Tracking method and device of dynamic resource dependency relationship and terminal equipment

Abstract

The invention relates to a method and a device for tracking a dynamic resource dependency relationship and terminal equipment. The method comprises the following steps: acquiring cloud resource dependency data of a cloud service platform and / or a data center in real time; constructing a cloud resource dynamic dependency graph according to the cloud resource dependency data; detecting the cyclic dependence of the cloud resource dynamic dependence graph by adopting a depth-first search algorithm to obtain a cyclic dependence detection result; and using a TGCN algorithm to predict the future cyclic dependence of the cloud resource dynamic dependence graph to obtain a cyclic dependence prediction result. According to the method, a real-time cloud resource dynamic dependency graph is constructed, a depth-first search algorithm and a TGCN algorithm are adopted to process the cloud resource dynamic dependency graph to obtain a cyclic dependency detection result and a cyclic dependency prediction result, and a basis is provided for determining corresponding resource cyclic dependency prevention measures and processing strategies. And data is provided for effective cloud resource management and optimization, so that problems and risks caused by cyclic dependence are reduced, and the reliability of cloud services is improved.

Description

Technical Field

[0001] The present application relates to the field of information technology, and in particular to a method, apparatus and terminal device for tracking dynamic resource dependencies. Background Art

[0002] With the widespread application and rapid development of cloud services, cloud resource management and optimization have become important issues faced by cloud service providers and users. Among them, circular dependency is a common but serious problem that may lead to a closed loop between cloud resources, which in turn leads to resource competition, performance degradation, and reduced reliability. In the existing technology, there are some problems with the detection and prediction methods of circular dependencies. The probability of circular dependency of cloud resources accounts for about 2% of all dependencies. Once a dependency occurs, it will consume 3 times more resources than normal cloud resources.

[0003] Traditional circular dependency detection methods usually require offline data processing, that is, static analysis or model building before the cloud resource system is running. This method cannot provide real-time circular dependency detection results and may not capture dynamically changing dependencies in time.

[0004] For large-scale cloud resource systems, the computational complexity required for offline processing is also high, which will lead to extended detection time and resource waste. The detection and prediction of circular dependencies have poor support for dynamics. For example, the existing technical documents provide a system and method based on Helm cloud resource orchestration to design the underlying resources through the cloud resource orchestration visualization service module and the cloud resources and PaaS services provided by it, and save the designed orchestration files to the database; the technical solution of this document does not propose an effective solution to the circular dependency of cloud resources. For another example, the hybrid cloud resource allocation method and device provided in the existing technical documents is based on the current cloud resource usage rate corresponding to various cloud resource types and the standard usage rate range corresponding to each cloud resource type under the preset multiple cloud resource types. The cloud resources of the request end under various cloud resource types are adjusted. The technical solution of this document only solves the cloud resource allocation problem and the usage rate problem. The technical solutions of the above two documents are both through cloud resource methods, but neither of them solves the problem of circular dependency well. Circular dependencies continue to exist between cloud resources, which in turn causes resource competition, performance degradation and reliability reduction. Furthermore, there are a lot of changes and interactions in a dynamic environment, and the detection and prediction of circular dependencies need to consider the relationships between multiple variables at the same time. As the number and complexity of variables increase, the computational complexity of detection and prediction will also increase significantly, making it difficult to detect and predict circular dependencies in real time and accurately. And considering that in a dynamic environment, the real-time and integrity of data are often not guaranteed. Summary of the invention

[0005] The present application provides a method, apparatus and terminal device for tracking dynamic resource dependencies, which are used to solve the technical problem that existing cloud resource circular dependency detection and prediction methods are difficult to detect and predict the circular dependencies of dynamic cloud resources in real time.

[0006] In order to achieve the above objectives, this application provides the following technical solutions:

[0007] On the one hand, a method for tracking dynamic resource dependencies is provided, comprising the following steps:

[0008] Acquire cloud resource dependency data of a cloud service platform and / or data center in real time, wherein the cloud resource dependency data includes resource type information, service type information, change information, and dependency relationship information;

[0009] Constructing a cloud resource dynamic dependency graph according to the cloud resource dependency data;

[0010] Using a depth-first search algorithm to detect the circular dependency of the cloud resource dynamic dependency graph, and obtaining a circular dependency detection result;

[0011] The TGCN algorithm is used to predict the future cyclic dependency of the cloud resource dynamic dependency graph to obtain a cyclic dependency prediction result.

[0012] Preferably, constructing a cloud resource dynamic dependency graph according to the cloud resource dependency data includes:

[0013] Extracting data attributes from each piece of information of the cloud resource dependent data to obtain a data attribute set;

[0014] Performing support and normalization processing on the data attribute set to obtain a preprocessed data set;

[0015] Acquire the number of features, group the data of resource type information and service type information in all the preprocessed data sets according to the number of features, obtain the number of groups and the feature data group corresponding to each group; and determine the judgment threshold according to the number of groups;

[0016] Determine the data objects of the corresponding group according to the characteristic data group of each group, and filter the cloud resource dependency data according to the data objects of each group to obtain the number of occurrences of the data objects of each group;

[0017] Determine dependency information between the corresponding data objects and resources according to the number of occurrences of each group of the data objects and the judgment threshold;

[0018] The cloud resource dynamic dependency graph is constructed according to all the dependency information with resources as nodes and edges as dependency relationships.

[0019] Preferably, the data attribute set is subjected to support and normalization processing to obtain a preprocessed data set, which includes:

[0020] Obtain the frequency, data length and data quantity of each data from the data attribute set, and calculate the support of the corresponding data in the data attribute set according to the frequency and data quantity of each data;

[0021] Acquire the mapping space position, and calculate according to the mapping space position, each data and the data length and the support corresponding to the data to obtain the pre-processed data after the data is normalized;

[0022] All the preprocessed data constitute a preprocessed data set.

