Disaster recovery resource dynamic scheduling method and related device

Abstract

The application discloses a kind of disaster recovery resource dynamic scheduling method and related equipment, the present application is dynamically scheduled with the minimum overall cost as target, can greatly improve resource utilization and reduce overall cost;And, by the multidimensional weight of business system automatically grading, the resource adaptive allocation of business system can be realized, and the resource mismatch problem can be effectively solved;Wherein, by the standardization of heterogeneous cloud resource through unified abstraction, then cross-cloud decision can be made to overcome the management complexity caused by the difference of API, performance, billing mode of different cloud vendors;In addition, by revising iteration, dynamic rescheduling can be realized, to intelligently cope with price fluctuations, performance degradation and other dynamic changes, and can be widely applied in data processing technical field.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and related equipment for dynamic scheduling of disaster recovery resources. Background Technology

[0002] Currently, traditional cloud disaster recovery solutions mainly fall into the following categories: 1. Disaster recovery within a single cloud vendor: Business systems and their disaster recovery are deployed in different availability zones or regions of the same cloud vendor.

[0003] 2. Primary / Secondary Cloud Mode: The production center is deployed on cloud provider A, and the disaster recovery center is deployed on cloud provider B.

[0004] 3. Static resource allocation mode: Fixed computing, storage and network resources are pre-allocated to the disaster recovery system, and the disaster recovery resources are always in a standby state regardless of the actual load of the production system.

[0005] However, existing technologies have at least the following drawbacks and limitations: low resource utilization and high costs: the traditional solutions mentioned above, especially the static resource allocation model, usually require reserving resources of the same or similar scale as the production system for the disaster recovery system to ensure availability at all times. These resources are idle for a long time in non-disaster recovery states, resulting in huge resource waste and high costs; lack of refined business hierarchical management: different business systems have drastically different importance, RTO, and RPO requirements, and existing technologies often adopt the same level of disaster recovery solutions; inability to adapt to the complexity and dynamism of multi-cloud environments: multi-cloud environments are inherently heterogeneous, with significant differences in API interfaces, resource specifications, performance, billing models, and network latency among different cloud vendors; rigid scheduling strategies and lack of intelligence: existing scheduling methods are mostly based on preset simple rules and cannot cope with complex, multi-objective optimization needs. Summary of the Invention

[0006] The main objective of this invention is to provide a method, apparatus, electronic device, storage medium, and program product for dynamic scheduling of disaster recovery resources, aiming to solve at least one problem in the prior art.

[0007] To achieve the above objectives, one aspect of this invention proposes a dynamic scheduling method for disaster recovery resources, the method comprising: By calling the application programming interfaces of multiple cloud service providers, raw data of heterogeneous cloud resources is collected, and the raw data is uniformly abstracted into a standardized internal resource model to generate a standardized global resource status library. In response to the configuration data of the target object, the business level of different business systems is quantitatively evaluated through multi-dimensional disaster recovery requirement indicators, and a business requirement list is generated; each business level is bound to a corresponding service level agreement constraint. Based on a global resource status database and a list of business requirements, with minimizing the overall disaster recovery cost as the core optimization objective, and taking the service level agreement constraints of each business system as constraints, a list of disaster recovery resource deployment strategies is generated through a multi-objective optimization algorithm; the list of disaster recovery resource deployment strategies includes the optimal disaster recovery resources corresponding to each business system. The disaster recovery resources corresponding to the disaster recovery resource deployment strategy list are parsed into API call instructions for each cloud service provider, so that each cloud service provider can execute the API call instructions to complete the scheduling operation of the disaster recovery resources; among which, the scheduling operation includes creation operation, configuration change operation, and release operation; The results of the scheduling operation are visualized to generate a cost-benefit simulation report and pushed to the target object. If the target object inputs a hypothetical parameter modification instruction, the global resource status database and / or business requirement list are updated in response to the hypothetical parameter modification instruction. The process then returns to execute steps based on the global resource status database and business requirement list, with the core optimization objective of minimizing the overall disaster recovery cost, and updates the cost-benefit simulation report.

[0008] In some embodiments, raw data of heterogeneous cloud resources is collected by calling the application programming interfaces of multiple cloud service providers, and the raw data is uniformly abstracted into a standardized internal resource model to generate a standardized global resource state library, including the following steps: Raw data from heterogeneous cloud resources is collected in real time or periodically from the application programming interfaces of multiple cloud service providers. Among them, the raw data from different cloud service providers adopts the format of heterogeneous instances, and the attribute information of the raw data includes resource metadata, performance indicators, pricing information and resource availability status data; The system cleans and aligns the raw data from different cloud service providers. Then, based on predefined standardized resource unit mapping rules, it maps heterogeneous instances from different cloud service providers into standard identifiers according to the target core capabilities. The system associates and stores attribute information with the standard identifiers to generate a standardized global resource status library.

[0009] In some embodiments, in response to the configuration data of the target object, the business level of different business systems is quantitatively evaluated through multi-dimensional disaster recovery requirement indicators to generate a business requirement list, including the following steps: The configuration data of the target object is received through the configuration interface. The configuration data includes the multi-dimensional disaster recovery requirement indicators configured by the target object for each business system, the weight and quantification logic of each disaster recovery requirement indicator, and the multi-dimensional disaster recovery requirement indicators include recovery time target requirements, recovery point target requirements, business importance / financial impact, compliance requirements and cost budget limit. The raw values ​​of each disaster recovery requirement indicator in the multi-dimensional disaster recovery requirement indicators are obtained from the business system. Based on the value range to which the raw values ​​belong, the raw scores are obtained by mapping and transformation through quantification logic. The original scores corresponding to each disaster recovery requirement indicator are weighted and summed to obtain the disaster recovery priority score. Finally, the disaster recovery priority score is compared with the preset score range to determine the business level corresponding to each business system. Based on the business level, a pre-defined service level agreement constraint is matched for each business system, generating a list of business requirements.

[0010] In some embodiments, based on a global resource status database and a list of business requirements, with minimizing the overall disaster recovery cost as the core optimization objective, and using the service level agreement constraints of each business system as constraints, a list of disaster recovery resource deployment strategies is generated through a multi-objective optimization algorithm, including the following steps: Based on the service level agreement constraints corresponding to each business system in the business requirements list, all candidate resources in the global resource status database are screened, and candidate resources that do not meet any service level agreement constraints are removed, thus constructing a feasible resource pool corresponding to each business system. By using a feasible resource pool to constrain resource selection, minimizing the total cost of disaster recovery resources for all business systems is taken as the objective function. A multi-objective optimization algorithm is used to solve for the optimal disaster recovery resources for each business system, and a list of disaster recovery resource deployment strategies is generated.

[0011] In some embodiments, the disaster recovery resources corresponding to the disaster recovery resource deployment strategy list are parsed into API call instructions corresponding to each cloud service provider, so that each cloud service provider executes the API call instructions to complete the scheduling operation of the disaster recovery resources, including the following steps: The optimal disaster recovery resources are used as standardized resources, and the correspondence between each business system and the standardized resources is determined based on the disaster recovery resource deployment strategy list. Based on the API call specification mapping table of each cloud service provider, the corresponding relationship is translated in reverse into the specific API request parameters of the corresponding cloud service provider as API call instructions; The corresponding cloud service provider's target API is invoked through API call instructions to sequentially execute each atomic task in the preset resource deployment workflow, and the execution results of each atomic task are collected to complete the scheduling operation of disaster recovery resources.

