Method and device for creating cloud sandbox experiment

Abstract

The invention discloses a method and device for creating a cloud sandbox experiment, and belongs to the technical field of cloud services. The method comprises the steps that a cloud management platform receives an experiment creation request sent by a tenant, provides an input interface of experiment basic information for the tenant based on the experiment creation request, obtains the experiment basic information input by the tenant from the input interface, then obtains experiment content of a cloud sandbox experiment based on the experiment basic information, and sends the experiment content to the tenant; and creating a cloud sandbox experiment for realizing the experiment content by using the infrastructure, and providing an operation interface of the cloud sandbox experiment for the tenant. The experiment content at least comprises an experiment task of the cloud sandbox experiment and an implementation step thereof, and the experiment task and the implementation step thereof are used for guiding the operation mode of the tenant on the cloud sandbox experiment. According to the invention, the automation degree and efficiency of creating the cloud sandbox experiment by the cloud management platform are improved.

Description

Technical Field

[0001] This application relates to the field of cloud service technology, and in particular to a method and apparatus for creating a cloud sandbox experiment. Background Technology

[0002] With the development of cloud computing technology, cloud providers are offering an increasing variety of cloud services. To help users quickly understand and experience cloud services, cloud providers offer cloud sandbox experiments. Through these experiments, users can practice, debug, and verify cloud services in the cloud. Cloud sandbox experiments (also known as sandbox tests) are cloud service-related experiments. Cloud providers offer experimental manuals for these sandbox experiments, allowing users to follow the instructions. The manuals include the experimental objectives, task descriptions, and experimental procedures.

[0003] Currently, cloud sandbox experiments provided by cloud vendors to users are all manually created. For example, before providing a cloud sandbox experiment to a user, the cloud vendor needs to complete the writing of the basic experimental information, experimental tasks and their implementation steps, configure the detection conditions for the experimental tasks, configure the cloud resources required to operate the cloud sandbox experiment, draw the cloud service topology diagram of the cloud sandbox experiment, and configure the desktop image of the cloud sandbox experiment, etc. However, all these operations need to be completed manually.

[0004] This indicates that the current efficiency of creating cloud sandbox experiments is relatively low. Summary of the Invention

[0005] This application provides a method and apparatus for creating cloud sandbox experiments. This application improves the automation and efficiency of creating cloud sandbox experiments on a cloud management platform. The technical solution provided by this application is as follows:

[0006] Firstly, this application provides a method for creating a cloud sandbox experiment. This method is executed by a cloud management platform. The cloud management platform manages the infrastructure that provides cloud services. The infrastructure is used to deploy cloud instances that implement cloud services. The method includes: the cloud management platform receiving an experiment creation request sent by a tenant, the experiment creation request instructing the cloud management platform to create a cloud sandbox experiment for the tenant, the cloud sandbox experiment being used for the tenant to experience cloud services; the cloud management platform providing the tenant with an input interface for basic experiment information based on the experiment creation request; the cloud management platform obtaining the basic experiment information input by the tenant from the input interface, the basic experiment information indicating one or more of the following: the experiment name, experiment objective, experiment content summary, experiment type, or, the cloud services involved in the cloud sandbox experiment; the cloud management platform obtaining the experiment content of the cloud sandbox experiment based on the basic experiment information, the experiment content including at least the experiment task and its implementation steps, the experiment task and its implementation steps being used to guide the tenant in operating the cloud sandbox experiment; and the cloud management platform using the infrastructure to create a cloud sandbox experiment to implement the experiment content, providing the tenant with an operation interface for the cloud sandbox experiment.

[0007] In this way, the cloud management platform can use the method provided in this application to automatically obtain the experimental content of the cloud sandbox experiment based on the basic experimental information of the cloud sandbox experiment provided by the tenant, and use the infrastructure to create cloud sandbox experiments to implement the experimental content. This improves the automation and efficiency of the cloud management platform in creating cloud sandbox experiments, thereby enhancing the learning experience and efficiency of the tenant.

[0008] Furthermore, since cloud sandbox experiments are created based on the basic experimental information provided by the tenant, they can meet the tenant's personalized experimental needs. This enables the cloud management platform to provide cloud sandbox experiments that meet the tenant's personalized experimental needs. It helps the cloud management platform to use this capability to provide personalized cloud sandbox experiments to its many tenants, enriching the types of cloud sandbox experiments that the cloud management platform can provide and improving the flexibility of the cloud sandbox experiments provided by the cloud management platform.

[0009] In one possible implementation, the method further includes: the cloud management platform receiving an entry instruction from a tenant via an operation interface; and displaying the cloud sandbox experiment platform and its contents to the tenant based on the entry instruction. The experiment platform is used by the tenant to operate the cloud sandbox experiment. By displaying the experiment platform to the tenant, the cloud management platform facilitates the tenant's operation of the cloud sandbox experiment. By displaying the experiment contents to the tenant, the cloud management platform provides experimental guidance to the tenant while they perform operations on the cloud sandbox experiment, thereby improving the user experience of using the cloud sandbox experiment.

[0010] In one possible implementation, the experiment content also includes: detection conditions for the experimental task, which indicate the conditions that must be met for the successful completion of the experimental task. During the execution of the cloud sandbox experiment, the cloud management platform can detect whether the tenant has successfully completed the experimental task based on the detection conditions. Therefore, creating a cloud sandbox experiment using infrastructure to implement the experimental content also includes: configuring an application on the cloud host to perform detection and obtain detection results based on the detection conditions.

[0011] Accordingly, the method further includes: the cloud management platform receiving operation instructions sent by the tenant, the operation instructions indicating the operations required to complete the experimental task in the cloud sandbox experiment; the cloud management platform executing the operations according to the operation instructions and generating an operation response; and, upon receiving a detection instruction sent by the tenant, the cloud management platform using detection conditions to detect whether the experimental task has been successfully completed in response, and providing feedback on the detection results to the tenant, whereby the detection instructions indicate the detection of the completion status of the experimental task. By using detection conditions to detect whether the experimental task has been successfully completed in response and providing feedback on the detection results to the tenant, the cloud management platform facilitates the tenant's understanding of whether their operation of the experimental task is correct, contributing to a better user experience when using the cloud sandbox experiment and increasing user stickiness.

