include steps 501, 502 and 503.

[0130] Step 501: Determine the resource allocation for each support service within the support area based on the super-resolution function.

[0131] In this embodiment, the supporting services may include: management services, storage services, network services, and operation and maintenance services.

[0132] After obtaining the super-resolution function, the resource allocation for each support service within the support area can be determined based on the super-resolution function.

[0133] In the specific implementation, the super-resolution ratio of the support area (management, cloud management, network, storage) services is calculated through the super-resolution function R(load). The resource function is designed as follows. Taking the cloud management service as an example, it is assumed that the load of the cloud management service is 0.2, and the initial number of CPUs used is Cinit. For CPU usage, in order to improve the business experience, it is usually necessary to reserve a certain level to prevent the CPU from running at 100%. The reserved level is represented by T, and T defaults to 0.7. The resource function (i.e., the resource allocation amount in this example) is designed as C(load), as shown in the following formula (2):

[0134]

[0135] In the above formula (2), C init The initial CPU resource usage for the service is 4 by default. The number of CPU resources for the cloud management system can be calculated as: C(0.2)=(4*0.2) / (0.7*2)=0.57 cores.

[0136] After determining the resource allocation for each support service within the support area based on the super-resolution function, proceed to step 502.

[0137] Step 502: Based on the resource allocation amount and super-division function of each support service, merge the service resources in the dedicated support area and the shared support area respectively to obtain the total shared resources of the shared support area and the total dedicated resources of the dedicated support area.

[0138] After determining the resource allocation amount for each support service within the support zone based on the super-resolution function, the service resources in the dedicated support zone and the shared support zone can be merged according to the resource allocation amount of each support service and the super-resolution function to obtain the total shared resources of the shared support zone and the total dedicated resources of the dedicated support zone.

[0139] Specifically, in the above process, the system is divided into dedicated support zones and shared support zones according to different loads. These zones are not necessarily contiguous. To improve system efficiency, resources in the dedicated and shared support zones need to be merged, and logical isolation is also required between them. This is achieved by merging service CPU resources with the same super-score.

[0140] Therefore, the total resource function of the shared support area can be derived as shown in formula (3):

[0141]

[0142] The total resource function of the dedicated support area is shown in the following formula (4):

[0143]

[0144] In formulas (3) and (4) above, C share (load) represents the total amount of shared resources, C monopoly (load) represents the total amount of dedicated resources.

[0145] To ensure service operation, the results of resource functions are all rounded up. For example, in C... share (load) = Γ5.1 = 6.

[0146] After merging the service resources in the dedicated support area and the shared support area according to the resource allocation amount and super-division function of each support service to obtain the total shared resources in the shared support area and the total dedicated resources in the dedicated support area, step 503 is executed.

[0147] Step 503: Based on the total amount of shared resources and the total amount of dedicated resources, plan the service resources for the support area and the business area.

[0148] After obtaining the total shared resources of the shared support zone and the total dedicated resources of the dedicated support zone, service resources for the support zone and business zone can be planned based on these totals. Specifically, the total resources required for the support zone can be determined based on the total shared and dedicated resources, and then service resources can be allocated according to the total resources. This implementation process can be combined with... Figure 6 The following is a detailed description.

[0149] Reference Figure 6 The flowchart illustrates the steps of a CPU core allocation method provided in an embodiment of this application, as shown below. Figure 6 As shown, the CPU core allocation method may include steps 601, 602 and 603.

[0150] Step 601: Determine the total amount of resources corresponding to the support area based on the total amount of shared resources and the total amount of exclusive resources.

[0151] In this embodiment, after obtaining the total amount of shared resources and the total amount of dedicated resources, the total amount of resources corresponding to the support area can be determined based on these two amounts. That is, the total amount of resources required by the support area. In other words, total resources = total amount of shared resources + total amount of dedicated resources.

[0152] After determining the total amount of resources corresponding to the support area based on the total amount of shared resources and the total amount of exclusive resources, proceed with steps 602 and 603.

[0153] Step 602: Allocate the number of CPU cores within the CPU that matches the total resource quantity to the support area.