[0023] Preferably, a depth-first search algorithm is used to detect the circular dependency of the cloud resource dynamic dependency graph, and the circular dependency detection result obtained includes:

[0024] S31. Access the next node from any node of the cloud resource dynamic dependency graph to obtain node access data, wherein the node access data includes access status and access path;

[0025] S32. Determine whether there is a circular dependency between two nodes of the access path according to the access status, and obtain the access result;

[0026] S33. According to the access result, there is a circular dependency between two nodes of the access path, the access status value of the access node is greater than the status threshold, and the access status values ​​of all nodes of the cloud resource dynamic dependency graph are non-0, the cloud resource dynamic dependency graph is traversed and detected, and all the access results are used as circular dependency detection results;

[0027] Among them, if the access result is that there is no circular dependency between the two nodes of the access path, the access status value of the access node is set to 1, and the access to the next node is continued according to steps S31 and S32 until all nodes in the cloud resource dynamic dependency graph are visited; if the access result is that there is a circular dependency between the two nodes of the access path, the access status value of the access node is increased by 1, and the access to the next node is continued according to steps S31 and S32.

[0028] Preferably, the TGCN algorithm is used to predict the future cyclic dependency of the cloud resource dynamic dependency graph, and the cyclic dependency prediction result includes:

[0029] Acquire graph structure information, node feature information, and extract time series information from the cloud resource dynamic dependency graph;

[0030] Analyze the cloud resource dynamic dependency graph according to the time series information to obtain node past dependency information; extract the characteristics of each node in the cloud resource dynamic dependency graph according to the node feature information to obtain node feature information; analyze the connection and dependency relationship between two nodes in the cloud resource dynamic dependency graph according to the graph structure information to obtain graph structure characteristics;

[0031] According to the time series information, the node feature information and the graph structure information, a TGCN algorithm is used to construct a model to obtain a TGCN model;

[0032] The TGCN model is used to predict future circular dependencies based on the past dependency information of the nodes, the feature information of the nodes, and the graph structure features to obtain a circular dependency prediction result.

[0033] Preferably, the dynamic resource dependency tracking system includes dividing the time series information by time steps, and obtaining dependency evolution data within each time step as the node past dependency information.

[0034] In yet another aspect, a dynamic resource dependency tracking device is provided, the dynamic resource dependency tracking device comprising a data acquisition module, a dependency graph construction module, a detection module, and a prediction module;

[0035] The data acquisition module is used to acquire cloud resource dependency data of the cloud service platform and / or data center in real time, wherein the cloud resource dependency data includes resource type information, service type information, change information and dependency relationship information;

[0036] The dependency graph construction module is used to construct a cloud resource dynamic dependency graph according to the cloud resource dependency data;

[0037] The detection module is used to detect the circular dependency of the cloud resource dynamic dependency graph by using a depth-first search algorithm to obtain a circular dependency detection result;

[0038] The prediction module is used to predict the future cyclic dependency of the cloud resource dynamic dependency graph by using the TGCN algorithm to obtain a cyclic dependency prediction result.

[0039] Preferably, the dependency graph construction module includes an attribute extraction submodule, a data processing submodule, a data determination submodule, a data screening submodule, a dependency information determination submodule and a dependency graph construction submodule;

[0040] The attribute extraction submodule is used to extract data attributes from each information of the cloud resource dependency data to obtain a data attribute set;

[0041] The data processing submodule is used to perform support and normalization processing on the data attribute set to obtain a preprocessed data set;

[0042] The data determination submodule is used to obtain the number of features, group the data of resource type information and service type information in all the pre-processed data sets according to the number of features, obtain the number of groups and the feature data group corresponding to each group; and determine the judgment threshold according to the number of groups;

[0043] The data screening submodule is used to determine the data objects of the corresponding group according to the characteristic data group of each group, and screen the cloud resource dependency data according to the data objects of each group to obtain the number of occurrences of the data objects of each group;

[0044] The dependency information determination submodule is used to determine the dependency information between the corresponding data object and the resource according to the number of occurrences of each group of the data objects and the judgment threshold;

[0045] The dependency graph construction submodule is used to construct the cloud resource dynamic dependency graph according to all the dependency information, with resources as nodes and edges as dependency relationships.

[0046] Preferably, the data processing submodule is also used to obtain the frequency, data length and data quantity of each data from the data attribute set, and calculate based on the frequency and data quantity of each data to obtain the support of the corresponding data in the data attribute set; obtain the mapping space position, and calculate based on the mapping space position, each data and the data length and support corresponding to the data to obtain the preprocessed data after normalization of the data; all the preprocessed data constitute a preprocessed data set.

[0047] In another aspect, a terminal device is provided, comprising a processor and a memory;

[0048] The memory is used to store program codes and transmit the program codes to the processor;

[0049] The processor is used to execute the above-mentioned dynamic resource dependency tracking method according to the instructions in the program code.

[0050] The method, apparatus and terminal device for tracking dynamic resource dependencies include real-time acquisition of cloud resource dependency data of a cloud service platform and / or a data center, wherein the cloud resource dependency data includes resource type information, service type information, change information and dependency information; constructing a cloud resource dynamic dependency graph based on the cloud resource dependency data; using a depth-first search algorithm to detect circular dependencies in the cloud resource dynamic dependency graph to obtain circular dependency detection results; and using a TGCN algorithm to predict future circular dependencies in the cloud resource dynamic dependency graph to obtain circular dependency prediction results.

[0051] It can be seen from the above technical solutions that the present application has the following advantages: the method for tracking dynamic resource dependencies constructs a cloud resource dynamic dependency graph through cloud resource dependency data, and uses a depth-first search algorithm to detect circular dependencies in the cloud resource dynamic dependency graph, thereby obtaining a circular dependency detection result; the TGCN algorithm is used to predict the circular dependency prediction result based on the cloud resource dynamic dependency graph, and the circular dependency detection result and the circular dependency prediction result provide a basis for determining the corresponding resource circular dependency prevention measures and processing strategies, and also provide data for effective cloud resource management and optimization to reduce the problems and risks caused by circular dependencies, thereby improving the reliability of cloud services, and solving the technical problem that the existing cloud resource circular dependency detection and prediction methods are difficult to detect and predict the circular dependencies of dynamic cloud resources in real time.

