A resource fragmentation management method

By aggregating dimensionality reduction and causal structure analysis of timing indicator data, the causal structure of Bayesian network is generated and resource creation is optimized, the problem of low resource utilization in the existing technology is solved, and efficient fragmentation management of resource pools is realized.

CN118672792BActive Publication Date: 2024-12-06CHINA UNICOM (LIAONING) IND INTERNET CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411162182.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-23
Publication Date
2024-12-06
Estimated Expiration
2044-08-23

AI Technical Summary

Technical Problem

The existing resource pool resource scheduling management cannot cope with instance creation with custom specifications, resulting in a decrease in the overall available resource utilization of the host after the resource instance is created.

Method used

By obtaining the timing indicator data of the target business, performing aggregated dimensionality reduction processing, obtaining adjacent matrices, and substituting them into the causal structure discovery algorithm, generating a directed acyclic graph of the causal structure of Bayesian network, and using SimpleEstimator algorithm to calculate the conditional probability table , combine the RPLoc algorithm to calculate the value of nodes under preset conditions, and optimize resource creation.

Benefits of technology

Reduce the fragmentation of the resource pool, reduce the fragmentation rate of the entire resource pool, and improve resource usage efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118672792B_ABST
    Figure CN118672792B_ABST
Patent Text Reader

Abstract

The embodiment of the present invention discloses a resource fragmentation management method, which relates to the technical field of cloud resource allocation. The method comprises: obtaining time series indicator data of a target business, and performing aggregation and dimensionality reduction processing on the time series indicator data to obtain an adjacent matrix; substituting the adjacent matrix into a causal structure discovery algorithm to obtain a directed acyclic graph of a Bayesian network causal structure; based on the adjacent matrix and the directed acyclic graph, calculating a conditional probability table of each node in the directed acyclic graph by an RPLoc algorithm to obtain a target directed acyclic graph with the conditional probability table; calculating the values ​​of each node in the directed acyclic graph under preset conditions according to the target directed acyclic graph and prior knowledge of the target business, and creating resources according to the values.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of cloud resource allocation, and in particular to a resource fragmentation management method. Background Art

[0002] Most of the existing resource pool resource scheduling management is to divide different resource pools using the same instance specifications. When creating resource instances, the instance creation request is forwarded to the corresponding resource pool with different specifications. This is a fixed behavior pattern and cannot cope with the creation of instances with custom specifications, resulting in a decrease in the overall available resource utilization rate of the host after the resource instance is created. Summary of the invention

[0003] In view of this, the present invention provides a resource fragmentation management method for solving the problem in the prior art that the overall available resource utilization rate of the host machine decreases after the instance is created.

[0004] In a first aspect, an embodiment of the present invention provides a resource fragmentation management method, the method comprising:

[0005] Acquire time series indicator data of the target business, and perform aggregation and dimensionality reduction processing on the time series indicator data to obtain an adjacent matrix;

[0006] Substituting the adjacency matrix into a causal structure discovery algorithm to obtain a directed acyclic graph of the Bayesian network causal structure;

[0007] Based on the adjacency matrix and the directed acyclic graph, a conditional probability table of each node in the directed acyclic graph is calculated by a SimpleEstimator algorithm under preset conditions to obtain a target directed acyclic graph with the conditional probability table;

[0008] According to the prior knowledge of the target directed acyclic graph and the target business, the value of each node in the directed acyclic graph under preset conditions is calculated by the RPLoc algorithm, and resources are created according to the value.

[0009] Optionally, the step of performing aggregation and dimensionality reduction processing on the time series indicator data to obtain an adjacent matrix includes:

[0010] Aggregating the time series indicator data by a non-parametric Bayesian method to obtain a strongly correlated data set in the time series indicator data;

[0011] The adjacent matrix is ​​constructed based on the strongly correlated data set, wherein the elements in the adjacent matrix represent the similarity data between the variable data in the strongly correlated data set.

[0012] Optionally, before the step of substituting the adjacency matrix into the causal structure discovery algorithm to obtain the directed acyclic graph of the Bayesian network causal structure, the method further includes:

[0013] The adjacent matrix is ​​discretized by preset parameters to discretize the data in the adjacent matrix into different designated areas to obtain a discretized target adjacent matrix. The target adjacent matrix is ​​used to substitute into the causal structure discovery algorithm to obtain a directed acyclic graph of the Bayesian network causal structure, wherein different designated areas are used to represent different similarity levels.