[0154] Step 603: Allocate the CPU cores other than the number of CPU cores corresponding to the total resource amount to the service area.

[0155] After determining the total resources corresponding to the support zone based on the total shared resources and the total dedicated resources, the number of CPU cores within the CPU that match the total resources can be allocated to the support zone. The remaining CPU cores, excluding those corresponding to the total resources, can be allocated to the service zone.

[0156] In this embodiment, cgroup and other technologies can be used to bind the resources of the service to form logically independent shared areas and dedicated areas, thereby ensuring the quality of service operation.

[0157] In this embodiment, the hyperconverged cloud dynamic adjustment system can also dynamically adjust the service resources of the support area and the business area. This implementation process can be combined with... Figure 7 The following is a detailed description.

[0158] Reference Figure 7 The flowchart illustrates the steps of a service resource adjustment method provided in an embodiment of this application, as shown below. Figure 7 As shown, the service resource adjustment method may include steps 701, 702 and 703.

[0159] Step 701: Based on the node load data, predict the predicted node load data of the CPU system in the future.

[0160] In this embodiment, after obtaining the node load data, the predicted node load data of the CPU system in the future can be predicted based on the node load data.

[0161] In its implementation, the hyperconverged cloud dynamic adjustment system continuously verifies the current resource usage of each service by monitoring the load data of the resource monitoring system. Currently, the data collection period is 1 minute. The system analyzes and predicts the system load by comprehensively analyzing the system's load usage over a past period. The system uses an LSTM neural network algorithm for load prediction. The LSTM neural network algorithm is mainly based on time series correlation prediction. In this embodiment, it mainly uses the load values ​​over a past period (default set to 3 days) to predict future values, making advance judgments on the system load situation and improving system response speed and service quality.

[0162] After predicting the node load data of the CPU system in the future based on the node load data, step 702 is executed.

[0163] Step 702: Based on the predicted node load data, determine the resource adjustment function of the CPU system.

[0164] After predicting the node load data of the CPU system in the future based on the node load data, the resource adjustment function of the CPU system can be determined based on the predicted node load data.

[0165] This embodiment pre-sets a resource adjustment function. Specifically, the resource adjustment function for the shared support area is shown in the following formula (5):

[0166]

[0167] The resource adjustment function for the dedicated support area is shown in the following formula (6):

[0168]

[0169] In formulas (5) and (6) above, T is the pressure water level, which is generally set to 0.7. It can be adjusted according to system requirements. The load value within the most recent period (default setting is 3 days) is input into the LSTM algorithm to obtain the predicted value. The predicted value is used to adjust the system partition boundary. To ensure service operation, the result of the resource adjustment function is rounded up, that is, if CJshare=Γ5.1=6.

[0170] After determining the CPU system resource adjustment function based on the predicted node load data, step 703 is executed.

[0171] Step 703: Adjust the service resources of the support area and the business area according to the resource adjustment function.

[0172] After determining the CPU system resource adjustment function based on the predicted node load data, the service resources of the support area and the business area can be adjusted according to the resource adjustment function.

[0173] Against the backdrop of the rapid development of cloud computing, various cloud computing models have made significant progress, with hyperconverged cloud architecture offering substantial advantages in resource utilization and deployment. However, the isolation and resource contention issues within hyperconverged cloud architectures have long been challenging problems. This embodiment effectively addresses resource contention in hyperconverged cloud architectures through partitioning and dynamic adjustment technologies, achieving this at a controllable engineering cost. This significantly improves service quality and resource utilization in hyperconverged clouds, enhancing product quality and service capabilities. The technical solution presented in this embodiment will play a significant role in hyperconverged cloud architectures, enabling rapid deployment in scenarios such as edge computing, mobile offices, highways, hospitals, and communities, addressing the shortcomings of hyperconverged cloud products and significantly enhancing their competitiveness.

[0174] The technical solutions of the embodiments of this application can be combined with Figure 9 The following is a detailed description.