[0052] The dynamic resource dependency tracking device realizes dynamic capture of cloud resource dependency data and real-time update of resource dynamic dependency graph through the cooperation of data acquisition module, dependency graph construction module, detection module and prediction module; it uses depth-first search algorithm to detect circular dependency, prevent deadlock and resource waste, and provide data for providing effective processing strategies; it also avoids circular dependency between resources to ensure the rationality of resource allocation and use. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present application 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 application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0054] Figure 1 A flowchart of the steps of the method for tracking dynamic resource dependencies described in an embodiment of the present application;

[0055] Figure 2 A flowchart of the steps of detecting circular dependencies in the method for tracking dynamic resource dependencies described in an embodiment of the present application;

[0056] Figure 3 A schematic diagram of a framework of a device for tracking dynamic resource dependencies described in an embodiment of the present application;

[0057] Figure 4 A schematic diagram of a terminal device described in an embodiment of the present application. DETAILED DESCRIPTION

[0058] In order to make the purpose, features, and advantages of the invention of this application more obvious and easy to understand, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the embodiments described below are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0059] In the description of the embodiments of the present application, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present application, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0060] In the embodiments of the present application, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral one; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the embodiments of the present application can be understood according to the specific circumstances.

[0061] Patent terms:

[0062] Cloud resources are a service model based on cloud computing, whose main features are virtualization, dynamic allocation and elastic expansion and contraction of resources. The standard definition of cloud resources refers to the clear and standardized definition of various aspects of cloud resources in the field of cloud computing, including the types, functions, performance, security and other aspects of cloud resources.

[0063] API interface (Application Program Interface): API interface can realize mutual communication between computer software. Developers can develop applications through API interface program, which can reduce the writing of useless programs and alleviate programming tasks. API is also a kind of middleware that provides data sharing for various platforms.

[0064] AWS (Amazon web service): is a cloud computing service platform under Amazon, providing a complete set of infrastructure and cloud solutions for customers in various countries and regions around the world.

[0065] EC2 API (EC2 (elastic compute cloud) virtual server in the cloud): is an Amazon Web service used to create and run virtual machines in the cloud. EC2 is a virtual computer with unlimited collection capacity that users can use to perform some processing tasks. EC2 is an optional virtual cluster service model.

[0066] Azure Machine Image (AMI): Contains the packages required by a server (including the operating system and other software).

[0067] ResourceManager (RM for short) is Linkis' computing resource management module.

[0068] The TGCN algorithm refers to a deep learning model suitable for time series graph data, and the neural network of the deep learning model is the existing time graph convolutional neural network.

[0069] VPC (Virtual Private Cloud) is a virtual private cloud service provided by Alibaba Cloud. It is a dynamically configured pool in which users can build a completely isolated virtual private cloud environment, providing higher security and flexibility.

[0070] ECS (Elastic Compute Service) is Alibaba Cloud's basic cloud computing service, providing cloud server instances of various specifications and configurations for users to choose from. Users can choose different instance specifications such as CPU, memory, storage, etc. according to their needs, and enjoy rich network and security features.

[0071] OSS (Object Storage Service) is Alibaba Cloud's distributed storage service for storing and managing massive amounts of data. It provides a highly reliable and scalable storage solution that allows users to easily upload, download, and manage files. OSS is particularly suitable for storing static files, such as large files such as pictures and videos, and can improve access speed and success rate through the content distribution network (CDN).

[0072] ‌Depth-first search algorithm (DFS) is an algorithm for traversing a tree or graph. Its core idea is to explore the branches of the tree as deeply as possible by backtracking and pruning until the target vertex is reached or all nodes are traversed. It can be understood that starting from the starting vertex, a path is selected to deeply explore the target vertex. When the target vertex does not exist on the path, the search will backtrack to the starting vertex and continue to select another path for depth exploration until the target vertex is found or all paths are exhausted. Since DFS traverses vertices in a last-in-first-out manner, a stack can be used to store vertices that have been searched.

[0073] The embodiments of the present application provide a method, apparatus and terminal device for tracking dynamic resource dependencies, which solve the technical problem that existing cloud resource circular dependency detection and prediction methods are difficult to detect and predict the circular dependencies of dynamic cloud resources in real time.

[0074] Embodiment 1:

[0075] Figure 1 This is a flowchart of the steps of the method for tracking dynamic resource dependencies described in an embodiment of the present application.

[0076] like Figure 1 As shown, the embodiment of the present application provides a method for tracking dynamic resource dependencies, comprising the following steps:

[0077] S1. Obtain cloud resource dependency data of the cloud service platform and / or data center in real time. The cloud resource dependency data includes resource type information, service type information, change information and dependency relationship information.

[0078] It should be noted that in step S1, cloud resource dependency data of the cloud service platform and / or data center is obtained to provide data for subsequent steps. In this embodiment, resource type information includes servers, storage devices, network devices, etc. Service type information includes ‌VPC (Virtual Private Cloud), ECS (Elastic Compute Service) and OSS (Object Storage Service), etc. Change information includes change history records, monitoring and log information such as creation, modification, and deletion of resources. Dependency information includes dependencies between applications, dependencies between services, scheduling relationships between applications, etc. In other embodiments, real-time acquisition of cloud resource dependency data of the cloud service platform and / or data center can automatically discover new cloud resources and their dependencies through the monitoring system, and collect cloud resource dependency data through the monitoring system. The data includes device status, performance indicators, configuration information, etc.

[0079] In an embodiment of the present application, real-time acquisition of cloud resource dependency data of a cloud service platform and / or a data center can be performed through the API interface of the service platform and / or the data center to obtain a unique identifier ID for all cloud resources; based on the unique identifier ID, the resource type information, service type information, change information, dependency information and other data of each resource can be obtained through the API interface.

[0080] S2. Build a cloud resource dynamic dependency graph based on cloud resource dependency data.