[0012] In some embodiments, the results of the scheduling operation are visualized to generate a cost-benefit simulation report and pushed to the target object. If the target object inputs a hypothetical parameter modification instruction, the global resource state database and / or business requirement list are updated in response to the hypothetical parameter modification instruction, including the following steps: The results of scheduling operations are rendered and displayed graphically, generating a cost-benefit simulation report; The cost-benefit simulation report includes a resource topology map, a cost trend curve, and a resource utilization heatmap. If the cost-benefit simulation report has been updated, a difference analysis report will be generated based on the comparison of the cost-benefit simulation reports before and after the update. The variance analysis report is annotated in the updated cost-benefit simulation report and pushed to the target audience.

[0013] In some embodiments, the method further includes the following steps: In response to the progression of time, real-time data from heterogeneous cloud resources is collected periodically by calling the application programming interfaces of multiple cloud service providers; If the parameter changes in real-time data compared to the original data trigger any of the preset monitoring event triggers, the global resource status database will be updated based on the real-time data. Based on the updated global resource status database and the original business requirement list, return to the execution steps based on the global resource status database and the business requirement list, with the core optimization goal of minimizing the overall disaster recovery cost, and update the disaster recovery resource deployment strategy list. Compare the differences between the disaster recovery resource deployment strategy list before and after the update, and generate an incremental change plan; The disaster recovery resources corresponding to the incremental change plan are parsed into incremental API call instructions for each cloud service provider, so that each cloud service provider can execute the incremental API call instructions to complete the incremental dynamic migration of disaster recovery resources.

[0014] To achieve the above objectives, another aspect of the present invention provides a disaster recovery resource dynamic scheduling device, the device comprising: The first module is used to collect raw data of heterogeneous cloud resources by calling the application programming interfaces of multiple cloud service providers, and to abstract the raw data into a standardized internal resource model to generate a standardized global resource status library. The second module is used to respond to the configuration data of the target object, quantitatively evaluate the business level of different business systems through multi-dimensional disaster recovery requirement indicators, and generate a list of business requirements; each business level is bound to a corresponding service level agreement constraint. The third module is used to generate a list of disaster recovery resource deployment strategies based on a global resource status database and a list of business requirements, with the core optimization objective of minimizing the overall disaster recovery cost. It takes the service level agreement constraints of each business system as constraints and uses a multi-objective optimization algorithm to generate the list of disaster recovery resource deployment strategies. The list of disaster recovery resource deployment strategies includes the optimal disaster recovery resources for each business system. The fourth module is used to parse the disaster recovery resources corresponding to the disaster recovery resource deployment strategy list into API call instructions corresponding to each cloud service provider, so that each cloud service provider can execute the API call instructions to complete the scheduling operation of the disaster recovery resources; among which, the scheduling operation includes creation operation, configuration change operation, and release operation; The fifth module is used to visualize the results of scheduling operations, generate a cost-benefit simulation report, and push it to the target object. If the target object inputs a hypothetical parameter modification instruction, it updates the global resource status database and / or business requirement list in response to the hypothetical parameter modification instruction, and returns to execute the operation of the third module to update the cost-benefit simulation report.

[0015] In some embodiments, the apparatus further includes a sixth module for performing the following operations: In response to the progression of time, real-time data from heterogeneous cloud resources is collected periodically by calling the application programming interfaces of multiple cloud service providers; If the parameter changes in real-time data compared to the original data trigger any of the preset monitoring event triggers, the global resource status database will be updated based on the real-time data. Based on the updated global resource status database and the original business requirement list, return to the execution of the third module's operation to update the disaster recovery resource deployment strategy list; Compare the differences between the disaster recovery resource deployment strategy list before and after the update, and generate an incremental change plan; The disaster recovery resources corresponding to the incremental change plan are parsed into incremental API call instructions for each cloud service provider, so that each cloud service provider can execute the incremental API call instructions to complete the incremental dynamic migration of disaster recovery resources.

[0016] To achieve the above objectives, another aspect of the present invention provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the aforementioned method.

[0017] To achieve the above objectives, another aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned method.

[0018] To achieve the above objectives, another aspect of the present invention provides a computer program product, including a computer program that, when executed by a processor, implements the aforementioned method.

[0019] The embodiments of this invention include at least the following beneficial effects: This invention provides a method, apparatus, electronic device, storage medium, and program product for dynamic scheduling of disaster recovery resources. This solution collects raw data from heterogeneous cloud resources by calling the application programming interfaces (APIs) of multiple cloud service providers, unifies and abstracts the raw data into a standardized internal resource model, and generates a standardized global resource status library. Responding to the configuration data of the target object, it quantitatively evaluates the business levels of different business systems through multi-dimensional disaster recovery requirement indicators, generating a business requirement list. Each business level is bound to a corresponding service level agreement (SLA) constraint. Based on the global resource status library and the business requirement list, with minimizing the overall disaster recovery cost as the core optimization objective, and using the SLA constraints of each business system as constraints, a multi-objective optimization algorithm is used to generate disaster recovery resources. The system comprises: a source deployment strategy list; a disaster recovery resource deployment strategy list including the optimal disaster recovery resources for each business system; parsing the disaster recovery resources corresponding to the disaster recovery resource deployment strategy list into API call instructions for each cloud service provider, enabling each cloud service provider to execute the API call instructions and complete the scheduling operation of the disaster recovery resources; the scheduling operation includes creation, configuration change, and release operations; visualizing the results of the scheduling operation to generate a cost-benefit simulation report and pushing it to the target object; if the target object inputs a hypothetical parameter modification instruction, updating the global resource status database and / or business requirement list in response to the hypothetical parameter modification instruction, and returning to execute steps based on the global resource status database and business requirement list with the core optimization objective of minimizing the overall disaster recovery cost, thereby updating the cost-benefit simulation report. This invention, through dynamic scheduling aimed at minimizing total cost, can significantly improve resource utilization and reduce total cost of ownership. Furthermore, by automatically classifying resources based on the multi-dimensional weights of business systems, it can achieve adaptive resource allocation, effectively solving the resource mismatch problem. Moreover, by standardizing heterogeneous cloud resources through unified abstraction, it can make cross-cloud decisions to overcome the management complexity caused by differences in APIs, performance, and billing models among different cloud vendors. In addition, through revision and iteration, this invention can achieve dynamic rescheduling to intelligently respond to dynamic changes such as price fluctuations and performance degradation. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of an implementation environment for the dynamic scheduling method for disaster recovery resources provided in this embodiment of the invention; Figure 2 This is a flowchart illustrating the dynamic scheduling method for disaster recovery resources provided in an embodiment of the present invention. Figure 3This is a schematic diagram of the overall architecture of the dynamic scheduling method for disaster recovery resources provided in this embodiment of the invention. Figure 4 This is a schematic diagram illustrating the overall process of the dynamic scheduling method for disaster recovery resources provided in this embodiment of the invention. Figure 5 This is a schematic diagram of the structure of the disaster recovery resource dynamic scheduling device provided in an embodiment of the present invention; Figure 6 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the embodiments of this invention; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this invention as detailed in the appended claims.

[0022] It is understood that the terms "first," "second," etc., used in this invention may be used to describe various concepts, but unless specifically stated otherwise, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of embodiments of this invention, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the words "if" or "when" as used herein may be interpreted as "when," "in response to determination," or "in the event of a determination."

[0023] The terms “at least one,” “multiple,” “each,” “any,” etc., used in this invention, “at least one” includes one, two, or more than two; “multiple” includes two or more than two; “each” refers to each of the corresponding multiple; and “any” refers to any one of the multiple.