[0012] In one possible implementation, the experiment content also includes: the type of cloud resources required for the experimental task and the specification constraints that the cloud resources need to meet. Therefore, creating a cloud sandbox experiment using infrastructure to implement the experimental content also includes: configuring the cloud host's ability to access cloud resources, such as configuring the applications required to access cloud resources.

[0013] Accordingly, the method further includes: the cloud management platform receiving a configuration instruction sent by the tenant, the configuration instruction indicating whether the required cloud resources need to be pre-configured for the experimental task; and, if the configuration instruction indicates that the required cloud resources need to be pre-configured for the experimental task, the cloud management platform configuring the required cloud resources for the experimental task according to the type of cloud resources required by the experimental task as indicated by the experimental content and the specification constraints that the cloud resources need to meet. By pre-configuring the cloud resources required for the cloud sandbox experiment, the cloud resources required for the cloud sandbox experiment can be prepared for the tenant in advance, so that the tenant does not need to configure the target cloud resources when operating the cloud sandbox experiment, thus avoiding the tenant spending too much effort and time on the environment preparation stage.

[0014] In one possible implementation, when the experimental content includes the experimental tasks and implementation steps of the cloud sandbox experiment, the cloud management platform obtains the experimental content of the cloud sandbox experiment based on the basic experimental information, including: the cloud management platform generates one or more experimental tasks of the cloud sandbox experiment based on the basic experimental information; and the cloud management platform generates the implementation steps required to complete each experimental task based on the one or more experimental tasks.

[0015] For example, based on the basic information of a cloud sandbox experiment input by a tenant, the cloud management platform can obtain the functions that the tenant can experience by operating the cloud sandbox experiment. Using cloud knowledge data, the cloud management platform can determine one or more operations required to achieve these functions, and the execution order of these operations. Based on the function of each operation, these operations can be divided into multiple operation categories, each category used to implement an experimental task, thus obtaining the experimental tasks of the cloud sandbox experiment. Based on the execution order of the operations, the implementation order between different experimental tasks can be obtained, as well as the implementation steps required to complete an experimental task using one or more operations. Here, cloud knowledge data is used to indicate the cloud services that the cloud management platform can provide.

[0016] Optionally, the cloud management platform can also configure diagrams for the experimental tasks to help tenants better understand how the experimental tasks are implemented. Based on the basic experimental information, the cloud management platform obtains the experimental content of the cloud sandbox experiment, which also includes: the cloud management platform configuring target diagrams for the experimental tasks, whereby the target diagrams are used to visually represent how the experimental tasks are implemented.

[0017] In one possible implementation, the cloud management platform queries a pre-defined image index library for target images that match the experimental task. If a target image matching the experimental task exists in the image index library, that target image is designated as the image configured for the experimental task, and the target image is configured for the experimental task. For example, each image in the image index library carries an image description, which indicates the content expressed by the image. When querying the image index library for a target image matching the experimental task, the image descriptions of all images in the image index library can be matched with the textual description of the experimental task. If the content expressed by the image description is the task that the experimental task needs to achieve, that image is designated as the target image matching the experimental task, and the image description of that image is designated as the image description of the target image.

[0018] In one possible implementation, the experimental content also includes: detection conditions for the experimental task. Detection conditions are used to indicate the conditions that should be met for the successful completion of the experimental task. Then, based on the basic experimental information, the cloud management platform obtains the experimental content of the cloud sandbox experiment, which also includes: the cloud management platform obtaining the expected state that should be achieved for the successful completion of the experimental task based on the experimental task; the cloud management platform obtaining the target detection type to which the target detection condition used to detect whether the expected state has been reached based on the expected state; the cloud management platform obtaining the target detection condition template based on the target detection type, where the target detection condition template is a template that all detection conditions belonging to the target detection type must satisfy; and the cloud management platform instantiating the parameters in the target detection condition template based on the expected state to obtain the target detection conditions, and using the target detection conditions as the detection conditions for the experimental task.

[0019] In one possible implementation, the cloud management platform can pre-configure detection condition templates for various detection types. After determining the target detection type of the target detection condition to which the desired state needs to be achieved after successfully completing the experimental task belongs, the cloud management platform can match this target detection type with the detection condition templates for various detection types. Once a detection condition template for detecting the same target detection type as the desired state is matched, this matched detection condition template is determined as the target detection condition template for the experimental task. Then, based on the differentiated parameters in the desired state, the parameters in the target detection condition template are instantiated to obtain the target detection conditions.

[0020] In one possible implementation, the experimental content also includes: the type of cloud resources required for the experimental task and the specification constraints that the cloud resources need to meet; the cloud management platform obtains the experimental content of the cloud sandbox experiment based on the basic experimental information; and the cloud management platform obtains the type of cloud resources required to complete the experimental task and the specification constraints that the cloud resources need to meet based on the experimental task; and the type of cloud resources and the specification constraints that the cloud resources need to meet match the experimental task.

[0021] In one possible implementation, the cloud management platform determines the type of cloud resources required to complete each experimental task in the cloud sandbox experiment, as well as the maximum demand for each type of cloud resource. When a certain cloud resource is used only to complete one experimental task in the cloud sandbox experiment, the maximum demand for that cloud resource to complete that task is the maximum specification of that cloud resource allowed to be configured to complete that task. When a certain cloud resource is used to complete multiple experimental tasks in the cloud sandbox experiment, the maximum value among the maximum demands for that cloud resource to complete these multiple experimental tasks is the maximum specification of that cloud resource allowed to be configured to complete each of these multiple experimental tasks. This limitation on the maximum specification of cloud resources allowed to be configured to complete experimental tasks constitutes the specification constraint condition that the cloud resources must meet. Optionally, the cloud management platform may pre-record the type of cloud resources required to complete each experimental task that the cloud management platform can provide, as well as the maximum demand for each type of cloud resource, and this maximum demand can be obtained through simulation experiments.