[0014] Optionally, the step of substituting the adjacency matrix into a causal structure discovery algorithm to obtain a directed acyclic graph of the Bayesian network causal structure includes:

[0015] Substituting the target adjacent matrix into the conditional independence test algorithm in the causal structure discovery algorithm to obtain the conditional independent variables of each data in the target adjacent matrix;

[0016] Determine the causal relationship connection direction of each conditional independent variable by using the V-structure algorithm and the Meek rules algorithm in the causal structure discovery algorithm;

[0017] A directed acyclic graph of a Bayesian network causal structure is determined based on the causal relationship connection direction and the conditional independent variables.

[0018] Optionally, the step of substituting the adjacency matrix into a causal structure discovery algorithm to obtain a directed acyclic graph of the Bayesian network causal structure further includes:

[0019] Determine the functional relationship between each variable data based on the similarity data;

[0020] Establishing the causal structure discovery algorithm according to the functional relationship;

[0021] Substituting the adjacent matrix into the causal structure discovery algorithm, a directed acyclic graph of the Bayesian network causal structure is obtained.

[0022] Optionally, the step of calculating, based on the adjacency matrix and the directed acyclic graph, a conditional probability table of each node in the directed acyclic graph under preset conditions by a SimpleEstimator algorithm to obtain a target directed acyclic graph with the conditional probability table includes:

[0023] When the preset condition is that the resource fragmentation rate is minimum, according to the adjacent matrix and the directed acyclic graph, the conditional probability data of each node in the directed acyclic graph at different possible values ​​are calculated by the SimpleEstimator algorithm to obtain a conditional probability table of each node in the directed acyclic graph;

[0024] Substituting the conditional probability table into the directed acyclic graph, a target directed acyclic graph with the conditional probability table is obtained.

[0025] Optionally, the method further includes:

[0026] When the preset condition is that the resource fragmentation rate is the minimum, the resource fragmentation rate calculation logic is:

[0027] .

[0028] in, represents the resource fragmentation rate, k is the total number of server nodes, represents the fragmentation rate of the i-th server node, Represents the i-th virtual resource usage instance.

[0029] On the other hand, the present application provides a resource fragmentation management device, the device comprising:

[0030] A data acquisition module is used to obtain the time series indicator data of the target business, and perform aggregation and dimensionality reduction processing on the time series indicator data to obtain an adjacent matrix;

[0031] A first computing module is used to substitute the adjacent matrix into a causal structure discovery algorithm to obtain a directed acyclic graph of a Bayesian network causal structure;

[0032] A second calculation module is used to calculate, based on the adjacent matrix and the directed acyclic graph, a conditional probability table of each node in the directed acyclic graph under preset conditions through a SimpleEstimator algorithm to obtain a target directed acyclic graph with the conditional probability table;

[0033] The allocation module is used to calculate the value of each node in the directed acyclic graph under preset conditions through the RPLoc algorithm according to the prior knowledge of the target directed acyclic graph and the target business, and create resources according to the value.

[0034] In a third aspect, an embodiment of the present invention further provides an electronic device, the electronic device comprising:

[0035] One or more processors;

[0036] A storage device for storing one or more programs;

[0037] When the one or more programs are executed by the one or more processors, the one or more processors implement the resource fragmentation management method as described in any of the embodiments of the present invention.

[0038] In a fourth aspect, an embodiment of the present invention further provides a storage medium comprising computer executable instructions, which, when executed by a computer processor, are used to execute a resource fragmentation management method as described in any of the embodiments of the present invention.

[0039] The technical solution of the embodiment of the present invention obtains the time series indicator data of the target business and performs aggregation and dimensionality reduction processing on the time series indicator data to obtain an adjacent matrix; substitutes the adjacent matrix into the causal structure discovery algorithm to obtain a directed acyclic graph of the causal structure of the Bayesian network; based on the adjacent matrix and the directed acyclic graph, the conditional probability table of each node in the directed acyclic graph under preset conditions is calculated by the SimpleEstimator algorithm to obtain a target directed acyclic graph with the conditional probability table; according to the target directed acyclic graph and the prior knowledge of the target business, the value of each node in the directed acyclic graph under preset conditions is calculated by the RPLoc algorithm, and resources are created according to the values. The resource pool creates resources according to the predicted results (the values ​​of each node in the directed acyclic graph under preset conditions are calculated by the RPLoc algorithm), reducing the generation of fragments in the resource pool and reducing the fragmentation rate of the entire resource pool. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0041] in:

[0042] Figure 1 A schematic diagram of a process of resource fragmentation management method in one embodiment;

[0043] Figure 2 is a schematic diagram of the structure of a resource fragment management device in one embodiment;

[0044] Figure 3 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application;

[0045] Figure 4 It is a structural diagram of a computer-readable storage provided in an embodiment of the present application. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] Before describing the technical solution of the embodiment of the present invention, the application scenario of the embodiment of the present invention is first exemplarily described:

[0048] Based on the massive high-dimensional time series indicators collected for target application scenarios, non-parametric Bayesian methods are used for clustering, and then feature-reduced data sets are extracted; then, the causal relationship between virtual machine instances of different specifications and server node resource occupancy rates is found through probability graphs. Based on the causal relationship law, the goal is to minimize the fragmentation rate, and the possible fragmentation in the future is calculated according to the historical data law (prior knowledge), and the resource fragmentation rate indicator FragRatio (0-100%) is used to measure the fragmentation of the cloud resource pool, and guide the scheduling algorithm to optimize the resource allocation strategy.

[0049] In view of this, an embodiment of the present invention provides a resource fragmentation management method, which obtains the time series indicator data of the target business and performs aggregation and dimensionality reduction processing on the time series indicator data to obtain an adjacent matrix; substitutes the adjacent matrix into the causal structure discovery algorithm to obtain a directed acyclic graph of the causal structure of the Bayesian network; based on the adjacent matrix and the directed acyclic graph, the conditional probability table of each node in the directed acyclic graph is calculated by the SimpleEstimator algorithm under preset conditions to obtain a target directed acyclic graph with the conditional probability table; according to the target directed acyclic graph and the prior knowledge of the target business, the value of each node in the directed acyclic graph under preset conditions is calculated by the RPLoc algorithm, and resources are created according to the values. The resource pool creates resources according to the predicted results (the values ​​of each node in the directed acyclic graph under preset conditions are calculated by the RPLoc algorithm), reducing the generation of fragments in the resource pool and reducing the fragmentation rate of the entire resource pool.

[0050] In one embodiment, the present invention provides a resource fragmentation management method, and the resource fragmentation management method of the present invention can be executed by a resource fragmentation management device. The resource fragmentation management device can be implemented by software and / or hardware.

[0051] like Figure 1 As shown, the resource fragmentation management method of the embodiment of the present invention specifically includes the following steps:

[0052] S110, acquiring time series indicator data of a target service, and performing aggregation and dimensionality reduction processing on the time series indicator data to obtain an adjacent matrix;

[0053] Exemplarily, the process of aggregation and dimensionality reduction is used to eliminate indicators with no causal correlation or weak correlation in the massive input time series indicators (time series indicator data), and find a data clustering set with strong correlation. Commonly used clustering algorithms can be divided into hierarchical clustering methods and partition clustering methods. This article mainly uses the partition clustering method, which is based on calculating the distance between sample data. The distance usually uses Euclidean distance, dynamic time warping distance, etc. to divide the time series indicator data into multiple sets.

[0054] In a possible implementation, the step of performing aggregation and dimensionality reduction processing on the time series indicator data to obtain an adjacent matrix includes:

[0055] Aggregating the time series indicator data by a non-parametric Bayesian method to obtain a strongly correlated data set in the time series indicator data;

[0056] The adjacent matrix is ​​constructed based on the strongly correlated data set, wherein the elements in the adjacent matrix represent the similarity data between the variable data in the strongly correlated data set.

[0057] Exemplarily, the time series indicator data are aggregated by a non-parametric Bayesian clustering method, such as the Chinese restaurant process CRP (Chinese Restaruant Process) based on the Dirichlet process and its derivative methods, to obtain a strongly correlated data set in the time series indicator data, and an adjacent matrix is ​​output based on the strongly correlated data set, wherein the elements in the adjacent matrix represent the similarity data between the variable data in the strongly correlated data set.