[0024] Unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this invention is for descriptive purposes only and is not intended to limit the invention.

[0025] To facilitate understanding of the technical solution of this invention, the following explanations are provided regarding the technical terms that may be involved in the technical solution of this invention: API: Application Programming Interface; RTO: Recovery Time Objective; RPO: Recovery Point Objective; SLA: Service Level Agreement; TCO: Total Cost of Ownership; In related technologies, there is an urgent need for a technical solution that can overcome the above-mentioned shortcomings. It should be able to intelligently perceive business needs and the status of multi-cloud environments, and make dynamic decisions, thereby significantly reducing the total cost of ownership (TCO) of disaster recovery while ensuring business continuity.

[0026] In view of this, this invention provides a method and related equipment for dynamic scheduling of disaster recovery resources. This method collects raw data from heterogeneous cloud resources by calling the application programming interfaces (APIs) of multiple cloud service providers, abstracts the raw data into a standardized internal resource model, and generates a standardized global resource status database. Responding to the configuration data of the target object, it quantitatively evaluates the business levels of different business systems through multi-dimensional disaster recovery requirement indicators, generating a business requirement list. Each business level is bound to a corresponding service level agreement (SLA) constraint. Based on the global resource status database and the business requirement list, with minimizing the overall disaster recovery cost as the core optimization objective, and using the SLA constraints of each business system as constraints, a disaster recovery resource deployment strategy list is generated through a multi-objective optimization algorithm. The disaster recovery resource deployment strategy list includes the optimal disaster recovery resources corresponding to each business system. The disaster recovery resources corresponding to the disaster recovery resource deployment strategy list are parsed into API call instructions corresponding to each cloud service provider, enabling each cloud service provider to execute the API call instructions and complete the scheduling operation of the disaster recovery resources. The scheduling operation includes creation, configuration change, and release operations. The results of the scheduling operation are visualized to generate a cost-benefit simulation report and pushed to the target object. If the target object inputs a hypothetical parameter modification instruction, the global resource status database and / or business requirement list are updated in response to the hypothetical parameter modification instruction. The process then returns to execute steps based on the global resource status database and business requirement list, with the core optimization objective of minimizing the overall disaster recovery cost, to update the cost-benefit simulation report. This invention, through dynamic scheduling aimed at minimizing total cost, can significantly improve resource utilization and reduce total cost of ownership. Furthermore, by automatically classifying resources based on the multi-dimensional weights of business systems, it can achieve adaptive resource allocation, effectively solving the resource mismatch problem. Moreover, by standardizing heterogeneous cloud resources through unified abstraction, it can make cross-cloud decisions to overcome the management complexity caused by differences in APIs, performance, and billing models among different cloud vendors. In addition, through revision and iteration, this invention can achieve dynamic rescheduling to intelligently respond to dynamic changes such as price fluctuations and performance degradation.

[0027] It is understood that the disaster recovery resource dynamic scheduling method provided by this invention can be applied to any computer device with data processing and computing capabilities, and this computer device can be various terminals or servers. When the computer device in the embodiment is a server, the server is an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms. Optionally, the terminal can be a smartphone, tablet, laptop, or desktop computer, but it is not limited to these.

[0028] like Figure 1 The diagram shown is a schematic representation of an implementation environment provided by an embodiment of the present invention. (Refer to...) Figure 1 The implementation environment includes at least one terminal 102 and a server 101. The terminal 102 and the server 101 can be connected via a network, either wirelessly or via a wired connection, to complete data transmission and exchange.

[0029] Server 101 can be a standalone physical server, a server cluster or distributed system consisting of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms.

[0030] Additionally, server 101 can also be a node server in a blockchain network. Blockchain is a novel application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms.

[0031] Terminal 102 can be a smartphone, tablet computer, laptop computer, desktop computer, smart speaker, smartwatch, etc., but is not limited to these. Terminal 102 and server 101 can be directly or indirectly connected via wired or wireless communication, and this embodiment of the invention does not impose any limitations.

[0032] For example, based on Figure 1 The implementation environment shown in this embodiment of the invention provides a dynamic scheduling method for disaster recovery resources. The following description uses the application of this dynamic scheduling method for disaster recovery resources in server 101 as an example. It can be understood that this dynamic scheduling method for disaster recovery resources can also be applied to terminal 102.

[0033] Reference Figure 2 , Figure 2 This is an optional flowchart of the dynamic scheduling method for disaster recovery resources provided in the embodiments of the present invention. The executing entity of the dynamic scheduling method for disaster recovery resources can be any of the aforementioned computer devices (including servers or terminals). Figure 2 The method may include, but is not limited to, steps S100 to S500.

[0034] Step S100: Collect raw data of heterogeneous cloud resources by calling the application programming interfaces of multiple cloud service providers, abstract the raw data into a standardized internal resource model, and generate a standardized global resource status library. It should be noted that in some embodiments, step S100 may include the following steps: collecting raw data of heterogeneous cloud resources in real time or periodically from the application programming interfaces of multiple cloud service providers; wherein the raw data of different cloud service providers adopts the format of heterogeneous instances, and the attribute information of the raw data includes resource metadata, performance indicators, price information and resource availability status data; cleaning and format alignment of the raw data of different cloud service providers, and then mapping the heterogeneous instances of different cloud service providers into standard identifiers according to the target core capabilities according to the predefined standardized resource unit mapping rules, associating the attribute information with the standard identifiers and storing them together to generate a standardized global resource status library.

[0035] For example, in some specific implementations, multi-cloud resource data collection and unified abstraction can be achieved as follows: This invention collects resource data by calling the API interfaces of various cloud platforms. Below is an example of the collected real-time data: 1) Alibaba Cloud: Region: East China 1 (City A); Instance type: ecs.g6.large (2vCPU, 8GiB Mem); Status: Available; Price: On-demand price 0.8 RMB / hour; Preemptive instance price 0.25 RMB / hour (recovery risk rate 5%); Network latency (to data center): <10ms.

[0036] 2) Amazon Web Services (AWS): Region: City B (cn-north-1); Instance type: m5.large (2 vCPUs, 8GiB Mem); Status: Available; Price: On-demand price of 1.1 yuan / hour; Network latency (to data center): <15ms.

[0037] 3) Microsoft Azure: Region: East Asia (City C); Instance type: Standard_D2s_v3 (2 vCPUs, 8GiB Mem); Status: Available; Price: On-demand price of 1.0 RMB / hour; Network latency (to data center): <35ms.

[0038] Unified Abstraction: The system unifies the above heterogeneous instances into an internal resource identifier: `Compute_Unit_V2` (representing a computing resource with 2vCPU / 8GiB specifications), and records their respective price and latency attributes.

[0039] Step S200: In response to the configuration data of the target object, the business level of different business systems is quantitatively evaluated through multi-dimensional disaster recovery requirement indicators, and a business requirement list is generated. Each service level is bound to a corresponding service level agreement constraint; It should be noted that in some embodiments, step S200 may include the following steps: receiving configuration data of the target object through a configuration interface; wherein, the configuration data includes multi-dimensional disaster recovery requirement indicators configured by the target object for each business system, the weights and quantification logic corresponding to each disaster recovery requirement indicator, and the multi-dimensional disaster recovery requirement indicators including recovery time target requirements, recovery point target requirements, business importance / financial impact, compliance requirements, and cost budget ceiling; obtaining the original value corresponding to each disaster recovery requirement indicator from the business system, and mapping and transforming it through quantification logic based on the value range to which the original value belongs to obtain the original score; using weights to perform weighted summation of the original scores corresponding to each disaster recovery requirement indicator to obtain a disaster recovery priority score; finally, comparing the disaster recovery priority score with a preset score range to determine the business level corresponding to each business system; matching preset service level agreement constraints for each business system according to the business level to generate a business requirement list.