[0022] In one possible implementation, the method further includes: the cloud management platform matching the topological relationships between cloud services recorded on the cloud management platform based on the cloud services involved in the cloud sandbox experiment to obtain a cloud service topology map of the cloud sandbox experiment, and displaying the cloud service topology map to the tenant. The cloud service topology map is used to indicate the topological relationships between the cloud services involved in the cloud sandbox experiment. For example, when multiple cloud services included in a certain topological relationship recorded by the cloud management platform correspond one-to-one with multiple cloud services required by the cloud sandbox experiment, the cloud management platform determines the cloud service topology map indicating that topological relationship as the cloud service topology map of the cloud sandbox experiment.

[0023] When tenants operate cloud sandbox experiments, the cloud management platform displays a cloud service topology map to them, which makes it easy for tenants to understand the cloud services required for the cloud sandbox experiment and the relationships between cloud services based on the cloud service topology map.

[0024] And / or, the cloud management platform matches the desktop images recorded on the cloud sandbox experiment with the experiment type to obtain the desktop image for the cloud sandbox experiment, and provides the desktop image to the tenant. The desktop image is used to instruct the tenant's client to display the configuration required by the cloud sandbox experiment platform. For example, the cloud management platform has pre-built mappings between various experiment types and their corresponding desktop images. When determining the desktop image for a cloud sandbox experiment, the cloud management platform first determines the experiment type, and then queries the mapping between the experiment type and the desktop image based on that experiment type to obtain the desktop image for the cloud sandbox experiment.

[0025] In one possible implementation, the cloud management platform obtains the experimental content of the cloud sandbox experiment based on the basic experimental information. This includes: the cloud management platform inputting the basic experimental information into the target large model and receiving the experimental content of the cloud sandbox experiment output by the target large model. For example, when the experimental content indicates the experimental task and its implementation steps of the cloud sandbox experiment, after the cloud management platform inputs the basic experimental information into the target large model, the target large model can output the experimental task and its implementation steps of the cloud sandbox experiment, thus the cloud management platform can obtain the experimental task and its implementation steps of the cloud sandbox experiment. When the experimental content indicates the experimental task and its implementation steps of the cloud sandbox experiment, the detection conditions of the experimental task, and the cloud resources required for the experimental task, after the cloud management platform inputs the basic experimental information into the target large model, the target large model can output the experimental task and its implementation steps of the cloud sandbox experiment, the detection conditions of the experimental task, and the cloud resources required for the experimental task, thus the cloud management platform can obtain the experimental content output by the target large model.

[0026] In one possible implementation, the target large model includes a first sub-model, a second sub-model, and a third sub-model. After the cloud management platform inputs basic experimental information into the first sub-model, the first sub-model can output the experimental task and its implementation steps for the cloud sandbox experiment. Then, after the cloud management platform inputs the experimental task and its implementation steps into the second sub-model, the second sub-model can output the detection conditions for the experimental task. After the cloud management platform inputs the experimental task and its implementation steps into the third sub-model, the third sub-model can output the cloud resources required for the experimental task. Alternatively, the first sub-model can be cascaded with the second and third sub-models respectively. This cascading relationship allows the first sub-model to output to the cloud management platform, the second sub-model, and the third sub-model respectively, eliminating the need for the cloud management platform to input the output of the first sub-model into the second and third sub-models.

[0027] The target large model is the initial large model obtained through incremental pre-training and fine-tuning training. Its training process includes the following two training phases:

[0028] The first training phase involves the cloud management platform using cloud knowledge data and historical cloud sandbox experiment content as the initial training data to incrementally pre-train the initial large model. The cloud knowledge data indicates the cloud services that the cloud management platform can provide. The purpose of this first training phase is to enable the trained initial large model to recognize the cloud knowledge data and the experimental content of the cloud sandbox experiments. The experimental content of the historical cloud sandbox experiments is typically obtained manually.

[0029] The second training phase involves the cloud management platform extracting key information from specified experimental content in historical cloud sandbox experiments as secondary training data. This data is then used to fine-tune the initial large-scale model, which has undergone incremental pre-training, to obtain the target large-scale model used to generate the specified experimental content. The purpose of this second training phase is to enable the trained target large-scale model to more accurately generate the required specified experimental content. Incremental pre-training refers to using cloud knowledge data and experimental content from historical cloud sandbox experiments as primary training data to train the initial large-scale model. Cloud knowledge data indicates the cloud services that the cloud management platform can provide. Fine-tuning training involves using key information extracted from specified experimental content from historical cloud sandbox experiments as secondary training data to train the initial large-scale model, which has undergone incremental pre-training.

[0030] In one possible implementation, the second training data includes: a first instruction fine-tuning dataset, which comprises multiple first input-output data pairs. Each first input-output data pair includes: basic information about historical experiments in the historical cloud sandbox experiment, historical experimental tasks and their implementation steps, and attached figures and their descriptions configured for the historical experimental tasks. The basic information about historical experiments serves as the input for training the initial large model sub-model, while the historical experimental tasks and their implementation steps, and the attached figures and their descriptions configured for the historical experimental tasks serve as the output for training the initial large model sub-model.

[0031] In one possible implementation, the second training data further includes: a second instruction fine-tuning dataset and / or a third instruction fine-tuning dataset; the second instruction fine-tuning dataset includes multiple second input-output data pairs, each second input-output data pair including: historical experimental tasks and their experimental steps and detection conditions from historical cloud sandbox experiments, wherein the historical experimental tasks and their experimental steps are the inputs used to train the initial large model sub-model, and the detection conditions of the historical experimental tasks are the outputs used to train the initial large model sub-model; the third instruction fine-tuning dataset includes multiple third input-output data pairs, each third input-output data pair including: historical experimental tasks and their implementation steps from historical cloud sandbox experiments, the cloud services used, and the specifications of the cloud services. The historical experimental tasks and their experimental steps are the inputs used to train the initial large model sub-model, and the cloud services used in the historical experimental tasks and the specifications of the cloud services are the outputs used to train the initial large model sub-model.