[0175] Reference Figure 9 This diagram illustrates a partition isolation system according to an embodiment of this application. This partition isolation system can logically isolate various services of the system based on data from a resource monitoring system, ensuring the service quality of each service. The logic of the partition isolation system can be as follows:

[0176] Step 1: Allocate high-load storage and network resources to dedicated support areas and isolate them. Calculate the super-division function R and perform super-division of resources.

[0177] Step 2: Merge the management and operation management areas with lower loads into a shared support area, and enable resource super-division through super-division function calculation;

[0178] Step 3: Allocate all remaining resources to the business area to maximize resource usage for the business.

[0179] The partition management method for hyperconverged architecture provided in this application divides the CPU system into support zones and service zones based on business nature. It collects node load data of hyperconverged nodes within the support zone across multiple dimensions. Based on this load data, the support zone is further divided into shared support zones and dedicated support zones. Service resources for the support zone and service zone are planned according to the total resource volume corresponding to each zone. This application effectively solves the resource contention problem in hyperconverged cloud architecture through partition binding technology, and can be implemented with controllable engineering costs. It effectively addresses service quality and resource utilization in hyperconverged clouds, improving product quality and service capabilities.

[0180] Reference Figure 10 The diagram illustrates a structural schematic of a partition management device for a hyperconverged architecture provided in an embodiment of this application. Figure 10 As shown, the partition management device 1000 of the hyperconverged architecture may include the following modules:

[0181] CPU system partitioning module 1001 is used to divide the CPU system into support area and service area according to the nature of the business;

[0182] The node load acquisition module 1002 is used to collect node load data of hyperconverged nodes in the support area in multiple dimensions.

[0183] The support area division module 1003 is used to divide the support area into a shared support area and a dedicated support area based on the node load data.

[0184] The service resource planning module 1004 is used to plan the service resources of the support area and the business area according to the total amount of resources corresponding to the shared support area and the dedicated support area respectively.

[0185] Optionally, the node load acquisition module includes:

[0186] The node load acquisition unit is used to collect node load data of the hyperconverged node in multiple dimensions according to a set data acquisition cycle.

[0187] Optionally, the node load acquisition unit includes:

[0188] The first node load acquisition subunit is used to acquire node load data of the hyperconverged node in multiple dimensions according to the set data acquisition cycle using a probe acquisition method.

[0189] The second node load acquisition subunit is used to collect node load data of the hyperconverged node in multiple dimensions according to the set data acquisition cycle using a set script.

[0190] Optionally, the support area division module includes:

[0191] The dimension weight acquisition unit is used to obtain the weights corresponding to multiple dimensions;

[0192] The super-resolution function calculation unit is used to calculate the super-resolution function corresponding to the hyper-converged node based on the weight and the corresponding node load data.

[0193] The support region division unit is used to divide the support region into the shared support region and the exclusive support region according to the super-resolution function.

[0194] Optionally, the service resource planning module includes:

[0195] The resource allocation determination unit is used to determine the resource allocation for each support service within the support area based on the super-resolution function.

[0196] The resource total acquisition unit is used to merge the service resources in the dedicated support area and the shared support area according to the resource allocation amount and the super-division function of each support service, so as to obtain the total shared resources of the shared support area and the total dedicated resources of the dedicated support area.

[0197] The service resource planning unit is used to plan the service resources of the support area and the business area based on the total amount of shared resources and the total amount of dedicated resources.

[0198] Optionally, the service resource planning unit includes:

[0199] The total resource quantity determination subunit is used to determine the total resource quantity corresponding to the support area based on the total shared resources and the total dedicated resources;

[0200] The first CPU core allocation subunit is used to allocate the number of CPU cores in the CPU that matches the total resource amount to the support area;

[0201] The second CPU core allocation subunit is used to allocate the number of CPU cores other than the number of CPU cores corresponding to the total resource amount to the service area.

[0202] Optionally, the supporting services include: management services, storage services, network services, and operation and maintenance services.

[0203] Optionally, the device further includes:

[0204] The load data prediction module is used to predict the node load data of the CPU system in the future based on the node load data.

[0205] The resource adjustment function determination module is used to determine the resource adjustment function of the CPU system based on the predicted node load data.