[0081] It should be noted that in step S2, a cloud resource dynamic dependency graph is constructed based on the cloud resource dependency data obtained in step S1, so as to provide a cloud resource dynamic dependency graph for subsequent use.

[0082] S3. Use the depth-first search algorithm to detect the circular dependency of the cloud resource dynamic dependency graph and obtain the circular dependency detection result.

[0083] It should be noted that in step S3, a depth-first search algorithm is used to detect the circular dependency in the cloud resource dynamic dependency graph constructed in step S2 to obtain a circular dependency detection result.

[0084] In the embodiment of the present application, in order to solve the problems of network delay and distributed computing, the dynamic resource dependency tracking method uses an optimized depth-first search algorithm to detect the circular dependency of the cloud resource dynamic dependency graph. The content of optimizing the depth-first search algorithm includes:

[0085] Obtain an estimate of the network delay and the delay of each path in a depth-first search algorithm, and prioritize paths with delays below a preset delay threshold;

[0086] By obtaining the location information of the data in the distributed system, the computing node with the best data locality is determined, and the computing node is used as the priority for search;

[0087] By obtaining real-time network status and system load, the load between two computing nodes is dynamically adjusted according to the network status and system load, and the search task allocation is determined according to the processing capacity and current load of each computing node;

[0088] Based on the preferred paths and computing nodes, the search tasks are assigned to several computing nodes for simultaneous search through parallel computing to determine the best search strategy and parameters;

[0089] The search strategy and parameters of the depth-first search algorithm are replaced according to the best search strategy and parameters to obtain an optimized depth-first search algorithm.

[0090] It should be noted that the optimized depth-first search algorithm uses asynchronous communication to perform concurrent network transmission and data processing during the search process. The optimized depth-first search algorithm uses the best search strategy and parameters to improve search efficiency and accuracy.

[0091] S4. The TGCN algorithm is used to predict the future cyclic dependencies of the cloud resource dynamic dependency graph and obtain the cyclic dependency prediction results.

[0092] It should be noted that in step S4, the cloud resource dynamic dependency graph is constructed according to step S2. The real-time dynamic change information is obtained according to step S1, and a prediction method is used to predict the possible circular dependencies in the future to obtain the circular dependency prediction result.

[0093] In an embodiment of the present application, the method for tracking dynamic resource dependencies further includes:

[0094] Determine whether to make a prediction based on the change information of the real-time cloud resource dependency data and obtain a judgment result;

[0095] If the judgment result is update information, the cloud resource dynamic dependency graph is updated according to the update information to obtain a first updated dynamic dependency graph; a depth-first search algorithm is used to perform a circular dependency detection on the first updated dynamic dependency graph to obtain an updated circular dependency detection result;

[0096] If the judgment result is prediction information, the cloud resource dynamic dependency graph is updated according to the prediction information to obtain a second updated dynamic dependency graph; the access state value corresponding to the prediction information in the second updated dynamic dependency graph is set to 0, and a depth-first search algorithm is used to perform circular dependency detection on the second updated dynamic dependency graph to obtain an updated circular dependency detection result.

[0097] If the updated circular dependency detection result is that there is no circular dependency relationship, the circular dependency prediction result is used as reference data for preventing and handling circular dependencies of the cloud service platform and / or data center.

[0098] It should be noted that the dynamic resource dependency tracking method obtains the change information of cloud resource dependency data (such as periodic or change information) in real time through the monitoring system, and updates the cloud resource status and dependency relationship in the cloud resource dynamic dependency graph. If the comparison result of the real-time change information obtained with the previous change information is a newly added node, the judgment result is the prediction information; if the comparison result of the real-time change information obtained with the previous change information is that the dependency relationship of the existing node is updated, the judgment result is the update information. After that, according to the judgment result of the prediction information, the access status value of the current node is set to 0, and the other nodes maintain the results after S3 traversal unchanged; continue to execute the detection program of the depth-first search algorithm, and determine whether there is a circular dependency. This execution method uses the detection program of the depth-first search algorithm of S3. If the judgment result is update information, the access status value of the entire node is set to 0, and the detection program of the depth-first search algorithm of S3 is continued. After the real-time dynamic change information is updated, the circular dependency judgment is made. If it is not updated but predicted information, the cloud resources and cloud resource dependencies need to be updated. Before the change, the unique identifier ID of the cloud resource of the change result, the type of each resource, service type information, change information and dependency are obtained first according to the result to be changed. Finally, according to the updated circular dependency detection result, if there is a circular dependency, it is returned to the monitoring system, and the monitoring system is prohibited from updating the change result. The cloud resource dynamic dependency graph and the depth-first search algorithm detection program execution results are rolled back; according to the updated circular dependency detection result, if there is no circular dependency, the result returned to the monitoring system is feasible, the monitoring system completes the change, and the cloud resource dynamic dependency graph and the depth-first search algorithm detection program execution results are retained. The dynamic resource dependency tracking method completes the circular dependency detection in the dynamic situation, and allows the monitoring system to obtain the predicted result to change the cloud resources, avoiding the undesirable situation of dealing with and solving the problem after it occurs.

[0099] In an embodiment of the present application, the method for tracking dynamic resource dependencies also includes: determining corresponding resource circular dependency prevention measures and processing strategies based on circular dependency detection results and circular dependency prediction results. Obtaining identifiers of nodes with circular dependencies based on the cloud resource dynamic dependency graph, the identifiers are fed back to a monitoring system connected to a service platform and / or data center; the monitoring system performs resource management and optimization based on the properties and dependencies of the nodes.