[0032] The target large model includes a first sub-model, a second sub-model, and a third sub-model. The first instruction fine-tuning dataset can be used to train the first sub-model, the second instruction fine-tuning dataset can be used to train the second sub-model, and the third instruction fine-tuning dataset can be used to train the third sub-model.

[0033] Secondly, this application provides an apparatus for creating cloud sandbox experiments. The apparatus is deployed on a cloud management platform. The cloud management platform manages the infrastructure providing cloud services. The infrastructure is used to deploy cloud instances that implement cloud services. The apparatus includes: an interaction module for receiving an experiment creation request sent by a tenant, the experiment creation request instructing the cloud management platform to create a cloud sandbox experiment for the tenant, the cloud sandbox experiment being used for the tenant to experience cloud services; the interaction module is also used to provide the tenant with an input interface for basic experiment information based on the experiment creation request; the interaction module is also used to obtain the basic experiment information input by the tenant from the input interface, the basic experiment information indicating one or more of the following: the experiment name, experiment objective, experiment content summary, experiment type, or, the cloud services involved in the cloud sandbox experiment; a processing module for obtaining the experiment content of the cloud sandbox experiment based on the basic experiment information, the experiment content including at least the experiment task and its implementation steps, the experiment task and its implementation steps being used to guide the tenant in operating the cloud sandbox experiment; a deployment module for using the infrastructure to create a cloud sandbox experiment to implement the experiment content; and the interaction module is also used to provide the tenant with an operation interface for the cloud sandbox experiment.

[0034] In one possible implementation, the interaction module is also used to receive an entry experiment instruction sent by the tenant based on the operation interface, and to display the experimental platform and experimental content of the cloud sandbox experiment to the tenant based on the entry experiment instruction. The experimental platform is used for the tenant to operate the cloud sandbox experiment.

[0035] In one possible implementation, the experiment content further includes: detection conditions for the experimental task, which indicate the conditions that must be met for the successful completion of the experimental task. The interaction module is further configured to receive operation instructions sent by the tenant, which instruct the operations required to complete the experimental task in the cloud sandbox experiment; the processing module is further configured to execute the operations according to the operation instructions for the cloud sandbox experiment and generate a response to the operations; the processing module is further configured to, upon receiving a detection instruction from the tenant in the interaction module, use the detection conditions to detect whether the experimental task has been successfully completed in response, where the detection instruction instructs the detection of the completion status of the experimental task; the interaction module is further configured to provide feedback on the detection results to the tenant.

[0036] In one possible implementation, the experiment content also includes: the type of cloud resources required for the experiment task and the specification constraints that the cloud resources need to meet. The interaction module is further configured to receive configuration instructions sent by the tenant, indicating whether the required cloud resources need to be pre-configured for the experiment task; the processing module is further configured to, if the configuration instructions indicate that the required cloud resources need to be pre-configured for the experiment task, configure the required cloud resources for the experiment task according to the type of cloud resources required for the experiment task and the specification constraints that the cloud resources need to meet as indicated in the experiment content.

[0037] In one possible implementation, the processing module is specifically used to: generate one or more experimental tasks for the cloud sandbox experiment based on the basic experimental information; generate the implementation steps required to complete each experimental task based on the one or more experimental tasks; and configure target illustrations for the experimental tasks based on the experimental tasks, the target illustrations being used to illustrate the implementation of the experimental tasks in the form of images.

[0038] In one possible implementation, the experiment content also includes: detection conditions for the experimental task. Detection conditions are used to indicate the conditions that should be met for the successful completion of the experimental task. The processing module is specifically used for: based on the experimental task, obtaining the expected state that should be achieved for the successful completion of the experimental task; based on the expected state, obtaining the target detection type to which the target detection condition used to detect whether the expected state has been reached belongs; based on the target detection type, obtaining the target detection condition template, which is a template that all detection conditions belonging to the target detection type must satisfy; based on the expected state, instantiating the parameters in the target detection condition template to obtain the target detection conditions, and using the target detection conditions as the detection conditions for the experimental task.

[0039] In one possible implementation, the experiment content also includes: the type of cloud resources required for the experiment task and the specification constraints that the cloud resources need to meet; and a processing module specifically used to: based on the experiment task, obtain the type of cloud resources required to complete the experiment task and the specification constraints that the cloud resources need to meet, and match the type of cloud resources and the specification constraints that the cloud resources need to meet with the experiment task.

[0040] In one possible implementation, the processing module is further configured to match the topological relationships between cloud services recorded on the cloud management platform based on the cloud services involved in the cloud sandbox experiment, to obtain a cloud service topology map of the cloud sandbox experiment. The cloud service topology map is used to indicate the topological relationships between the cloud services involved in the cloud sandbox experiment. The interaction module is further configured to display the cloud service topology map to the tenant.

[0041] And / or, the processing module is also used to match the desktop image recorded by the cloud management platform based on the experiment type of the cloud sandbox experiment to obtain the desktop image of the cloud sandbox experiment. The desktop image is used to instruct the tenant's client to display the configuration required by the experiment platform of the cloud sandbox experiment. The interaction module is also used to provide the desktop image to the tenant.

[0042] In one possible implementation, the processing module is specifically used to: input basic experimental information into the target large model, and receive the experimental content of the cloud sandbox experiment output by the target large model. The target large model is an initial large model obtained through incremental pre-training and fine-tuning training. Incremental pre-training refers to using cloud knowledge data and the experimental content of historical cloud sandbox experiments as the first training data to train the initial large model. The cloud knowledge data is used to indicate the cloud services that the cloud management platform can provide. Fine-tuning training refers to using key information extracted from the specified experimental content of historical cloud sandbox experiments as the second training data to train the initial large model that has undergone incremental pre-training.