[0040] For example, in some specific implementations, the quantification and automatic classification of disaster recovery requirements for business systems can be achieved as follows: Suppose there are the following three business systems to be protected: 1) Order-service: RTO requirement: <5 minutes (weight: 40%); RPO requirement: ≈0 (weight: 40%); Maximum tolerable loss per minute: 1000 yuan (weight: 20%); The system automatically calculates its disaster recovery priority score: (1×0.4 + 1×0.4 + 1×0.2)×100 = 100 points (the higher the score, the more important it is); Automatic grading result: Core level (Tier-0).

[0041] 2) Product search service: RTO requirement: <30 minutes (weight: 40%); RPO requirement: <1 minute (weight: 40%); Maximum tolerable loss per minute: 100 yuan (weight: 20%); The system automatically calculates its disaster recovery priority score: (0.7×0.4 + 0.8×0.4 + 0.8×0.2)×100 = 76 points; Automatic classification result: Tier-1.

[0042] 3) Log analysis service: RTO requirement: <4 hours (weight: 40%); RPO requirement: <1 hour (weight: 40%); Maximum tolerable loss per minute: 10 yuan (weight: 20%); The system automatically calculates its disaster recovery priority score: (0.3×0.4 + 0.5×0.4 + 0.5×0.2)×100 = 42 points; Automatic classification result: General level (Tier-2).

[0043] The original scores for RTO requirements, RPO requirements, and maximum tolerable loss per minute are obtained by mapping the numerical values ​​to a range (mapped to the range [0,1]). The shorter the time corresponding to the RTO and RPO requirements, the higher their original scores; the larger the amount corresponding to the maximum tolerable loss per minute, the higher their original scores. (The scores for the above three services are only examples; the mapping relationship between the numerical ranges of RPO requirements and maximum tolerable loss per minute and the original scores can be adjusted according to actual application settings.) Finally, the disaster recovery priority score is compared with the preset score range (e.g., core level: 90-100 points, important level: 60-89 points, general level: 0-59 points) to determine the corresponding business level, or the disaster recovery priority score is directly sorted according to the score to determine the business level.

[0044] Step S300: Based on the global resource status database and business requirement list, with minimizing the overall disaster recovery cost as the core optimization objective, the service level agreement constraints of each business system are used as constraints, and a disaster recovery resource deployment strategy list is generated through a multi-objective optimization algorithm. The list of disaster recovery resource deployment strategies includes the optimal disaster recovery resources for each business system. It should be noted that in some embodiments, step S300 may include the following steps: based on the service level agreement constraints corresponding to each business system in the business requirements list, all candidate resources in the global resource status database are screened, and candidate resources that do not meet any service level agreement constraints are removed, thereby constructing a feasible resource pool corresponding to each business system; resource selection constraints are applied through the feasible resource pool, and the minimum total cost of all business disaster recovery resources is used as the objective function. A multi-objective optimization algorithm is then used to solve for the optimal disaster recovery resource corresponding to each business system, thereby generating a disaster recovery resource deployment strategy list.

[0045] For example, in some specific implementations, for each business system, based on the SLA constraints corresponding to the business requirement list, all candidate resources in the standardized global resource status library are screened, and resources that do not meet any constraint are eliminated (e.g., for businesses that prohibit the use of preemptive instances, resources with a recovery risk rate > 0 are screened), thus constructing a feasible resource pool for each business. Then, with minimizing the total cost of disaster recovery resources for all businesses as the objective function and with the constraint that each business must and can only select one resource from the feasible resource pool, a multi-objective optimization algorithm (e.g., greedy algorithm, dynamic programming, or reinforcement learning model) is used to solve the problem, allocating an optimal standardized resource unit for each business. The constraints for each business level can be configured as follows: for example, core-level businesses are forced to choose on-demand instances, which have higher costs but stable performance and no recovery risk; general-level businesses prioritize preemptive instances, which have the lowest costs but can tolerate a certain recovery risk. Finally, a disaster recovery resource deployment strategy list is generated through constraint and cost optimization solutions, and its format may include business identifiers and standardized resource unit identifiers @specific cloud vendor / region / availability zone / instance billing type.

[0046] Step S400: Parse the disaster recovery resources corresponding to the disaster recovery resource deployment strategy list into API call instructions corresponding to each cloud service provider, so that each cloud service provider can execute the API call instructions to complete the scheduling operation of the disaster recovery resources. The scheduling operations include creation operations, configuration change operations, and release operations; It should be noted that in some embodiments, step S400 may include the following steps: taking the optimal disaster recovery resource as a standardized resource, determining the correspondence between each business system and the standardized resource based on the disaster recovery resource deployment strategy list; based on the API call specification mapping table of each cloud service provider, translating the correspondence into the specific API request parameters of the corresponding cloud service provider as API call instructions; calling the target API of the corresponding cloud service provider through the API call instructions to sequentially execute each atomic task in the preset resource deployment workflow, and collecting the execution results of each atomic task to complete the scheduling operation of the disaster recovery resource.

[0047] For example, in some specific implementations, firstly, based on the correspondence between each business system and standardized resources in the policy list, the API call specification mapping table of each cloud service provider is reverse-translated into specific API request parameters that can be recognized by the corresponding cloud service provider (for example, mapping the standardized resource unit Compute_Unit_V2 to Alibaba Cloud's ecs.g6.large instance specification, and combining parameters such as region, image, and security group); then, the complete resource deployment operation (such as the creation, configuration, data synchronization, and traffic switching of disaster recovery instances) is decomposed into a series of atomic tasks, and an ordered workflow is formed according to the dependencies; finally, by calling operation and maintenance automation tools (including Ansible or Terraform) or directly calling the APIs of each cloud service provider, the atomic tasks in the workflow are executed sequentially, and the execution status of each task is recorded, thereby recording the execution result log.

[0048] Step S500: Visualize the results of the scheduling operation to generate a cost-benefit simulation report and push it to the target object. If the target object inputs a hypothetical parameter modification instruction, update the global resource status database and / or business requirement list in response to the hypothetical parameter modification instruction, and return to execute the steps based on the global resource status database and business requirement list with the core optimization goal of minimizing the overall disaster recovery cost, so as to update the cost-benefit simulation report. It should be noted that in some embodiments, the results of the scheduling operation are visualized to generate a cost-benefit simulation report and pushed to the target object. If the target object inputs a hypothetical parameter modification instruction, the global resource status database and / or business requirement list are updated in response to the hypothetical parameter modification instruction. This may include the following steps: rendering and displaying the results of the scheduling operation graphically to generate a cost-benefit simulation report; wherein, the cost-benefit simulation report includes a resource topology map, a cost trend curve, and a resource utilization heatmap; if the cost-benefit simulation report has undergone an update operation, a difference analysis report is generated based on the comparison of the cost-benefit simulation reports before and after the update; the difference analysis report is marked in the updated cost-benefit simulation report and pushed to the target object.

[0049] For example, in some specific implementations, taking an e-commerce platform as an example, the visualization processing and simulation of the embodiments of the present invention can be achieved as follows: the performance and cost data of the deployed disaster recovery resources can be obtained in real time through WebSocket, and then rendered on the front end using the ECharts chart library. Specifically, the resource topology map can display the instance nodes of each cloud region through a force-directed graph, wherein different health statuses are indicated by node colors; the cost trend curve can display the daily cost of each business through a stacked area graph; the resource utilization heatmap can display CPU / memory usage through different colors or gradients of color depth.