[0100] It should be noted that the corresponding resource circular dependency prevention measures and processing strategies are determined according to the circular dependency detection results and circular dependency prediction results, including: providing a basis for resource allocation strategy based on the dependency relationship between resources and the circular dependency prediction results, avoiding placing interdependent resources in the same circular dependency chain; through constructor injection, the acquisition and use rights of resources are handed over to external management to avoid direct circular dependencies between resources; the constructor is used to initialize the instance of the class when creating an object; for resources with circular dependencies, middleware (such as program interfaces) is introduced, and the middleware serves as a decoupling layer between resources to cut off circular dependencies and use resources; for potential circular dependency problems in resource initialization and loading, a delayed loading strategy is adopted, that is, loading resources when they are needed to avoid circular dependencies of resources; regular resource circular dependency detection is performed to ensure the rationality of resource allocation and use. Among them, the delayed loading strategy is used to avoid circular dependencies of resources, and the content of the delayed loading strategy includes: according to the delayed loading strategy, determine the resources that need to be loaded and mark them as unloaded. Through the cache mechanism, it is determined whether the resources that need to be loaded already exist in the cache. If the resources already exist in the cache, the resources are directly obtained from the cache. If the resource does not exist in the cache, determine whether the resource has a circular dependency. By parsing the circular dependency detection result or the circular dependency prediction result dependency, determine the loading order of the resources to avoid the occurrence of circular dependency problems. According to the loading order, use asynchronous loading to load the resources. During the loading process, use delayed initialization to delay the actual initialization of the resource until the resource is accessed for the first time. Store the loaded resources in the cache and use weak references to reference the resources.

[0101] The present application provides a method for tracking dynamic resource dependencies, including real-time acquisition of cloud resource dependency data of a cloud service platform and / or a data center, the cloud resource dependency data including resource type information, service type information, change information and dependency information; constructing a cloud resource dynamic dependency graph according to the cloud resource dependency data; using a depth-first search algorithm to detect the cyclic dependency of the cloud resource dynamic dependency graph to obtain a cyclic dependency detection result; using a TGCN algorithm to predict the future cyclic dependency of the cloud resource dynamic dependency graph to obtain a cyclic dependency prediction result. The method for tracking dynamic resource dependencies constructs a cloud resource dynamic dependency graph through cloud resource dependency data, and uses a depth-first search algorithm to detect the cyclic dependency in the cloud resource dynamic dependency graph, thereby obtaining a cyclic dependency detection result; using a TGCN algorithm to predict the cyclic dependency prediction result according to the cloud resource dynamic dependency graph, the cyclic dependency detection result and the cyclic dependency prediction result provide a basis for determining the corresponding resource cyclic dependency prevention measures and processing strategies, and also provide data for effective cloud resource management and optimization to reduce the problems and risks caused by cyclic dependencies, thereby improving the reliability of cloud services; it solves the technical problem that the existing cloud resource cyclic dependency detection and prediction methods are difficult to detect and predict the cyclic dependencies of dynamic cloud resources in real time.

[0102] It should be noted that the dynamic resource dependency tracking method updates the resource dynamic dependency graph in real time by acquiring cloud resource dependency data that dynamically captures cloud resources; a depth-first search algorithm is used to detect circular dependencies to prevent deadlocks and resource waste, providing data for providing effective processing strategies; and circular dependencies between resources are also avoided to ensure the rationality of resource allocation and use.

[0103] In one embodiment of the present application, constructing a cloud resource dynamic dependency graph according to cloud resource dependency data includes:

[0104] Extract data attributes for each piece of information on cloud resource dependent data to obtain a data attribute set;

[0105] Perform support and normalization processing on the data attribute set to obtain the preprocessed data set;

[0106] Obtain the number of features, group the data of resource type information and service type information in all preprocessed data sets according to the number of features, obtain the number of groups and the feature data group corresponding to each group; and determine the judgment threshold according to the number of groups;

[0107] Determine the data objects of the corresponding group according to the characteristic data group of each group, and filter the data objects of each group according to the cloud resource dependency data to obtain the number of occurrences of the data objects of each group;

[0108] Determine the dependency information between the corresponding data objects and resources according to the number of occurrences of each group of data objects and the judgment threshold;

[0109] Build a cloud resource dynamic dependency graph based on all dependency information, with resources as nodes and edges as dependency relationships;

[0110] Among them, the data attribute set is supported and normalized, and the preprocessed data set includes:

[0111] Obtain the frequency, data length and data quantity of each data from the data attribute set, and calculate the support of the corresponding data in the data attribute set based on the frequency and data quantity of each data;

[0112] Obtain the mapping space position, and calculate the pre-processed data after normalization of the data according to the mapping space position, each data and the data length and support corresponding to the data;

[0113] All preprocessed data form a preprocessed dataset.

[0114] It should be noted that the constructed cloud resource dynamic dependency graph is established with resources as nodes and edges as dependency relationships. In this embodiment, the cloud resource dependency data is denoted as I, , where n is the number of information contained in the cloud resource dependency data. Data attributes are extracted for each information in the cloud resource dependency data to obtain the data attribute set J, Data attribute changes: resource creation, modification, deletion and other change history records, monitoring and log information. Then calculate the support S of each piece of information in the cloud resource dependent data in the data attribute set, S=P / n, P is the frequency of the data in the data attribute set. Then, according to the mapping space position, each data and the data length and support calculation corresponding to the data, the pre-processed data Z after normalization of the data is obtained. , where ω is the mapping space position, Z i is the preprocessed data of the i-th data in the data attribute set, J i is the i-th data in the data attribute set, L i is the data length of the i-th data in the data attribute set, S i is the support of the i-th data in the data attribute set. After normalization, the data in the data attribute set is mapped to the range of 0~1. After obtaining the preprocessed data set, the number of features d is obtained, and the data in the preprocessed data set is divided into M feature data groups according to the number of features. The M feature data groups constitute the data feature space, and the judgment threshold m is determined according to the number of groups M. Among them, , In the data feature space, the center point of the data feature space is selected as the determined data feature point, and the data closest to the data feature point is screened out from each data feature group to extract the data object of the data. Then, the number of occurrences corresponding to the data object is extracted from the cloud resource dependency data I. If the number of occurrences exceeds the judgment threshold m, the dependency information is that there is a dependency relationship between the data object and the resource; if the number of occurrences does not exceed the judgment threshold m, the dependency information is that there is no dependency relationship between the data object and the resource.