[0043] In one possible implementation, the second training data includes: a first instruction fine-tuning dataset, which comprises multiple first input-output data pairs. Each first input-output data pair includes: basic information about historical experiments in the historical cloud sandbox experiment, historical experimental tasks and their implementation steps, and attached figures and their descriptions configured for the historical experimental tasks. The basic information about historical experiments serves as the input for training the initial large model sub-model, while the historical experimental tasks and their implementation steps, and the attached figures and their descriptions configured for the historical experimental tasks serve as the output for training the initial large model sub-model.

[0044] In one possible implementation, the second training data further includes: a second instruction fine-tuning dataset and / or a third instruction fine-tuning dataset. The second instruction fine-tuning dataset includes multiple second input-output data pairs, each pair including: historical experimental tasks from historical cloud sandbox experiments, their experimental steps, and detection conditions. The historical experimental tasks and their experimental steps serve as input for training the initial large model sub-model, and the detection conditions of the historical experimental tasks serve as output for training the initial large model sub-model. The third instruction fine-tuning dataset includes multiple third input-output data pairs, each pair including: historical experimental tasks from historical cloud sandbox experiments, their implementation steps, the cloud services used, and the specifications of the cloud services. The historical experimental tasks and their experimental steps serve as input for training the initial large model sub-model, and the cloud services used and their specifications serve as output for training the initial large model sub-model.

[0045] Thirdly, this application provides a computing device including a memory and a processor, the memory storing program instructions, and the processor executing the program instructions to perform the methods provided in the first aspect of this application and any possible implementation thereof.

[0046] Fourthly, this application provides a computing device cluster, including multiple computing devices, each computing device including multiple processors and multiple memories, the multiple memories storing program instructions, and the multiple processors executing the program instructions, causing the computing device cluster to perform the methods provided in the first aspect of this application and any possible implementation thereof.

[0047] Fifthly, this application provides a computer-readable storage medium that is a non-volatile computer-readable storage medium, which includes program instructions that, when executed on a computing device, cause the computing device to perform the methods provided in the first aspect of this application and any of its possible implementations.

[0048] Sixthly, this application provides a computer program product containing instructions that, when run on a computer, cause the computer to perform the methods provided in the first aspect of this application and any possible implementation thereof. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the interface of a cloud sandbox experiment provided in an embodiment of this application;

[0050] Figure 2 This is a schematic diagram of the implementation scenario of a method for creating a cloud sandbox experiment provided in this application embodiment;

[0051] Figure 3 This is a schematic diagram illustrating the deployment of basic resources in a data center, provided in an embodiment of this application.

[0052] Figure 4 This is a flowchart illustrating a method for creating a cloud sandbox experiment, as provided in an embodiment of this application.

[0053] Figure 5 This is a schematic diagram of a user interface provided by a cloud management platform according to an embodiment of this application, which allows tenants to send experiment creation requests;

[0054] Figure 6 This is a schematic diagram of a user interface provided by a cloud management platform according to an embodiment of this application, which allows tenants to input basic experimental information for cloud sandbox experiments;

[0055] Figure 7 This is a schematic diagram of an experimental manual for a cloud sandbox experiment provided in an embodiment of this application;

[0056] Figure 8 This is a schematic diagram of an experiment provided in an embodiment of this application;

[0057] Figure 9This is a schematic diagram illustrating the process of generating experimental tasks and their implementation steps provided in an embodiment of this application;

[0058] Figure 10 This is a schematic diagram illustrating another process for generating experimental tasks and its implementation steps provided in an embodiment of this application;

[0059] Figure 11 This is a schematic diagram illustrating a process for generating detection conditions for an experimental task, as provided in an embodiment of this application.

[0060] Figure 12 This is a schematic diagram illustrating a process for generating cloud resources required for an experimental task, as provided in an embodiment of this application.

[0061] Figure 13 This is a schematic diagram illustrating another process for generating cloud resources required for experimental tasks, provided in an embodiment of this application.

[0062] Figure 14 This is a schematic diagram illustrating how a large model is used to obtain experimental content, as provided in an embodiment of this application.

[0063] Figure 15 This is a schematic diagram illustrating the training of a first sub-model using a dataset fine-tuned by a first instruction, as provided in an embodiment of this application.

[0064] Figure 16 This is a schematic diagram illustrating a method for logging into a cloud account, as provided in an embodiment of this application.

[0065] Figure 17 This is a schematic diagram illustrating the creation of an OBS bucket according to an embodiment of this application;

[0066] Figure 18 This is a schematic diagram illustrating an option to create a file, as provided in an embodiment of this application.

[0067] Figure 19 This is a schematic diagram illustrating how to name a created file according to an embodiment of this application;

[0068] Figure 20 This is a schematic diagram of an example of uploading a file provided in this application embodiment;

[0069] Figure 21 This is a schematic diagram illustrating a file download method provided in an embodiment of this application;

[0070] Figure 22 This is a schematic diagram illustrating the use of a second instruction to fine-tune a dataset for training a second sub-model, as provided in an embodiment of this application.

[0071] Figure 23 This is a schematic diagram illustrating the use of a third instruction to fine-tune a dataset for training a third sub-model, as provided in an embodiment of this application.

[0072] Figure 24 This is a schematic diagram of a VPC creation interface provided in an embodiment of this application;

[0073] Figure 25 This is a schematic diagram of a VPC basic information configuration interface provided in an embodiment of this application;

[0074] Figure 26 This is a schematic diagram of an interface for purchasing Elastic Cloud Server (ECS) provided in an embodiment of this application;

[0075] Figure 27 This is a schematic diagram of a network configuration interface for an Elastic Cloud Server (ECS) provided in an embodiment of this application;

[0076] Figure 28 This is a schematic diagram of an advanced configuration interface for an Elastic Cloud Server (ECS) provided in an embodiment of this application.