[0050] In this scenario, suppose operations personnel input a change to reduce the RTO requirement for the order service from 5 minutes to 1 minute in the simulation panel. Then, a replica of the application's current in-memory business requirement list is created, and the same multi-objective optimization algorithm is called again to obtain an updated cost-benefit simulation report. This updated report is then compared with the previous cost-benefit simulation report to quantify the differences. For example, the order service, due to its stricter RTO, is allocated to a lower-latency AWS instance, leading to increased costs (specific comparative values ​​can be determined based on the parameters in the previous and current cost-benefit simulation reports). The comparison format can be, for example, a dual-bar chart comparing the actual and simulated costs for each business, or a radar chart comparing SLA compliance margins. The final comparison report is automatically pushed to operations personnel to assist their decision on whether to adopt the new strategy.

[0051] It should be noted that in some embodiments, the method may further include the following steps: in response to the progression of time, periodically collect real-time data of heterogeneous cloud resources by calling the application programming interfaces of multiple cloud service providers; if the parameter changes of the real-time data compared with the original data trigger any of the preset monitoring event triggers, update the global resource status database based on the real-time data; based on the updated global resource status database and the original business requirement list, return to execute the steps based on the global resource status database and the business requirement list, with the core optimization goal of minimizing the overall disaster recovery cost, to update the disaster recovery resource deployment strategy list; compare the differences between the disaster recovery resource deployment strategy lists before and after the update, and generate an incremental change plan; parse the disaster recovery resources corresponding to the incremental change plan into incremental API call instructions corresponding to each cloud service provider, so that each cloud service provider executes the incremental API call instructions to complete the incremental dynamic migration of disaster recovery resources.

[0052] For example, in some specific implementations, a set of monitoring event triggers is first preset, such as: the risk rate of preemptive instance reclamation exceeding a threshold, network latency exceeding limits for core business, and significant fluctuations in cloud resource prices. Then, real-time performance, price, and status data of deployed disaster recovery resources are continuously collected by periodically calling the APIs of various cloud service providers. When any monitoring event is triggered by a change in real-time data compared to historical benchmarks, the global resource status database is automatically updated, marking the affected resources as "high-risk" or "unavailable." Subsequently, based on the updated status database and the original business requirement list, multi-objective optimization decisions are re-executed to generate a new list of disaster recovery resource deployment strategies. Then, the old and new strategies are compared, and only the changed parts are extracted to form an incremental change plan, rather than a full rebuild. Finally, the standardized resources in the incremental change plan are parsed into API call instructions from specific cloud vendors and executed in atomic task order (first creating new resources, then synchronizing data, then switching traffic, and finally releasing old resources), achieving uninterrupted or minimally interrupted dynamic migration of disaster recovery resources.

[0053] To explain in detail the principle of the technical solution of the present invention, the overall process of the present invention will be described below with reference to some specific embodiments. It is easy to understand that the following is an explanation of the technical principle of the present invention and should not be regarded as a limitation of the present invention.

[0054] In view of the shortcomings of the prior art, the purpose of this invention is to overcome the deficiencies of the prior art and provide a method for dynamic scheduling of intelligent disaster recovery resources for multi-cloud environments.

[0055] Specifically, the present invention aims to achieve the following objectives: 1. This paper provides an intelligent business classification and quantification method, which aims to solve the problem of "one-size-fits-all" disaster recovery strategies in existing technologies. By constructing a multi-dimensional weighted evaluation model, different businesses are automatically and accurately classified into core, important, and general levels, and the disaster recovery resource requirements and cost constraints of each level are quantified, thereby achieving refined resource management.

[0056] 2. This paper provides a unified abstraction and perception solution for heterogeneous resources in multi-cloud environments, aiming to solve the management complexity caused by the heterogeneity of APIs, performance and billing models in multi-cloud environments. By building a unified resource abstraction layer, it can perceive and standardize the resource status, performance indicators and price information of various cloud vendors in real time, and provide a reliable data foundation for intelligent scheduling.

[0057] 3. A dynamic scheduling decision-making mechanism based on optimization algorithms is provided to solve the problems of rigid and costly static resource allocation. By introducing intelligent algorithms (such as reinforcement learning, dynamic programming or greedy algorithms), the optimal disaster recovery resource deployment strategy is dynamically and automatically selected for each service based on the service level, real-time resource price and performance. Under the premise of absolutely guaranteeing the SLA of high-level services, the overall disaster recovery cost is significantly reduced.

[0058] 4. Provides a cost-benefit visualization simulation function, which aims to solve the pain point of lack of data support for enterprise decision-making. It allows users to simulate the costs and effects under different disaster recovery strategies, providing a scientific basis for resource planning and realizing the continuous optimization of the disaster recovery system.

[0059] The ultimate goal of this invention is to achieve a transformation of disaster recovery resources from "static configuration" to "dynamic intelligent scheduling" through the comprehensive application of the above methods, thereby fundamentally solving the core contradiction in the field of multi-cloud disaster recovery where it is difficult to balance cost, security and flexibility.

[0060] like Figure 3 The diagram shown is a schematic representation of a system architecture example corresponding to the dynamic scheduling method for disaster recovery resources provided by this invention. In some specific embodiments, the dynamic scheduling method for disaster recovery resources of this invention can be implemented through the following system architecture: 1. Data Acquisition & Abstraction Module: Function: Responsible for API integration with major cloud service providers (such as Alibaba Cloud, AWS, Azure, etc.).

[0061] Specific tasks include: collecting metadata, performance metrics (CPU utilization, network latency, disk IOPS), and real-time pricing information (on-demand, reserved, and preemptible instance pricing) of heterogeneous cloud resources in real time or periodically. The collected heterogeneous data will then be uniformly abstracted into a standardized, internally identifiable data model (e.g., defining a unified "computing unit" to benchmark vCPUs from different vendors).

[0062] 2. Business Perception & Tiering Module: Function: This is the core of achieving refined management.

[0063] Specific tasks: Provide a configuration interface that allows users to set multi-dimensional weights for each business system (such as RTO, RPO, business importance, compliance, and cost budget). This module has a built-in quantitative evaluation algorithm that automatically calculates the "disaster recovery priority score" for each business based on the user's configuration and automatically classifies it into "core," "important," and "general" levels according to the score range.

[0064] 3. Intelligent Dynamic Scheduling Engine: Function: The brain of this invention is responsible for making optimal decisions.

[0065] Specific tasks: Built-in scheduling decision algorithm (e.g., a multi-objective optimization algorithm based on a cost function). This engine receives standardized data from the two modules mentioned above, prioritizing meeting business RTO / RPO constraints and using the lowest overall cost as the core optimization objective, dynamically calculating the optimal disaster recovery resource deployment location (including cloud vendor, region, availability zone, and instance type) for each business. This engine can be executed periodically or triggered by events (such as price changes or performance fluctuations).

[0066] 4. Resource Orchestration & Execution Module: Function: Responsible for translating decisions into action.

[0067] Specific task: By calling the APIs of various cloud vendors, automatically execute the decision results of the scheduling engine to complete lifecycle management operations such as resource creation, release, and configuration changes, forming a closed loop.

[0068] 5. Visualization & Simulation Module: Functionality: Provides a management interface to enhance user experience and decision support capabilities.