[0058] S120, substituting the adjacent matrix into a causal structure discovery algorithm to obtain a directed acyclic graph of a Bayesian network causal structure;

[0059] In a possible implementation, the step of substituting the adjacency matrix into a causal structure discovery algorithm to obtain a directed acyclic graph of a Bayesian network causal structure includes:

[0060] Substituting the target adjacent matrix into the conditional independence test algorithm in the causal structure discovery algorithm to obtain the conditional independent variables of each data in the target adjacent matrix;

[0061] Determine the causal relationship connection direction of each conditional independent variable by using the V-structure algorithm and the Meek rules algorithm in the causal structure discovery algorithm;

[0062] A directed acyclic graph of a Bayesian network causal structure is determined based on the causal relationship connection direction and the conditional independent variables.

[0063] Exemplarily, the constraint based method is implemented by using the conditional independence test method (d / m-separation) to find the correlation between the DAG / MAG verification variables (conditional independent variables), find the conditional independent variables and eliminate the causal connection relationship between them, and then use the V-structure and Meek rules method to find the causal connection direction of each conditional independent variable, and then determine the directed acyclic graph of the Bayesian network causal structure.

[0064] In a possible implementation manner, before the step of substituting the adjacency matrix into the causal structure discovery algorithm to obtain the directed acyclic graph of the Bayesian network causal structure, the step further includes:

[0065] The adjacent matrix is ​​discretized by preset parameters to discretize the data in the adjacent matrix into different designated areas to obtain a discretized target adjacent matrix. The target adjacent matrix is ​​used to substitute into the causal structure discovery algorithm to obtain a directed acyclic graph of the Bayesian network causal structure, wherein different designated areas are used to represent different similarity levels.

[0066] Exemplarily, the data in the adjacency matrix is ​​discretized into discrete sets of different designated areas (such as 5 areas), and the similarity or distance is divided into different levels (high, medium, and low) to obtain a discretized target adjacency matrix.

[0067] In a possible implementation, the step of substituting the adjacency matrix into a causal structure discovery algorithm to obtain a directed acyclic graph of a Bayesian network causal structure further includes:

[0068] Determine the functional relationship between each variable data based on the similarity data;

[0069] Establishing the causal structure discovery algorithm according to the functional relationship;

[0070] Substituting the adjacent matrix into the causal structure discovery algorithm, a directed acyclic graph of the Bayesian network causal structure is obtained.

[0071] Exemplarily, the causal structure discovery algorithm can also be inferred through a hypothesis method, and the functional relationship between each variable data is determined based on the discretized target adjacent matrix. Specifically, the causal structure discovery algorithm is established based on the functional relationship according to the elements of the target adjacent matrix that represent the similarity data between each variable data in the strongly correlated data set, and the adjacent matrix is ​​substituted into the causal structure discovery algorithm to obtain a directed acyclic graph of the Bayesian network causal structure.

[0072] S130, based on the adjacency matrix and the directed acyclic graph, calculating, by the SimpleEstimator algorithm, under preset conditions, a conditional probability table of each node in the directed acyclic graph, to obtain a target directed acyclic graph with the conditional probability table;

[0073] In a possible implementation, the step of calculating, based on the adjacency matrix and the directed acyclic graph, the conditional probability table of each node in the directed acyclic graph under preset conditions by the SimpleEstimator algorithm to obtain a target directed acyclic graph with the conditional probability table includes:

[0074] When the preset condition is that the resource fragmentation rate is minimum, according to the adjacent matrix and the directed acyclic graph, the conditional probability data of each node in the directed acyclic graph at different possible values ​​are calculated by the SimpleEstimator algorithm to obtain a conditional probability table of each node in the directed acyclic graph;

[0075] The SimpleEstimator algorithm is a prediction algorithm that estimates a result based on given data or rules. In minimizing the resource fragmentation rate, the SimpleEstimator algorithm is used to estimate the approximate value of resource fragmentation under different configurations to assist in decision-making.

[0076] It can be understood that when the SimpleEstimator algorithm is used to estimate the approximate value of resource fragmentation under different configurations, the operation logic is: when the preset condition is that the resource fragmentation is minimum, the SimpleEstimator algorithm is applied to the nodes in the acyclic graph (DAG) to calculate the conditional probability data of each node at different possible values; the conditional probability data refers to the probability of change of the resource fragmentation rate under a given resource allocation or node state change. The specific operation logic includes: initialization: read the structure of the directed acyclic graph, including nodes and edges; set the possible value of each node (resource allocation, state, etc.) and the corresponding resource fragmentation estimation function; traverse the node: traverse each node in the DAG; for each node, consider all its possible values; conditional probability calculation: for each possible value of each node, use the estimation function to calculate the resource fragmentation rate; calculate the probability that the resource fragmentation meets a specific condition under given conditions (that is, when the node takes a certain value), which needs to be estimated based on the probability distribution of historical data; construct a conditional probability table: organize the conditional probability data of each node into a table form, where the row represents the node, the column represents the possible value of the node, and the elements in the table are the corresponding conditional probability values.