[0077] Figure 29 This is a schematic diagram of a cloud hard drive EVS purchase interface provided in an embodiment of this application;

[0078] Figure 30 This is a schematic diagram of a configuration interface for basic information of the cloud disk EVS provided in an embodiment of this application;

[0079] Figure 31 This is a schematic diagram of an interface for mounting a cloud disk (EVS) according to an embodiment of this application;

[0080] Figure 32 This is a schematic diagram of the structure of a device for creating a cloud sandbox experiment provided in an embodiment of this application;

[0081] Figure 33 This is a schematic diagram of the structure of a computing device provided in an embodiment of this application;

[0082] Figure 34 This is a schematic diagram of the structure of a computing device cluster provided in an embodiment of this application;

[0083] Figure 35 This is a schematic diagram of another computing device cluster structure provided in an embodiment of this application. Detailed Implementation

[0084] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0085] To facilitate understanding, the technologies and background involved in the embodiments of this application will be introduced below.

[0086] Cloud computing is a type of distributed computing that refers to a network that centrally manages and schedules a large number of computing and storage resources to provide on-demand services to users. These computing and storage resources are provided through clusters of computing devices located in data centers. Furthermore, cloud computing can provide users with various types of services, such as Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS). Infrastructure as a Service provides virtual machines or other resources as a service to tenants. Platform as a Service provides a development platform as a service to tenants. Software as a Service provides applications (Apps) as a service to customers.

[0087] A physical machine (PM) is the physical resource used to host virtualization technology. It is also called a physical server. Typically, a physical machine is used to deploy virtual instances. A physical machine has multiple physical devices. For example, a physical server has physical devices such as processors and memory. Multiple virtual instances can be deployed on a single physical machine, sharing the machine's physical resources. Depending on the use case, multiple virtual instances deployed on a single physical machine can belong to the same tenant or to different tenants.

[0088] Virtualization is a resource management technology. Virtualization abstracts and transforms various physical resources of a host, such as computing, network, and storage resources, breaking down the indivisible barriers between the host's physical structures. This allows tenants to utilize these resources in a better way than the original configuration. Resources obtained through virtualization are called virtualized resources, and virtualized resources are not limited by the existing physical resource deployment methods, geographical location, or physical configuration.

[0089] Virtualized resources are typically provided to tenants in the form of virtual instances. Virtual instances utilize the host's hardware resources and run on the host's operating system. Applications run within virtual instances to implement the tenant's business logic. The host's hardware resources can be allocated to one or more tenants at the virtual instance level. Different virtual instances are isolated from each other, allowing tenants to use physical resources conveniently and flexibly while maintaining security and isolation, significantly improving the utilization of physical resources. Typically, virtual instances can be virtual machines, containers, or independent processes (such as functions). Virtual instances are also known as Elastic Compute Service (ECS, also called cloud servers) or Elastic Instances (different cloud service providers may use different names).

[0090] A virtual machine (VM) is a complete computer system with full hardware system functionality, simulated using virtualization technology and running in a completely isolated environment. A subset of the instructions in a VM can be processed on the host machine, while other instructions can be executed in a simulated manner. A VM is also called a virtual server. A VM can be viewed as a collection of virtual devices, which possess full hardware system functionality and run in a completely isolated environment. Virtual devices are created by virtualizing physical devices that can share resources. For example, a virtual processor, created by virtualizing a processor, is a virtual device. Similarly, a training card, created by virtualizing a field-programmable gate array (FPGA), is also a virtual device. For instance, the VM in this application can be a kernel-based virtual machine (KVM). Any task that can be performed on a server can also be performed in a VM. When creating a virtual machine on a server, a portion of the physical machine's hard drive and memory capacity is used as the virtual machine's hard drive and memory capacity. Each virtual machine has its own independent hard drive and operating system, and virtual machine tenants can operate the virtual machine as if it were a server. The runtime environments (such as virtual machine applications, operating systems, and virtual hardware) in different virtual machines are completely isolated, and communication between different virtual machines requires the virtual machine manager to forward network packets.

[0091] Containers utilize the namespace and cgroup technologies supported by the Linux kernel to isolate application processes and their dependencies (the runtime environment's bins / libs, specifically all files required to run the application) within an independent runtime environment. Containers provide a lightweight virtual runtime environment. Containers are created by packaging all the code, libraries, and dependencies of a tenant's application into an image. When the image is executed, it runs in a virtual runtime environment. At this point, the container is a runtime instance of the image, similar to a lightweight sandbox, which can be started, stopped, and deleted. The infrastructure for containers can be server hardware or virtual machines in the cloud (i.e., containers can also be deployed within virtual machines). The operating system uses the Linux kernel and supports namespaces and cgroups. Namespaces are used to isolate processes, while cgroups are used to allocate process resources, specifically virtual processors and memory allocated to the process. The container engine, similar to a virtual machine manager, runs within the operating system and is used to manage containers. Compared to virtual machines, which come with their own operating system, containers do not have an operating system. Instead, containers run as processes within the host machine's operating system. As a result, containers start up faster than virtual machines, making them particularly suitable for lightweight applications. Furthermore, a single host machine can run thousands of containers (processes) simultaneously.

[0092] Cloud vendors can provide users with cloud sandbox experiments. By operating the cloud sandbox experiments, users can practice, debug, and verify cloud services in the cloud, allowing them to quickly understand and experience cloud services. Cloud sandbox experiments (sandbox experiments) are experiments related to cloud services. Figure 1 This is a schematic diagram of the interface for a cloud sandbox experiment provided in an embodiment of this application. For example... Figure 1 As shown, the interface mainly presents three parts: the user account login interface in the upper left corner, the experiment manual in the lower left corner, and the experiment operation interface in the right corner.