[0069] Specific tasks: To graphically display the current multi-cloud disaster recovery architecture, resource utilization, and cost structure. Provide a "What-If" simulation function, allowing users to modify business tiers or scheduling policies and immediately see the predicted cost and performance changes.

[0070] The following is combined Figure 4 The present invention will be further described in detail with specific implementation examples of the dynamic scheduling method for disaster recovery resources, taking the disaster recovery of an e-commerce company's business system in a multi-cloud environment as an example: 1. Multi-cloud resource data collection and unified abstraction: This invention collects resource data by calling the API interfaces of various cloud platforms. Below is an example of the collected real-time data: 1) Alibaba Cloud: Region: East China 1 (City A); Instance type: ecs.g6.large (2vCPU, 8GiB Mem); Status: Available; Price: On-demand price 0.8 RMB / hour; Preemptive instance price 0.25 RMB / hour (recovery risk rate 5%); Network latency (to data center): <10ms.

[0071] 2) Amazon Web Services (AWS): Region: City B (cn-north-1); Instance type: m5.large (2 vCPUs, 8GiB Mem); Status: Available; Price: On-demand price of 1.1 yuan / hour; Network latency (to data center): <15ms.

[0072] 3) Microsoft Azure: Region: East Asia (City C); Instance type: Standard_D2s_v3 (2 vCPUs, 8GiB Mem); Status: Available; Price: On-demand price of 1.0 RMB / hour; Network latency (to data center): <35ms.

[0073] Unified Abstraction: The system unifies the above heterogeneous instances into an internal resource identifier: `Compute_Unit_V2` (representing a computing resource with 2vCPU / 8GiB specifications), and records their respective price and latency attributes.

[0074] 2. Quantification and automatic classification of disaster recovery requirements for business systems: Suppose there are the following three business systems to be protected: 1) Order-service: RTO requirement: <5 minutes (weight: 40%); RPO requirement: ≈0 (weight: 40%); Maximum tolerable loss per minute: 1000 yuan (weight: 20%); The system automatically calculates its disaster recovery priority score: (1×0.4 + 1×0.4 + 1×0.2)×100 = 100 points (the higher the score, the more important it is); Automatic grading result: Core level (Tier-0).

[0075] 2) Product search service: RTO requirement: <30 minutes (weight: 40%); RPO requirement: <1 minute (weight: 40%); Maximum tolerable loss per minute: 100 yuan (weight: 20%); The system automatically calculates its disaster recovery priority score: (0.7×0.4 + 0.8×0.4 + 0.8×0.2)×100 = 76 points; Automatic classification result: Tier-1.

[0076] 3) Log analysis service: RTO requirement: <4 hours (weight: 40%); RPO requirement: <1 hour (weight: 40%); Maximum tolerable loss per minute: 10 yuan (weight: 20%); The system automatically calculates its disaster recovery priority score: (0.3×0.4 + 0.5×0.4 + 0.5×0.2)×100 = 42 points; Automatic classification result: General level (Tier-2).

[0077] 3. Intelligent dynamic scheduling decision generation: Based on the above information, the scheduling engine makes decisions with the goal of "minimizing total cost while meeting SLA".

[0078] For 'order-service' (core level): extremely low network latency and extremely high stability are required. Preemptive instances cannot be used. Among Alibaba Cloud and AWS, which meet the latency requirements, Alibaba Cloud On-Demand instances, which have lower prices, can meet all the strict constraints and have relatively low costs.

[0079] For 'search-service' (heavyweight): good performance and stability are required, with slight cost optimization allowed, so Alibaba Cloud On-Demand Instances are chosen. Although AWS and Azure also meet the requirements, Alibaba Cloud has the lowest latency and is more expensive than AWS.

[0080] For 'log-service' (general level): Since it is not sensitive to latency and interruption, and cost is the primary factor, Alibaba Cloud preemptive instances are chosen. The price is extremely low (0.25 yuan / hour). Although there is a risk of reclaiming, its RTO requirement is 4 hours, which is more than enough time to restart an instance after reclamation, perfectly meeting the SLA.

[0081] The final scheduling scheme example is shown in Table 1 below: Table 1

[0082] Referring to Table 1 above, compared with the traditional solution (all using on-demand instances): the total monthly cost is approximately 1728 yuan. The cost savings of this invention are: (1728 - 1332) / 1728 ≈ 23%.

[0083] 4. Scheduling plan execution and status monitoring: The intelligent scheduling engine uses operational automation tools such as Ansible or Terraform to call various cloud APIs and execute the above decisions: 1) Create on-demand instances for 'order-service' and 'search-service' on Alibaba Cloud Hangzhou; 2) Create a preemptible instance for 'log-service' on Alibaba Cloud Hangzhou; 3) The system continuously monitors the health status of all disaster recovery instances and price / resource changes on various cloud platforms; 4) Monitoring Feedback Example: Suppose the monitoring module detects that the preemptible instance inventory in Alibaba Cloud Hangzhou is insufficient, and the risk of instance reclamation rises to 50%. The scheduling engine is immediately triggered, and step 3 is re-executed. The new decision may be to migrate the disaster recovery instance of 'log-service' to Alibaba Cloud South China 1 region where there are available preemptible instances, and generate a migration record and reason explanation in the management interface.

[0084] In summary, this invention constructs a configurable and quantifiable business tiering model. The model's input includes multiple dimensions of indicators and their weights, such as RTO, RPO, financial impact, and compliance requirements. An algorithm automatically calculates the overall importance score of each business and categorizes them into different tiers based on preset thresholds. A differentiated disaster recovery resource strategy template is automatically matched to each tier. Furthermore, this invention uses a unified resource abstraction layer independent of cloud vendors to convert heterogeneous resources from different cloud vendors into standardized internal resource units (such as CU and SU). It also collects real-time status, performance (such as latency and IOPS), and pricing information of each cloud resource, converting it into a unified format and storing it in a resource status database, providing a consistent and real-time global resource view for upper-level decision-making. Simultaneously, this invention constructs an optimized scheduling decision engine. This engine, with the core objective of "minimizing overall disaster recovery costs" and the condition of "meeting the RTO / RPO requirements defined for each business tier," dynamically selects the optimal disaster recovery resource deployment strategy for each business. In addition, this invention provides a workflow engine that can translate scheduling decisions into specific cloud API call instructions. It decomposes complex scheduling operations into atomic steps and executes them automatically. This invention achieves dynamic scheduling through intelligent algorithms, matching the most economical resources in real time according to business needs. It enables differentiated disaster recovery strategies, ensuring the continuity of core businesses while optimizing the costs of non-critical businesses. Furthermore, through a unified resource abstraction layer, it standardizes information from various cloud vendors in real time, providing a consistent and reliable global view for intelligent decision-making. In addition, it introduces an intelligent optimization algorithm engine to dynamically calculate the optimal deployment strategy, achieving a leap from fixed rules to intelligent decision-making. Ultimately, it can achieve full closed-loop automation from resource perception to decision execution.

[0085] Specifically, this invention can be applied to the financial industry to ensure the high availability of core transaction systems while significantly reducing disaster recovery compliance costs; it can also be applied to e-commerce platforms to cope with the elastic expansion of traffic during peak promotion periods; in addition, it can provide differentiated data protection for different important public services such as medical insurance and social security within a limited budget; at the same time, this invention can provide cost-effective disaster recovery solutions for large enterprises' ERP and CRM systems migrating to the cloud.