[0077] For example, assume that there is a simple directed acyclic graph containing three nodes A, B, and C, which represent different resource allocation units. Each node has two possible states: high allocation (H) and low allocation (L). The SimpleEstimator algorithm is used to estimate the minimum value of the resource fragmentation rate under different allocation schemes. The results are shown in Table 1:

[0078] Table 1: Minimum value table of resource fragmentation rate under different allocation schemes

[0079]

[0080] Use the SimpleEstimator algorithm to traverse all possible allocation schemes (such as AAA=HHH, HHL, HLH, ...), calculate the sum of resource fragmentation rates under each scheme, and find the scheme that minimizes the sum of resource fragmentation rates (target scheme). Then, based on the resource fragmentation rate of the target scheme, the conditional probability table of each node can be constructed.

[0081] Substituting the conditional probability table into the directed acyclic graph, a target directed acyclic graph with the conditional probability table is obtained.

[0082] Exemplarily, based on the generated causal network structure, the SimpleEstimator algorithm is executed to generate a conditional probability table CPT (any node X in the network has a corresponding conditional probability table (ConditionalProbabilityTable, CPT), which is used to represent the conditional probability of node X when its parent node takes various possible values. If node X has no parent node, then the CPT of X is its prior probability distribution, and the prior probability distribution is obtained by collecting historical data on resource allocation in the resource pool and obtaining the historical probability distribution (prior probability distribution) according to the historical data. The structure of the Bayesian network and the CPT of each node define the probability distribution of each variable data in the network. Repeat the potential node selection and parent node configuration traversal, perform conditional probability calculation, fill the calculated conditional probability into the CPT, and form a complete CPT of node X until all nodes in the network generate corresponding CPTs, substitute the conditional probability tables CPT of all nodes into the directed acyclic graph, and obtain the target directed acyclic graph with the conditional probability table.

[0083] S140. According to the target directed acyclic graph and the prior knowledge of the target business, the value of each node in the directed acyclic graph under preset conditions is calculated by the RPLoc algorithm, and resources are created according to the value.

[0084] Exemplarily, a target host server with the lowest fragmentation rate is obtained according to the value, and resource creation is performed by calling a resource creation interface of a cloud management system and specifying a resource creation target instance.

[0085] Among them, the RPLog algorithm is an algorithm specific to resource allocation. Based on a given directed acyclic graph (DAG), it calculates the optimal value (with the minimum probability of fragmentation) of each node in the DAG under preset conditions, and then specifies the creation and allocation of resources. The operation logic of the RPLog algorithm includes: initialization: assigning an initial resource configuration to each node, which may be empty or a default configuration based on historical data; optimization calculation: traversing each node of the DAG; evaluating the resource fragmentation rate and resource utilization rate under different resource configuration schemes for each node; selecting a configuration scheme that can minimize the resource fragmentation rate and maximize resource utilization at the same time; updating the resource configuration of the node in the DAG; repeating the process until all nodes in the DAG are optimized; result output: inputting the final resource configuration scheme for each node, and the corresponding resource creation and allocation plan.

[0086] By acquiring the time series indicator data of the target business and performing aggregation and dimensionality reduction processing on the time series indicator data, an adjacent matrix is ​​obtained; the adjacent matrix is ​​substituted into the causal structure discovery algorithm to obtain a directed acyclic graph of the Bayesian network causal structure; based on the adjacent matrix and the directed acyclic graph, the conditional probability table of each node in the directed acyclic graph is calculated by the SimpleEstimator algorithm under preset conditions to obtain a target directed acyclic graph with the conditional probability table; according to the target directed acyclic graph and the prior knowledge of the target business, the value of each node in the directed acyclic graph under preset conditions is calculated by the RPLoc algorithm, and resources are created according to the values. The resource pool creates resources according to the predicted results (the value of each node in the directed acyclic graph under preset conditions is calculated by the RPLoc algorithm), reducing the generation of fragments in the resource pool and reducing the fragmentation rate of the entire resource pool.

[0087] In a possible implementation, the method further includes:

[0088] When the preset condition is that the resource fragmentation rate is the minimum, the resource fragmentation rate calculation logic is:

[0089] .