[0115] In an embodiment of the present application, during the process of constructing a dynamic dependency graph of cloud resources, the service type information and dependency relationships in the dynamic dependency graph of cloud resources can be updated based on the change information obtained in real time; when the service type information or dependency relationship changes, the corresponding event is triggered, the dynamic dependency graph of cloud resources is updated, and the relevant monitoring system or user is notified. The monitoring system automatically discovers new cloud resources and their dependencies, adds the new cloud resources and their dependencies to the dynamic dependency graph of cloud resources, and updates the data such as the cloud resource status and dependency relationships in the dynamic dependency graph of cloud resources. These data include automatic deployment of new cloud resources, automatic configuration of network connections, automatic configuration of dependencies between services, etc. The monitoring system also collects data on cloud resources and dependency relationships. The data on cloud resources includes device status, performance indicators, configuration information, etc. Data with dependencies include call relationships between applications and dependencies between services, etc.

[0116] In the embodiment of the present application, new cloud resources and their dependencies are added to the cloud resource dynamic dependency graph, and the resource type information, service type status and dependency relationships in the cloud resource dynamic dependency graph are updated. For example, each cloud resource A is composed of a service performance modules, and each server performance module has B k performance indicators, therefore, the ideal space of cloud resource service performance module requirements is expressed as C k , then the change information of each service performance module of cloud resources (such as the optimal demand value represented by D k ), where k = 1, 2, …, a. Calculate the optimal demand value D k The content includes:

[0117] Obtain the parameter information coordinates of each service performance module of manufacturing resources through the object-element description model , where k = 1, 2, ..., a, j = 1, 2, ..., B k , is the value of the jth service performance indicator in the kth service performance module of the cloud resource;

[0118] According to the object-element description model of user needs, the optimal service performance parameter coordinates of cloud resource requirements are obtained ,Therefore, the ideal interval coordinates of each service performance module of the cloud resource demand are obtained.,This matter-element description model can quantitatively evaluate the parameter information of different ,service performance modules and express it in coordinate form.

[0119] Figure 2 This is a flow chart of the steps of circular dependency detection in the dynamic resource dependency tracking method described in an embodiment of the present application.

[0120] like Figure 2 As shown, in one embodiment of the present application, a depth-first search algorithm is used to detect the circular dependency of the cloud resource dynamic dependency graph, and the circular dependency detection result obtained includes:

[0121] S31. Access the next node from any node in the cloud resource dynamic dependency graph to obtain node access data, which includes access status and access path;

[0122] S32. Determine whether there is a circular dependency between two nodes of the access path according to the access status, and obtain the access result;

[0123] S33. According to the access result, there is a circular dependency between two nodes of the access path, the access status value of the access node is greater than the status threshold, and the access status values ​​of all nodes in the cloud resource dynamic dependency graph are non-0, the cloud resource dynamic dependency graph is traversed and detected, and all access results are used as circular dependency detection results;

[0124] Among them, if the access result is that there is no circular dependency between the two nodes of the access path, the access status value of the access node is set to 1, and the access to the next node is continued according to steps S31 and S32 until all nodes in the cloud resource dynamic dependency graph are visited; if the access result is that there is a circular dependency between the two nodes of the access path, the access status value of the access node is increased by 1, and the access to the next node is continued according to steps S31 and S32.

[0125] It should be noted that when using the depth-first search algorithm to traverse the circular dependency of the cloud resource dynamic dependency graph, the depth-first search algorithm is used to start traversing from each node of the cloud resource dynamic dependency graph; during the traversal process, each access path is maintained and the nodes that have been visited are recorded. When a node is accessed, it is determined whether the node is already in the current access path; if so, there is a circular dependency; if not, the node is added to the access path and its subsequent nodes are continued to be traversed. If a circular dependency is detected, a circular dependency exception is thrown and the processing is interrupted; if there is no circular dependency, the cloud resource dynamic dependency graph continues to be traversed until all nodes of the cloud resource dynamic dependency graph are visited. In this embodiment, the depth-first search algorithm detects the circular dependency in the cloud resource dynamic dependency graph by traversing in the following steps: first, the access status values ​​of all nodes in the cloud resource dynamic dependency graph are set to 0, and then the access is started from any node, and the access status value of the current node is determined to be 0 according to the node access data. After the access, the access status value is set to +1, and then the next node is accessed according to the dependency relationship; secondly, during the traversal process, if the access status value of the next accessed node is not 0, there is a circular dependency relationship, and an alarm is issued to indicate the circular dependency relationship and which node it is. At this time, the access status value of the node is increased by 1; the traversal continues; finally, the traversal continues until there is an access status value greater than 5 in the access node, then check whether the access status values ​​of all nodes in the cloud resource dynamic dependency graph are non-0, if they are all non-0, the traversal of the current cloud resource dynamic dependency graph is completed; if there is a 0 situation, jump to the node with an access status value of 0 and continue to traverse until the access status values ​​of all nodes are non-0.

[0126] In one embodiment of the present application, the TGCN algorithm is used to predict the future cyclic dependency of the cloud resource dynamic dependency graph, and the cyclic dependency prediction results include:

[0127] Obtain graph structure information, node feature information, and extract time series information from the cloud resource dynamic dependency graph;

[0128] The cloud resource dynamic dependency graph is analyzed according to the time series information to obtain the node past dependency information; the characteristics of each node in the cloud resource dynamic dependency graph are extracted according to the node feature information to obtain the node feature information; the connection and dependency relationship between two nodes in the cloud resource dynamic dependency graph is analyzed according to the graph structure information to obtain the graph structure characteristics;

[0129] The TGCN algorithm is used to construct the model based on the time series information, node feature information and graph structure information to obtain the TGCN model;

[0130] The TGCN model is used to predict the future occurrence of circular dependencies based on the node past dependency information, node feature information, and graph structure features, and the circular dependency prediction results are obtained;

[0131] Among them, the time series information is divided according to the time step, and the dependency evolution data within each time step is obtained as the past dependency information of the node.