[0086] Compared with the prior art, the present invention has at least the following beneficial effects: 1. Through intelligent scheduling algorithms, the most economical resources are accurately matched for different levels of business, breaking the wasteful model of traditional 1:1 static resource allocation, significantly reducing the total cost of ownership (TCO) of disaster recovery and improving resource utilization.

[0087] 2. It has enabled the refinement and differentiation of disaster recovery strategies, ensuring the continuity of core businesses.

[0088] 3. Enhanced adaptability and management agility in complex multi-cloud environments.

[0089] 4. The introduction of intelligent optimization algorithms upgrades scheduling decisions from being based on fixed rules to being based on multiple objectives, thereby improving the intelligence and scientific nature of the decisions.

[0090] 5. It achieves full-process automation, reducing the complexity of operation and maintenance.

[0091] like Figure 5 As shown, this embodiment of the invention also provides a disaster recovery resource dynamic scheduling device 900, which can implement the above-described method. This device may include: The first module 901 is used to collect raw data of heterogeneous cloud resources by calling the application programming interfaces of multiple cloud service providers, and to abstract the raw data into a standardized internal resource model to generate a standardized global resource status library. The second module 902 is used to respond to the configuration data of the target object, quantitatively evaluate the business level of different business systems through multi-dimensional disaster recovery requirement indicators, and generate a business requirement list; among them, each business level is bound to a corresponding service level agreement constraint. The third module, 903, is used to generate a list of disaster recovery resource deployment strategies based on a global resource status database and a list of business requirements, with the core optimization objective of minimizing the overall disaster recovery cost. It takes the service level agreement constraints of each business system as constraints and uses a multi-objective optimization algorithm to generate a list of disaster recovery resource deployment strategies. The list of disaster recovery resource deployment strategies includes the optimal disaster recovery resources for each business system. Module 4, 904, is used to parse the disaster recovery resources corresponding to the disaster recovery resource deployment strategy list into API call instructions corresponding to each cloud service provider, so that each cloud service provider can execute the API call instructions to complete the scheduling operation of the disaster recovery resources; among which, the scheduling operation includes creation operation, configuration change operation, and release operation; Module 5, 905, is used to visualize the results of scheduling operations, generate a cost-benefit simulation report, and push it to the target object. If the target object inputs a hypothetical parameter modification instruction, it updates the global resource status database and / or business requirement list in response to the hypothetical parameter modification instruction, and returns to execute the operation of Module 3 to update the cost-benefit simulation report.

[0092] In some embodiments, the apparatus further includes a sixth module for performing the following operations: In response to the progression of time, real-time data from heterogeneous cloud resources is collected periodically by calling the application programming interfaces of multiple cloud service providers; If the parameter changes in real-time data compared to the original data trigger any of the preset monitoring event triggers, the global resource status database will be updated based on the real-time data. Based on the updated global resource status database and the original business requirement list, return to the execution of the third module's operation to update the disaster recovery resource deployment strategy list; Compare the differences between the disaster recovery resource deployment strategy list before and after the update, and generate an incremental change plan; The disaster recovery resources corresponding to the incremental change plan are parsed into incremental API call instructions for each cloud service provider, so that each cloud service provider can execute the incremental API call instructions to complete the incremental dynamic migration of disaster recovery resources.

[0093] It is understood that the content of the above method embodiments is applicable to the present device embodiments. The specific functions implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0094] This invention also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the method described above. This electronic device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0095] It is understood that the content of the above method embodiments is applicable to this device embodiment. The specific functions implemented by this device embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0096] like Figure 6 As shown, Figure 6 The hardware structure of an electronic device 1000 according to another embodiment is illustrated. The electronic device 1000 includes: The processor 1001 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (aSIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present invention. The memory 1002 can be implemented as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RaM). The memory 1002 can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 1002 and is called and executed by the processor 1001. Input / output interface 1003 is used to implement information input and output; The communication interface 1004 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). Bus 1005 transmits information between various components of the device (e.g., processor 1001, memory 1002, input / output interface 1003, and communication interface 1004); The processor 1001, memory 1002, input / output interface 1003 and communication interface 1004 are connected to each other within the device via bus 1005.

[0097] The electronic device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0098] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0099] It is understood that the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0100] This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0101] It is understood that the content of the above method embodiments is applicable to the embodiments of this program product. The specific functions implemented by the embodiments of this program product are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0102] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0103] The disaster recovery resource dynamic scheduling method, device, electronic device, storage medium, and program product provided in this invention collect raw data of heterogeneous cloud resources by calling the application programming interfaces of multiple cloud service providers, abstract the raw data into a standardized internal resource model, and generate a standardized global resource status library; responding to the configuration data of the target object, the business level of different business systems is quantitatively evaluated through multi-dimensional disaster recovery requirement indicators to generate a business requirement list; wherein, each business level is bound to a corresponding service level agreement constraint; based on the global resource status library and the business requirement list, with minimizing the overall disaster recovery cost as the core optimization objective, the service level agreement constraints of each business system are used as constraints, and a disaster recovery resource deployment strategy list is generated through a multi-objective optimization algorithm; The disaster recovery resource deployment strategy list includes the optimal disaster recovery resources for each business system. The disaster recovery resources in the deployment strategy list are parsed into API call instructions for each cloud service provider, enabling each cloud service provider to execute these API calls and complete the scheduling of disaster recovery resources. The scheduling operations include creation, configuration change, and release. The results of the scheduling operations are visualized to generate a cost-benefit simulation report, which is then pushed to the target object. If the target object inputs a hypothetical parameter modification instruction, the global resource status database and / or business requirement list are updated in response to the instruction. The process then returns to execute steps based on the global resource status database and business requirement list, with the core optimization objective of minimizing overall disaster recovery costs, thereby updating the cost-benefit simulation report. This invention, through dynamic scheduling aimed at minimizing total cost, can significantly improve resource utilization and reduce total cost of ownership. Furthermore, by automatically classifying resources based on the multi-dimensional weights of business systems, it can achieve adaptive resource allocation, effectively solving the resource mismatch problem. Moreover, by standardizing heterogeneous cloud resources through unified abstraction, it can make cross-cloud decisions to overcome the management complexity caused by differences in APIs, performance, and billing models among different cloud vendors. In addition, through revision and iteration, this invention can achieve dynamic rescheduling to intelligently respond to dynamic changes such as price fluctuations and performance degradation.

[0104] The embodiments described in this invention are for the purpose of more clearly illustrating the technical solutions of the embodiments of this invention, and do not constitute a limitation on the technical solutions provided by the embodiments of this invention. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this invention are also applicable to similar technical problems.

[0105] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present invention, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0106] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0107] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0108] The preferred embodiments of the present invention have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present invention. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and spirit of the present invention should be within the scope of the claims of the present invention.

Claims

1. A method for dynamic scheduling of disaster recovery resources, characterized in that, The method includes the following steps: By calling the application programming interfaces of multiple cloud service providers, raw data of heterogeneous cloud resources is collected, and the raw data is uniformly abstracted into a standardized internal resource model to generate a standardized global resource status library. In response to the configuration data of the target object, the business level of different business systems is quantitatively evaluated through multi-dimensional disaster recovery requirement indicators to generate a business requirement list; wherein, each of the business levels is bound to a corresponding service level agreement constraint. Based on the global resource status database and the business requirement list, with minimizing the overall disaster recovery cost as the core optimization objective, and taking the service level agreement constraints of each of the business systems as constraints, a disaster recovery resource deployment strategy list is generated through a multi-objective optimization algorithm; wherein, the disaster recovery resource deployment strategy list includes the optimal disaster recovery resources corresponding to each business system; The disaster recovery resources corresponding to the disaster recovery resource deployment strategy list are parsed into API call instructions corresponding to each of the cloud service providers, so that each of the cloud service providers executes the API call instructions to complete the scheduling operation of the disaster recovery resources; wherein, the scheduling operation includes creation operation, configuration change operation and release operation; The results of the scheduling operation are visualized to generate a cost-benefit simulation report, which is then pushed to the target object. If the target object inputs a hypothetical parameter modification instruction, the global resource status database and / or the business requirement list are updated in response to the hypothetical parameter modification instruction. The process then returns to the execution of the step based on the global resource status database and the business requirement list, with the core optimization objective of minimizing the overall disaster recovery cost, in order to update the cost-benefit simulation report.