[0090] in, represents the resource fragmentation rate, k is the total number of server nodes, represents the fragmentation rate of the i-th server node, Represents the i-th virtual resource usage instance.

[0091] On the other hand, Figure 2 As shown, the present application provides a resource fragmentation management device, the device comprising:

[0092] The data collection module 201 is used to obtain the time series indicator data of the target business, and perform aggregation and dimensionality reduction processing on the time series indicator data to obtain an adjacent matrix;

[0093] A first calculation module 202 is used to substitute the adjacent matrix into a causal structure discovery algorithm to obtain a directed acyclic graph of a Bayesian network causal structure;

[0094] A second calculation module 203 is used to calculate, based on the adjacent matrix and the directed acyclic graph, a conditional probability table of each node in the directed acyclic graph under preset conditions by a SimpleEstimator algorithm to obtain a target directed acyclic graph with the conditional probability table;

[0095] The allocation module 204 is used to calculate the value of each node in the directed acyclic graph under preset conditions through the RPLoc algorithm according to the prior knowledge of the target directed acyclic graph and the target business, and create resources according to the value.

[0096] It is worth noting that the various modules included in the above-mentioned device are only divided according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be realized; in addition, the specific names of the various functional modules are only for the convenience of distinguishing each other, and are not used to limit the protection scope of the embodiments of the present invention.

[0097] In another embodiment of the present invention, an electronic device is also provided. Figure 3 A block diagram of an exemplary electronic device 50 suitable for implementing exemplary embodiments of the present invention is shown. Figure 3 The electronic device 50 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0098] like Figure 3 As shown, the electronic device 50 is in the form of a general-purpose computing device. The components of the electronic device 50 may include, but are not limited to: one or more processors or processing units 501, a system memory 502, and a bus 503 connecting different system components (including the system memory 502 and the processing unit 501).

[0099] Bus 503 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor or a local bus using any of a variety of bus structures. For example, these architectures include but are not limited to Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MAC) bus, Enhanced ISA bus, Video Electronics Standards Association (VESA) local bus and Peripheral Component Interconnect (PCI) bus.

[0100] The electronic device 50 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the electronic device 50, including volatile and non-volatile media, removable and non-removable media.

[0101] The system memory 502 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 504 and / or cache memory 505. The electronic device 50 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 506 may be used to read and write non-removable, non-volatile magnetic media ( Figure 3 not shown, usually called a "hard drive"). Although Figure 3Not shown in the figure, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, a DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to the bus 503 via one or more data medium interfaces. The memory 502 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of the various embodiments of the present invention.

[0102] A program / utility 508 having a set (at least one) of program modules 507 may be stored, for example, in the memory 502, such program modules 507 including but not limited to an operating system, one or more application programs, other program modules, and program data, each of which or some combination thereof may include an implementation of a network environment. The program modules 507 generally perform the functions and / or methods of the embodiments described herein.

[0103] The electronic device 50 may also communicate with one or more external devices 509 (e.g., keyboards, pointing devices, displays 510, etc.), may communicate with one or more devices that enable a user to interact with the electronic device 50, and / or may communicate with any device that enables the electronic device 50 to communicate with one or more other computing devices (e.g., a network card, a modem, etc.). Such communication may be performed via an input / output (I / O) interface 511. Furthermore, the electronic device 50 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 512. As shown, the network adapter 512 communicates with other modules of the electronic device 50 via the bus 503. It should be understood that although Figure 3 Not shown, other hardware and / or software modules may be used in conjunction with the electronic device 50, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0104] The processing unit 501 executes various functional applications and data processing by running the programs stored in the system memory 502, such as implementing the resource fragmentation management method provided in the embodiment of the present invention.

[0105] In another embodiment of the present invention, Figure 4 As shown, a storage medium 400 including a computer program 411 is also provided. When the computer program 411 is executed by a computer processor, it is used to perform a resource fragmentation management method. The method includes:

[0106] Acquire time series indicator data of the target business, and perform aggregation and dimensionality reduction processing on the time series indicator data to obtain an adjacent matrix;

[0107] Substituting the adjacency matrix into a causal structure discovery algorithm to obtain a directed acyclic graph of the Bayesian network causal structure;

[0108] Based on the adjacency matrix and the directed acyclic graph, a conditional probability table of each node in the directed acyclic graph is calculated by a SimpleEstimator algorithm under preset conditions to obtain a target directed acyclic graph with the conditional probability table;

[0109] According to the prior knowledge of the target directed acyclic graph and the target business, the value of each node in the directed acyclic graph under preset conditions is calculated by the RPLoc algorithm, and resources are created according to the value.