[0093] The experiment manual includes basic information such as the experiment name and objectives, an introduction to the experiment task, experimental steps, detection conditions, and required cloud services. The detection conditions indicate the conditions that must be met for successful completion of the experiment. During the cloud sandbox experiment, the cloud management platform can check whether the tenant has successfully completed the experiment task based on these detection conditions. The experiment manual is displayed in the interface to provide guidance, enabling users to operate the cloud sandbox experiment according to the instructions. The experiment operation interface displays a cloud service topology diagram of the cloud sandbox experiment. This diagram indicates the topological relationships between the cloud services used in the experiment. Cloud service topology diagrams are typically scalable vector graphics (SVG). Figure 1 As shown, the cloud service topology diagram indicates that the cloud services required for the cloud sandbox experiment include cloud container engine (CCE), elastic load balancer (ELB), virtual private cloud (VPC), and elastic IP address (EIP). The cloud service topology diagram also shows the topological relationships between these cloud services.

[0094] Currently, cloud sandbox experiments provided by cloud vendors to tenants are all manually created. For example, before providing a cloud sandbox experiment, cloud vendors need to complete the writing of basic experimental information, experimental tasks and their implementation steps, configure detection conditions for the experimental tasks, configure the cloud resources required to operate the cloud sandbox experiment, draw the cloud service topology diagram of the cloud sandbox experiment, and configure the desktop image of the cloud sandbox experiment. All these operations need to be completed manually. The desktop image is used to instruct the tenant's client to display the configuration required for the cloud sandbox experiment platform. For example, the desktop image refers to the experimental environment upon which the experimental desktop displayed on the tenant's client depends when the tenant operates the cloud sandbox experiment.

[0095] This indicates that the current efficiency of creating cloud sandbox experiments is relatively low. Furthermore, the number and types of cloud sandbox experiments currently available are limited, the content is rigid, and they cannot flexibly adapt to the personalized learning needs of tenants.

[0096] In view of this, embodiments of this application provide a method for creating a cloud sandbox experiment. This method is executed by a cloud management platform, which manages the infrastructure providing cloud services. The infrastructure is used to deploy cloud instances that implement cloud services. The method includes: the cloud management platform receiving an experiment creation request sent by a tenant, the experiment creation request instructing the cloud management platform to create a cloud sandbox experiment for the tenant; based on the experiment creation request, the cloud management platform provides the tenant with an input interface for basic experiment information, obtains the basic experiment information input by the tenant from the input interface, then obtains the experiment content of the cloud sandbox experiment based on the basic experiment information, creates a cloud sandbox experiment using the infrastructure to implement the experiment content, and provides the tenant with an operation interface for the cloud sandbox experiment. The basic experiment information indicates one or more of the following: the experiment name, experiment objective, experiment content summary, experiment type, or the cloud services involved in the cloud sandbox experiment. The experiment content includes at least the experiment task and its implementation steps. The experiment task and its implementation steps are used to guide the tenant in how to operate the cloud sandbox experiment.

[0097] In this way, the cloud management platform can use the method provided in this application to automatically obtain the experimental content of the cloud sandbox experiment based on the basic experimental information of the cloud sandbox experiment provided by the tenant, and use the infrastructure to create cloud sandbox experiments to implement the experimental content. This improves the automation and efficiency of the cloud management platform in creating cloud sandbox experiments, thereby enhancing the learning experience and efficiency of the tenant.

[0098] Furthermore, since cloud sandbox experiments are created based on the basic experimental information provided by the tenant, they can meet the tenant's personalized experimental needs. This enables the cloud management platform to provide cloud sandbox experiments that meet the tenant's personalized experimental needs. It helps the cloud management platform to use this capability to provide personalized cloud sandbox experiments to its many tenants, enriching the types of cloud sandbox experiments that the cloud management platform can provide and improving the flexibility of the cloud sandbox experiments provided by the cloud management platform.

[0099] This article provides a detailed introduction to the technical solution of this application from multiple perspectives, including implementation scenarios, methods and processes, hardware devices, and software devices.

[0100] The following are examples illustrating the implementation scenarios of the embodiments of this application.

[0101] Figure 2 This is a structural diagram illustrating an implementation scenario of a method for creating a cloud sandbox experiment provided in this application. For example... Figure 2As shown, the implementation scenario includes: data center 1 and client 2. Data center 1 and client 2 can establish a communication connection via a network. Optionally, this network can be the Internet, or other networks; this embodiment is not limited to any particular network. Tenants can interact with data center 1 through client 2. For example, a tenant can send cloud service requests and other information to data center 1 through client 2. Data center 1 responds based on the information sent by client 2.

[0102] Data Center 1 houses a large amount of infrastructure owned by the cloud service provider, such as computing resources, storage resources, and network resources. For example, computing resources can be computing devices (such as servers) capable of providing computing power. Figure 2 As shown, data center 1 includes a cloud management platform and infrastructure ( Figure 2 (Not shown in the diagram). The cloud management platform and the infrastructure are connected via an internal data center network. The cloud management platform is used to manage the infrastructure. The infrastructure is used to provide public cloud services. The infrastructure includes multiple servers. Cloud services are optionally deployed on the servers. Cloud services are implemented by running virtual instances, and are therefore also referred to as virtual instances deployed on servers to implement tenant services. Tenants can send cloud service requests and related information to the server through their client 2. The server can process the cloud service requests and related information and provide cloud services to the tenant based on the processed cloud service requests and related information. For example, the server can use the method for creating a cloud sandbox experiment provided in this application embodiment to obtain the experimental content of the cloud sandbox experiment based on the basic experimental information of the cloud sandbox experiment provided by the tenant, use the infrastructure to create a cloud sandbox experiment to implement the experimental content, and provide the tenant with an operation interface for the cloud sandbox experiment so that the tenant can operate the cloud sandbox experiment based on the operation interface.