2. The method according to claim 1, characterized in that, The process of collecting raw data from heterogeneous cloud resources by calling application programming interfaces (APIs) of multiple cloud service providers, abstracting the raw data into a standardized internal resource model, and generating a standardized global resource state library includes the following steps: Raw data from heterogeneous cloud resources is collected in real time or periodically from the application programming interfaces of multiple cloud service providers. The raw data from different cloud service providers are in the format of heterogeneous instances, and the attribute information of the raw data includes resource metadata, performance indicators, price information and resource availability status data. The raw data from different cloud service providers is cleaned and formatted. Then, according to the predefined standardized resource unit mapping rules, the heterogeneous instances from different cloud service providers are uniformly mapped to standard identifiers according to the target core capabilities. The attribute information is associated with the standard identifiers and stored to generate the standardized global resource state library.

3. The method according to claim 1, characterized in that, The configuration data in response to the target object is used to quantitatively assess the business level of different business systems through multi-dimensional disaster recovery requirement indicators, and a business requirement list is generated, including the following steps: The configuration data of the target object is received through the configuration interface; wherein, the configuration data includes the multi-dimensional disaster recovery requirement indicators configured by the target object for each business system, the weight and quantification logic corresponding to each disaster recovery requirement indicator, and the multi-dimensional disaster recovery requirement indicators include recovery time target requirements, recovery point target requirements, business importance / financial impact, compliance requirements and cost budget limit; The system obtains the original value of each disaster recovery requirement indicator from the multi-dimensional disaster recovery requirement indicators, and based on the value range to which the original value belongs, it performs mapping and transformation through the quantification logic to obtain the original score. The original scores corresponding to each disaster recovery requirement indicator are weighted and summed using the weights to obtain a disaster recovery priority score. Finally, the disaster recovery priority score is compared with the preset score range to determine the business level corresponding to each business system. Based on the business level, a preset service level agreement constraint is matched for each business system to generate the business requirement list.

4. The method according to claim 1, characterized in that, Based on the global resource status database and the business requirement list, and with minimizing the overall disaster recovery cost as the core optimization objective, the process involves generating a disaster recovery resource deployment strategy list through a multi-objective optimization algorithm, using the service level agreement constraints of each business system as constraints, and employing the following steps: Based on the service level agreement constraints corresponding to each business system in the business requirement list, all candidate resources in the global resource status database are screened, and candidate resources that do not meet any of the service level agreement constraints are removed to construct a feasible resource pool corresponding to each business system. By using the feasible resource pool to impose resource selection constraints, minimizing the total cost of all business disaster recovery resources is taken as the objective function. A multi-objective optimization algorithm is used to solve for the optimal disaster recovery resource corresponding to each business system, and a list of disaster recovery resource deployment strategies is generated.

5. The method according to claim 1, characterized in that, The step of parsing the disaster recovery resources corresponding to the disaster recovery resource deployment strategy list into API call instructions corresponding to each of the cloud service providers, so that each of the cloud service providers executes the API call instructions to complete the scheduling operation of the disaster recovery resources, includes the following steps: The optimal disaster recovery resource is used as a standardized resource, and the correspondence between each business system and the standardized resource is determined based on the disaster recovery resource deployment strategy list. Based on the API call specification mapping table of each cloud service provider, the correspondence is translated in reverse into the specific API request parameters of the corresponding cloud service provider as the API call instruction; The API call instruction invokes the target API of the corresponding cloud service provider to sequentially execute each atomic task in the preset resource deployment workflow, and collects the execution result of each atomic task to complete the scheduling operation of disaster recovery resources.

6. The method according to claim 1, characterized in that, The process of visualizing the results of the scheduling operation to generate a cost-benefit simulation report and pushing it to the target object, and updating the global resource status database and / or the business requirement list in response to the hypothetical parameter modification instruction if the target object inputs a hypothetical parameter modification instruction, includes the following steps: The results of the scheduling operation are rendered and displayed graphically to generate the cost-benefit simulation report; The cost-benefit simulation report includes a resource topology map, a cost trend curve, and a resource utilization heatmap. If the cost-benefit simulation report has been updated, a difference analysis report is generated based on the comparison of the cost-benefit simulation reports before and after the update. The difference analysis report is annotated in the updated cost-benefit simulation report and pushed to the target object.

7. The method according to claim 1, characterized in that, The method further includes the following steps: In response to the progression of time, real-time data of heterogeneous cloud resources are collected periodically by calling the application programming interfaces of multiple cloud service providers; If the parameter change of the real-time data compared to the original data triggers any of the preset monitoring event triggers, the global resource status database is updated based on the real-time data; Based on the updated global resource status database and the original business requirement list, return to execute the step of minimizing the overall disaster recovery cost as the core optimization objective, based on the global resource status database and the business requirement list, in order to update the disaster recovery resource deployment strategy list; Compare the differences between the disaster recovery resource deployment strategy list before and after the update, and generate an incremental change plan; The disaster recovery resources corresponding to the incremental change plan are parsed into incremental API call instructions corresponding to each of the cloud service providers, so that each of the cloud service providers executes the incremental API call instructions to complete the incremental dynamic migration of disaster recovery resources.

8. A dynamic scheduling device for disaster recovery resources, characterized in that, The apparatus, applicable to the method of any one of claims 1 to 7, comprises: The first module is used to collect raw data of heterogeneous cloud resources by calling the application programming interfaces of multiple cloud service providers, and to abstract the raw data into a standardized internal resource model to generate a standardized global resource status library. The second module is used to respond to the configuration data of the target object, quantitatively evaluate the business level of different business systems through multi-dimensional disaster recovery requirement indicators, and generate a business requirement list; wherein, each of the business levels is bound to a corresponding service level agreement constraint. The third module is used to generate a list of disaster recovery resource deployment strategies based on the global resource status database and the business requirement list, with minimizing the overall disaster recovery cost as the core optimization objective, and taking the service level agreement constraints of each of the business systems as constraints, through a multi-objective optimization algorithm; wherein, the list of disaster recovery resource deployment strategies includes the optimal disaster recovery resources corresponding to each business system; The fourth module is used to parse the disaster recovery resources corresponding to the disaster recovery resource deployment strategy list into API call instructions corresponding to each of the cloud service providers, so that each of the cloud service providers can execute the API call instructions to complete the scheduling operation of the disaster recovery resources; wherein, the scheduling operation includes creation operation, configuration change operation and release operation; The fifth module is used to visualize the results of the scheduling operation, generate a cost-benefit simulation report, and push it to the target object. If the target object inputs a hypothetical parameter modification instruction, the module updates the global resource status database and / or the business requirement list in response to the hypothetical parameter modification instruction, and returns to execute the operation of the third module to update the cost-benefit simulation report.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 7.

Images