[0110] The computer storage medium of the embodiment of the present invention may adopt any combination of one or more computer-readable media. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination thereof. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, device or device.

[0111] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0112] Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0113] Computer program code for performing the operation of the embodiments of the present invention may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0114] The above disclosure is only the preferred embodiment of the present invention, which certainly cannot be used to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.

Claims

1. A resource fragmentation management method, It is characterized in that include: Acquire time series indicator data of the target business for generating an adjacent matrix, and perform aggregation and dimensionality reduction processing on the time series indicator data to obtain an adjacent matrix for generating a directed acyclic graph; The step of aggregating and reducing the dimension of the time series indicator data to obtain an adjacent matrix includes: aggregating the time series indicator data by a non-parametric Bayesian method to obtain a strongly correlated data set in the time series indicator data; constructing the adjacent matrix based on the strongly correlated data set, wherein the elements in the adjacent matrix represent the similarity data between the variable data in the strongly correlated data set; Substituting the adjacent matrix into a causal structure discovery algorithm to obtain a directed acyclic graph of a Bayesian network causal structure for estimating approximate values ​​of resource fragments under different configurations; The step of substituting the adjacent matrix into the causal structure discovery algorithm to obtain a directed acyclic graph of the Bayesian network causal structure for estimating the approximate value of resource fragments under different configurations also includes: Determine the functional relationship between each variable data based on the similarity data; Establishing the causal structure discovery algorithm according to the functional relationship; Substituting the adjacent matrix into a causal structure discovery algorithm to obtain a directed acyclic graph of a Bayesian network causal structure for estimating approximate values ​​of resource fragments under different configurations; Based on the adjacency matrix and the directed acyclic graph, a SimpleEstimator algorithm for estimating approximate values ​​of resource fragments under different configurations is used to calculate, under preset conditions, a conditional probability table of each node in the directed acyclic graph, to obtain a target directed acyclic graph with the conditional probability table; The step of calculating, based on the adjacency matrix and the directed acyclic graph, the conditional probability table of each node in the directed acyclic graph under preset conditions by the SimpleEstimator algorithm to obtain a target directed acyclic graph with the conditional probability table comprises: When the preset condition is that the resource fragmentation rate is minimum, according to the adjacent matrix and the directed acyclic graph, the conditional probability data of each node in the directed acyclic graph at different possible values ​​are calculated by the SimpleEstimator algorithm to obtain a conditional probability table of each node in the directed acyclic graph; Substituting the conditional probability table into the directed acyclic graph to obtain a target directed acyclic graph with the conditional probability table; According to the prior knowledge of the target directed acyclic graph and the target business, the value of each node in the directed acyclic graph under preset conditions is calculated by the RPLoc algorithm specific to resource allocation, and resources are created according to the values. Specifically, the RPLoc algorithm calculates the value of each node in the directed acyclic graph under the minimum resource fragmentation rate, and the process of creating resources according to the values ​​includes: initialization: assigning an initial resource configuration to each node in the target directed acyclic graph, which may be empty or a default configuration based on historical data; optimization calculation: traversing each node in each node in the target directed acyclic graph; evaluating the resource fragmentation rate and resource utilization rate under different resource configuration schemes for each node; selecting a configuration scheme that can simultaneously minimize the resource fragmentation rate and maximize the resource utilization rate; updating the resource configuration of each node in the target directed acyclic graph; repeating the process until all nodes in the target directed acyclic graph are optimized; result output: inputting the final resource configuration scheme for each node, and the corresponding resource creation and allocation plan.

2. The method according to claim 1, It is characterized in that Before the step of substituting the adjacent matrix into the causal structure discovery algorithm to obtain the directed acyclic graph of the Bayesian network causal structure, the method further includes: The adjacent matrix is ​​discretized by preset parameters to discretize the data in the adjacent matrix into different designated areas to obtain a discretized target adjacent matrix. The target adjacent matrix is ​​used to substitute into the causal structure discovery algorithm to obtain a directed acyclic graph of the Bayesian network causal structure, wherein different designated areas are used to represent different similarity levels.