[0132] It should be noted that the evolution of past dependencies and node features can be obtained based on time series information. Node feature information includes node type and resource usage.

[0133] Embodiment 2:

[0134] Figure 3 A schematic diagram of the framework of a device for tracking dynamic resource dependencies described in an embodiment of the present application.

[0135] like Figure 3 As shown, the embodiment of the present application provides a dynamic resource dependency tracking device, including a data acquisition module 100, a dependency graph construction module 200, a detection module 300 and a prediction module 400;

[0136] The data acquisition module 100 is used to acquire cloud resource dependency data of the cloud service platform and / or data center in real time, where the cloud resource dependency data includes resource type information, service type information, change information, and dependency relationship information;

[0137] A dependency graph construction module 200 is used to construct a cloud resource dynamic dependency graph according to cloud resource dependency data;

[0138] The detection module 300 is used to detect the circular dependency of the cloud resource dynamic dependency graph by using a depth-first search algorithm to obtain a circular dependency detection result;

[0139] The prediction module 400 is used to predict the future cyclic dependency of the cloud resource dynamic dependency graph by using the TGCN algorithm to obtain a cyclic dependency prediction result.

[0140] It should be noted that the content of the module in the dynamic resource dependency tracking device corresponds to the step content of the dynamic resource dependency tracking method. The step content of the dynamic resource dependency tracking method has been described in Example 1, and the content of the module in the dynamic resource dependency tracking device will not be repeated in this embodiment. The dynamic resource dependency tracking device dynamically captures cloud resource dependency data of cloud resources and updates the resource dynamic dependency graph in real time through the cooperation of the data acquisition module, dependency graph construction module, detection module and prediction module; uses a depth-first search algorithm to detect circular dependencies, prevent deadlocks and resource waste, and provide data for providing effective processing strategies; it also avoids circular dependencies between resources to ensure the rationality of resource allocation and use.

[0141] In one embodiment of the present application, the dependency graph construction module 200 includes an attribute extraction submodule, a data processing submodule, a data determination submodule, a data screening submodule, a dependency information determination submodule, and a dependency graph construction submodule;

[0142] The attribute extraction submodule is used to extract data attributes from each information of cloud resource dependent data to obtain a data attribute set;

[0143] The data processing submodule is used to perform support and normalization processing on the data attribute set to obtain a preprocessed data set;

[0144] The data determination submodule is used to obtain the number of features, group the data of resource type information and service type information in all pre-processed data sets according to the number of features, obtain the number of groups and the feature data group corresponding to each group; and determine the judgment threshold according to the number of groups;

[0145] The data screening submodule is used to determine the data objects of the corresponding group according to the characteristic data group of each group, and screen the data objects of each group according to the cloud resource dependency data to obtain the number of occurrences of the data objects of each group;

[0146] A dependency information determination submodule, used to determine dependency information between corresponding data objects and resources according to the number of occurrences of each group of data objects and a judgment threshold;

[0147] The dependency graph construction submodule is used to construct a dynamic dependency graph of cloud resources based on all dependency information, with resources as nodes and edges as dependency relationships.

[0148] In an embodiment of the present application, the data processing submodule is also used to obtain the frequency, data length and data quantity of each data from the data attribute set, and calculate the support of the corresponding data in the data attribute set based on the frequency and data quantity of each data; obtain the mapping space position, and calculate the preprocessed data after normalization of the data based on the mapping space position, each data and the data length and support corresponding to the data; all preprocessed data constitute a preprocessed data set.

[0149] Embodiment three:

[0150] Figure 4 A schematic diagram of a terminal device described in an embodiment of the present application.

[0151] like Figure 4 As shown, an embodiment of the present application provides a terminal device, including a processor and a memory;

[0152] A memory, used for storing program codes and transmitting the program codes to a processor;

[0153] The processor is used to execute the above-mentioned dynamic resource dependency tracking method according to the instructions in the program code.

[0154] It should be noted that the processor is used to execute the steps in the above-mentioned embodiment of a method for tracking dynamic resource dependencies according to the instructions in the program code. Alternatively, the processor implements the functions of each module / unit in the above-mentioned system / device embodiments when executing the computer program.

[0155] Exemplarily, the computer program may be divided into one or more modules / units, one or more modules / units are stored in a memory and executed by a processor to complete the present application. One or more modules / units may be a series of computer program instruction segments capable of completing a specific function, and the instruction segments are used to describe the execution process of the computer program in a terminal device.

[0156] The terminal device may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The terminal device may include, but is not limited to, a processor and a memory. Those skilled in the art will appreciate that this does not constitute a limitation on the terminal device, and may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the terminal device may also include input and output devices, network access devices, buses, etc.

[0157] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (dSIC), field-programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0158] The memory may be an internal storage unit of the terminal device, such as a hard disk or memory of the terminal device. The memory may also be an external storage device of the terminal device, such as a plug-in hard disk, a smart memory card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the terminal device. Furthermore, the memory may include both an internal storage unit of the terminal device and an external storage device. The memory is used to store computer programs and other programs and data required by the terminal device. The memory may also be used to temporarily store data that has been output or is to be output.

[0159] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0160] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0161] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0162] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0163] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

[0164] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for tracking dynamic resource dependencies, characterized in that: The following steps are involved: Acquire cloud resource dependency data of a cloud service platform and / or data center in real time, wherein the cloud resource dependency data includes resource type information, service type information, change information, and dependency relationship information; Constructing a cloud resource dynamic dependency graph according to the cloud resource dependency data; Using a depth-first search algorithm to detect the circular dependency of the cloud resource dynamic dependency graph, and obtaining a circular dependency detection result; The TGCN algorithm is used to predict the future cyclic dependency of the cloud resource dynamic dependency graph to obtain a cyclic dependency prediction result.