[0206] The service resource adjustment module is used to adjust the service resources of the support area and the business area according to the resource adjustment function.

[0207] The hyperconverged infrastructure partition management device provided in this application divides the CPU system into a support zone and a service zone according to the nature of the business. It collects node load data of hyperconverged nodes within the support zone across multiple dimensions. Based on the node load data, the support zone is further divided into a shared support zone and a dedicated support zone. Service resources for the support zone and the service zone are planned according to the total resource volume corresponding to each zone. This application embodiment effectively solves the resource contention problem in hyperconverged cloud architecture through partition binding technology, and it can be implemented with controllable engineering costs. It can effectively address the service quality and resource utilization of hyperconverged clouds, improving product quality and service capabilities.

[0208] This application also provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the partition management method of the hyperconverged architecture described above.

[0209] Figure 11 A schematic diagram of the structure of an electronic device 1100 according to an embodiment of the present invention is shown. For example... Figure 11As shown, the electronic device 1100 includes a central processing unit (CPU) 1101, which can perform various appropriate actions and processes according to computer program instructions stored in read-only memory (ROM) 1102 or loaded from storage unit 1108 into random access memory (RAM) 1103. The RAM 1103 may also store various programs and data required for the operation of the electronic device 1100. The CPU 1101, ROM 1102, and RAM 1103 are interconnected via bus 1104. An input / output (I / O) interface 1105 is also connected to bus 1104.

[0210] Multiple components in electronic device 1100 are connected to I / O interface 1105, including: input unit 1106, such as keyboard, mouse, microphone, etc.; output unit 1107, such as various types of monitors, speakers, etc.; storage unit 1108, such as disk, optical disk, etc.; and communication unit 1109, such as network card, modem, wireless transceiver, etc. Communication unit 1109 allows electronic device 1100 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0211] The various processes and handling described above can be executed by processing unit 1101. For example, the methods of any of the above embodiments can be implemented as computer software programs tangibly contained in a computer-readable medium, such as storage unit 1108. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 1100 via ROM 1102 and / or communication unit 1109. When the computer program is loaded into RAM 1103 and executed by CPU 1101, one or more actions of the methods described above can be performed.

[0212] Additionally, embodiments of this application also provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the partition management method of the hyperconverged architecture described above.

[0213] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0214] Those skilled in the art will understand that embodiments of this application can be provided as methods, apparatus, or computer program products. Therefore, embodiments of this application can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of this application can take the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0215] This application describes embodiments with reference to flowchart illustrations and / or block diagrams of methods, terminals (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0216] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0217] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal, causing a series of operational steps to be executed on the computer or other programmable terminal to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0218] Although preferred embodiments of the present application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present application.

[0219] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal that includes said element.

[0220] The foregoing has provided a detailed description of a partition management method, a partition management device, an electronic device, and a computer-readable storage medium for a hyperconverged architecture. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A partition management method for a hyperconverged architecture, characterized in that, The method includes: The CPU system is divided into a support area and a service area based on the nature of the business. Among them, the support services in the support area are services that do not directly provide computing power to the business; The business services in the aforementioned business area are those that directly provide computing power to the business. Collect node load data of hyperconverged nodes in the support area across multiple dimensions; Based on the node load data, the support area is divided into a shared support area and a dedicated support area; Based on the total resources corresponding to the shared support area and the dedicated support area, plan the service resources for the support area and the business area respectively; The step of dividing the support area into a shared support area and a dedicated support area based on the node load data includes: Obtain the weights corresponding to multiple dimensions; Calculate the super-resolution function corresponding to the hyperconverged node based on the weights and the corresponding node load data; Based on the super-resolution function, the support region is divided into the shared support region and the exclusive support region; The step of planning service resources for the support zone and the business zone based on the total resources corresponding to the shared support zone and the dedicated support zone respectively includes: Based on the super-resolution function, determine the resource allocation for each support service within the support area; Based on the resource allocation amount and super-division function of each support service, the service resources in the dedicated support area and the shared support area are merged to obtain the total shared resources of the shared support area and the total dedicated resources of the dedicated support area. Based on the total amount of shared resources and the total amount of dedicated resources, plan the service resources for the support area and the business area.