3. The method according to claim 2, It is characterized in that The step of substituting the adjacent matrix into the causal structure discovery algorithm to obtain a directed acyclic graph of the Bayesian network causal structure includes: Substituting the target adjacent matrix into the conditional independence test algorithm in the causal structure discovery algorithm to obtain the conditional independent variables of each data in the target adjacent matrix; Determine the causal relationship connection direction of each conditional independent variable by using the V-structure algorithm and the Meek rules algorithm in the causal structure discovery algorithm; A directed acyclic graph of a Bayesian network causal structure is determined based on the causal relationship connection direction and the conditional independent variables.

4. The method according to claim 1, It is characterized in that The method further comprises: When the preset condition is that the resource fragmentation rate is the minimum, the resource fragmentation rate calculation logic is: in, represents the resource fragmentation rate, k is the total number of server nodes, represents the fragmentation rate of the i-th server node, Represents the i-th virtual resource usage instance.

5. A resource fragmentation management device, It is characterized in that The device comprises: A data acquisition module, used to obtain time series indicator data of a target business for generating an adjacent matrix, and to perform aggregation and dimensionality reduction processing on the time series indicator data to obtain an adjacent matrix for generating a directed acyclic graph; The step of aggregating and reducing the dimension of the time series indicator data to obtain an adjacent matrix includes: aggregating the time series indicator data by a non-parametric Bayesian method to obtain a strongly correlated data set in the time series indicator data; constructing the adjacent matrix based on the strongly correlated data set, wherein the elements in the adjacent matrix represent the similarity data between the variable data in the strongly correlated data set; A first calculation module is used to substitute the adjacent matrix into a causal structure discovery algorithm to obtain a directed acyclic graph of a Bayesian network causal structure for estimating approximate values ​​of resource fragments under different configurations; The step of substituting the adjacent matrix into the causal structure discovery algorithm to obtain a directed acyclic graph of the Bayesian network causal structure for estimating the approximate value of resource fragments under different configurations also includes: Determine the functional relationship between each variable data based on the similarity data; Establishing the causal structure discovery algorithm according to the functional relationship; Substituting the adjacent matrix into a causal structure discovery algorithm to obtain a directed acyclic graph of a Bayesian network causal structure for estimating approximate values ​​of resource fragments under different configurations; A second calculation module is used to calculate, based on the adjacent matrix and the directed acyclic graph, a conditional probability table of each node in the directed acyclic graph under preset conditions by using a SimpleEstimator algorithm for estimating approximate values ​​of resource fragments under different configurations, to obtain a target directed acyclic graph with the conditional probability table; The step of calculating, based on the adjacency matrix and the directed acyclic graph, the conditional probability table of each node in the directed acyclic graph under preset conditions by the SimpleEstimator algorithm to obtain a target directed acyclic graph with the conditional probability table comprises: When the preset condition is that the resource fragmentation rate is minimum, according to the adjacent matrix and the directed acyclic graph, the conditional probability data of each node in the directed acyclic graph at different possible values ​​are calculated by the SimpleEstimator algorithm to obtain a conditional probability table of each node in the directed acyclic graph; Substituting the conditional probability table into the directed acyclic graph to obtain a target directed acyclic graph with the conditional probability table; The allocation module is used to calculate the value of each node in the directed acyclic graph under preset conditions according to the prior knowledge of the target directed acyclic graph and the target business through the RPLoc algorithm specific to resource allocation, and create resources according to the values. Specifically, the RPLoc algorithm calculates the value of each node in the directed acyclic graph under the minimum resource fragmentation rate, and creates resources according to the values. The process includes: initialization: assigning an initial resource configuration to each node in the target directed acyclic graph, which may be empty or a default configuration based on historical data; optimization calculation: traversing each node in each node in the target directed acyclic graph; evaluating the resource fragmentation rate and resource utilization rate under different resource configuration schemes for each node; selecting a configuration scheme that can simultaneously minimize the resource fragmentation rate and maximize the resource utilization rate; updating the resource configuration of each node in the target directed acyclic graph; repeating the process until all nodes in the target directed acyclic graph are optimized; result output: inputting the final resource configuration scheme of each node, and the corresponding resource creation and allocation plan.

6. An electronic device, It is characterized in that The electronic device comprises: One or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the resource fragmentation management method as described in any one of claims 1-4.

7. A storage medium containing computer executable instructions, It is characterized in that The computer executable instructions are used to execute the resource fragmentation management method as described in any one of claims 1-4 when executed by a computer processor.