2. The method for tracking dynamic resource dependencies according to claim 1, characterized in that: Constructing a cloud resource dynamic dependency graph according to the cloud resource dependency data includes: Extracting data attributes from each piece of information of the cloud resource dependent data to obtain a data attribute set; Performing support and normalization processing on the data attribute set to obtain a preprocessed data set; Acquire the number of features, group the data of resource type information and service type information in all the preprocessed data sets according to the number of features, obtain the number of groups and the feature data group corresponding to each group; and determine the judgment threshold according to the number of groups; Determine the data objects of the corresponding group according to the characteristic data group of each group, and filter the cloud resource dependency data according to the data objects of each group to obtain the number of occurrences of the data objects of each group; Determine dependency information between the corresponding data objects and resources according to the number of occurrences of each group of the data objects and the judgment threshold; The cloud resource dynamic dependency graph is constructed according to all the dependency information with resources as nodes and edges as dependency relationships.

3. The method for tracking dynamic resource dependencies according to claim 2, characterized in that: The data attribute set is subjected to support and normalization processing to obtain a preprocessed data set including: Obtain the frequency, data length and data quantity of each data from the data attribute set, and calculate the support of the corresponding data in the data attribute set according to the frequency and data quantity of each data; Acquire the mapping space position, and calculate according to the mapping space position, each data and the data length and the support corresponding to the data to obtain the pre-processed data after the data is normalized; All the preprocessed data constitute a preprocessed data set.

4. The method for tracking dynamic resource dependencies according to claim 1, characterized in that: The depth-first search algorithm is used to detect the circular dependency of the cloud resource dynamic dependency graph, and the circular dependency detection results obtained include: S31. Access the next node from any node of the cloud resource dynamic dependency graph to obtain node access data, wherein the node access data includes access status and access path; S32. Determine whether there is a circular dependency between two nodes of the access path according to the access status, and obtain the access result; S33. According to the access result, there is a circular dependency between two nodes of the access path, the access status value of the access node is greater than the status threshold, and the access status values ​​of all nodes of the cloud resource dynamic dependency graph are non-0, the cloud resource dynamic dependency graph is traversed and detected, and all the access results are used as circular dependency detection results; Among them, if the access result is that there is no circular dependency between the two nodes of the access path, the access status value of the access node is set to 1, and the access to the next node is continued according to steps S31 and S32 until all nodes in the cloud resource dynamic dependency graph are visited; if the access result is that there is a circular dependency between the two nodes of the access path, the access status value of the access node is increased by 1, and the access to the next node is continued according to steps S31 and S32.

5. The method for tracking dynamic resource dependencies according to claim 4, characterized in that: The TGCN algorithm is used to predict the future cyclic dependency of the cloud resource dynamic dependency graph, and the cyclic dependency prediction results include: Acquire graph structure information, node feature information, and extract time series information from the cloud resource dynamic dependency graph; Analyze the cloud resource dynamic dependency graph according to the time series information to obtain node past dependency information; extract the characteristics of each node in the cloud resource dynamic dependency graph according to the node feature information to obtain node feature information; analyze the connection and dependency relationship between two nodes in the cloud resource dynamic dependency graph according to the graph structure information to obtain graph structure characteristics; According to the time series information, the node feature information and the graph structure information, a TGCN algorithm is used to construct a model to obtain a TGCN model; The TGCN model is used to predict future circular dependencies based on the past dependency information of the nodes, the feature information of the nodes, and the graph structure features to obtain a circular dependency prediction result.

6. The method for tracking dynamic resource dependencies according to claim 5, characterized in that: include: The time series information is divided according to time steps, and dependency evolution data within each time step is obtained as past dependency information of the node.

7. A dynamic resource dependency tracking device, characterized in that: It includes data acquisition module, dependency graph construction module, detection module and prediction module; The data acquisition module is used to acquire cloud resource dependency data of the cloud service platform and / or data center in real time, wherein the cloud resource dependency data includes resource type information, service type information, change information and dependency relationship information; The dependency graph construction module is used to construct a cloud resource dynamic dependency graph according to the cloud resource dependency data; The detection module is used to detect the circular dependency of the cloud resource dynamic dependency graph by using a depth-first search algorithm to obtain a circular dependency detection result; The prediction module is used to predict the future cyclic dependency of the cloud resource dynamic dependency graph by using the TGCN algorithm to obtain a cyclic dependency prediction result.

8. The dynamic resource dependency tracking device according to claim 7, characterized in that: The dependency graph construction module includes an attribute extraction submodule, a data processing submodule, a data determination submodule, a data screening submodule, a dependency information determination submodule and a dependency graph construction submodule; The attribute extraction submodule is used to extract data attributes from each information of the cloud resource dependency data to obtain a data attribute set; The data processing submodule is used to perform support and normalization processing on the data attribute set to obtain a preprocessed data set; The data determination submodule is used to obtain the number of features, group the data of resource type information and service type information in all the pre-processed data sets according to the number of features, obtain the number of groups and the feature data group corresponding to each group; and determine the judgment threshold according to the number of groups; The data screening submodule is used to determine the data objects of the corresponding group according to the characteristic data group of each group, and screen the cloud resource dependency data according to the data objects of each group to obtain the number of occurrences of the data objects of each group; The dependency information determination submodule is used to determine the dependency information between the corresponding data object and the resource according to the number of occurrences of each group of the data objects and the judgment threshold; The dependency graph construction submodule is used to construct the cloud resource dynamic dependency graph according to all the dependency information, with resources as nodes and edges as dependency relationships.

9. The dynamic resource dependency tracking device according to claim 8, characterized in that: The data processing submodule is further used to obtain the frequency, data length and data quantity of each data from the data attribute set, and calculate the support of the corresponding data in the data attribute set according to the frequency and data quantity of each data; The mapping space position is obtained, and the preprocessed data after the data is normalized is obtained according to the mapping space position, each data and the data length and the support calculation corresponding to the data; all the preprocessed data constitute a preprocessed data set.

10. A terminal device, characterized in that: including a processor and a memory; The memory is used to store program codes and transmit the program codes to the processor; The processor is used to execute the dynamic resource dependency tracking method as described in any one of claims 1-6 according to the instructions in the program code.

Images