2. The method according to claim 1, characterized in that, The collection of node load data for hyperconverged nodes within the support zone across multiple dimensions includes: According to the set data collection cycle, the node load data of the hyperconverged node is collected in multiple dimensions.

3. The method according to claim 2, characterized in that, The process of collecting node load data of the hyperconverged node across multiple dimensions according to a set data collection cycle includes: Using a probe-based acquisition method, node load data of the hyperconverged node is collected in multiple dimensions according to the set data acquisition cycle; or Using a pre-defined script, node load data of the hyperconverged node is collected in multiple dimensions according to the pre-defined data collection cycle.

4. The method according to claim 1, characterized in that, The step of planning service resources for the support area and the business area based on the total amount of shared resources and the total amount of dedicated resources includes: The total amount of resources corresponding to the support area is determined based on the total amount of shared resources and the total amount of dedicated resources; Allocate the number of CPU cores within the CPU that matches the total resource amount to the support area; Allocate the CPU cores other than the number of CPU cores corresponding to the total resource amount to the service area.

5. The method according to claim 1, characterized in that, The supporting services include: management services, storage services, network services, and operation and maintenance services.

6. The method according to claim 1, characterized in that, After planning the service resources of the support area and the business area based on the total resources corresponding to the shared support area and the dedicated support area respectively, the method further includes: Based on the node load data, predict the node load data of the CPU system in the future. Based on the predicted node load data, determine the resource adjustment function of the CPU system; The service resources of the support area and the business area are adjusted according to the resource adjustment function.

7. A partition management device for a hyperconverged architecture, characterized in that, The device includes: The CPU system partitioning module is used to divide the CPU system into a support zone and a service zone according to the nature of the business; wherein, the support services in the support zone are services that do not directly provide computing power to the business; and the service services in the service zone are services that directly provide computing power to the business. The node load acquisition module is used to collect node load data of hyperconverged nodes in the support area in multiple dimensions; The support area division module is used to divide the support area into a shared support area and a dedicated support area based on the node load data. The service resource planning module is used to plan the service resources of the support area and the business area based on the total resources corresponding to the shared support area and the dedicated support area, respectively. The support area division module includes: The dimension weight acquisition unit is used to obtain the weights corresponding to multiple dimensions; The super-resolution function calculation unit is used to calculate the super-resolution function corresponding to the hyper-converged node based on the weight and the corresponding node load data. A support region division unit is used to divide the support region into the shared support region and the exclusive support region according to the super-resolution function; The service resource planning module includes: The resource allocation determination unit is used to determine the resource allocation for each support service within the support area based on the super-resolution function. The resource total acquisition unit is used to merge the service resources in the dedicated support area and the shared support area according to the resource allocation amount and the super-division function of each support service, so as to obtain the total shared resources of the shared support area and the total dedicated resources of the dedicated support area. The service resource planning unit is used to plan the service resources of the support area and the business area based on the total amount of shared resources and the total amount of dedicated resources.

8. The apparatus according to claim 7, characterized in that, The node load acquisition module includes: The node load acquisition unit is used to collect node load data of the hyperconverged node in multiple dimensions according to a set data acquisition cycle.

9. The apparatus according to claim 8, characterized in that, The node load acquisition unit includes: The first node load acquisition subunit is used to acquire node load data of the hyperconverged node in multiple dimensions according to the set data acquisition cycle using a probe acquisition method. The second node load acquisition subunit is used to collect node load data of the hyperconverged node in multiple dimensions according to the set data acquisition cycle using a set script.

10. An electronic device, characterized in that, include: A processor, a memory, and a computer program stored on the memory and executable on the processor, wherein the processor, when executing the program, implements the partition management method of the hyperconverged architecture according to any one of claims 1 to 6.

11. A computer-readable storage medium, characterized in that, When the instructions in the storage medium are executed by the processor of the electronic device, the electronic device is able to perform the partition management method of the hyperconverged architecture according to any one of claims 1 to 6.

Images