[0103] Cloud management platform functions: It provides access interfaces (such as interfaces or APIs), allowing tenants to remotely access the cloud management platform through client devices. Tenants can register a cloud account and password on the cloud management platform and log in. After successful authentication of the cloud account and password, tenants can further select and purchase virtual machines with specific specifications (processor, memory, disk) on the cloud management platform. After successful purchase, the cloud management platform provides the remote login account and password for the purchased virtual machine. Clients can remotely log in to the virtual machine and install and run the tenant's applications on it.

[0104] The cloud management platform can be logically divided into: tenant console, compute management service, network management service, storage management service, authentication service, and image management service. The tenant console provides a user interface or application programming interface (API) for interaction with tenants. The compute management service manages servers running virtual instances and bare metal servers. The network management service manages network services (such as gateways and firewalls). The storage management service manages storage services (such as data bucket services). The authentication service manages tenant accounts and passwords. The image management service manages virtual instance images.

[0105] exist Figure 2 In the illustrated implementation scenario, a data center contains multiple servers. The servers consist of a hardware layer and a software layer. The hardware layer comprises the standard server configuration, including processors, memory, network interface cards (NICs), disks, and buses. The software layer includes the operating system installed and running on the server. This operating system, relative to the virtual machine, can be called the host operating system. The host operating system runs a virtual machine manager (also known as a hypervisor). The hypervisor's role is to implement compute virtualization, network virtualization, and storage virtualization, and to manage the virtual machines.

[0106] The virtual machine manager runs a cloud management platform client. This client receives control plane commands from the cloud management platform, creates virtual instances on the server based on these commands, and manages the virtual instances throughout their lifecycle. For example, the client can monitor the hardware resource usage of the server in real time and report it to the cloud management platform. When the cloud management platform confirms that a virtual instance needs to be created on a specific server, it sends a virtual instance creation command to the client on that server. Upon receiving the command, the client creates the virtual instance on that server. In this way, tenants can create, manage, log in to, and operate virtual instances within the data center through the cloud management platform.

[0107] Servers can run virtual machines of different specifications. Virtual machine specifications are categorized as: general-purpose computing, memory-optimized, ultra-large memory, etc., with specific specifications under each type. After a tenant selects a virtual machine specification, the cloud management platform selects a server in the data center that supports that specification and ensures sufficient idle hardware resources on that server. Then, it creates and configures the virtual machine with that specification on that server. Configuring servers through the cloud management platform allows for the analysis and planning of server hardware resources. Based on the server's hardware performance, it plans the corresponding computing products for the physical hardware, such as planning virtual machines of different specifications, to meet the diverse needs of different tenants. Furthermore, differentiated pricing strategies can be implemented based on the performance differences of virtual machines of different specifications. For example, high-performance virtual instances can be sold at a higher price, while ordinary performance virtual instances can be sold at a lower price, allowing tenants to purchase virtual instances as needed.

[0108] In one implementation, such as Figure 3 As shown, the location of basic resources in a data center can be described using cloud resource deployment regions (regions) and availability zones (AZs). Tenants can choose to deploy cloud services based on resources within a specific region or AZ. A region is defined by geographical location and network latency. Using the same resource pool within the same region can be understood as sharing common services such as elastic computing, block storage, object storage, virtual private cloud (VPC) networks, elastic internet protocol (EIP) addresses, and images. Regions are divided into general regions and dedicated regions. A general region refers to a region that provides general cloud services to public tenants. A dedicated region refers to a dedicated region that carries the same type of business or provides business services to specific tenants. A region typically includes multiple AZs. Multiple AZs within a region are connected via high-speed fiber optic cables to meet the needs of tenants building high-availability systems across AZs. An AZ is one or more... Figure 3 The data center shown is a collection of data centers. Within an Availability Zone (AZ), computing, networking, and storage resources are logically divided into multiple clusters.

[0109] Tenants can send instructions to the cloud management platform through their client 2 to create, manage, log in to, and operate virtual instances on the server, and use the cloud services provided by these virtual instances. For example, the cloud management platform can provide an access interface. This interface can be provided either as a user interface or an API. Tenants can operate their client to remotely access the access interface to register a cloud account and password on the cloud management platform, and then log in using these accounts and passwords. The cloud management platform can also authenticate the cloud account and password. After successful authentication, the tenant can further select and purchase a virtual instance with specific specifications (processor, memory, disk) on the cloud management platform. After the tenant successfully purchases the virtual instance, the cloud management platform provides the tenant with a remote login account and password for the purchased virtual instance. The tenant can use the remote login account and password to remotely log in to the virtual instance on their client, install and run their application within the virtual instance, and use the application to implement their business operations.

[0110] Client 2 can be selected from computers, personal computers, laptops, mobile phones, smartphones, tablets, cloud servers, portable mobile terminals, multimedia players, e-book readers, wearable devices, smart home appliances, artificial intelligence devices, smart wearable devices, smart in-vehicle devices, or Internet of Things devices, etc.

[0111] In one implementation, the method for creating a cloud sandbox experiment provided in this application embodiment can be implemented by running an executable program on a computing device in data center 1. Optionally, the method for creating a cloud sandbox experiment provided in this application embodiment can be applied to a system for creating cloud sandbox experiments. This system for creating cloud sandbox experiments is deployed on a server managed by a cloud management platform. By running the executable program of the method for creating cloud sandbox experiments provided in this application embodiment, the system for creating cloud sandbox experiments can implement the method for creating cloud sandbox experiments provided in this application embodiment. Furthermore, the executable program for implementing the method for creating cloud sandbox experiments can optionally be presented in the form of an application installation package. After the server installs the application installation package, it can implement the method for creating cloud sandbox experiments provided in this application embodiment by running the executable program therein.

[0112] It should be understood that the above content is an exemplary description of the implementation scenarios of the method for creating a cloud sandbox experiment provided in the embodiments of this application, and does not constitute a limitation on the implementation scenarios of the method for creating a cloud sandbox experiment. Those skilled in the art will know that, as business needs change, the implementation scenarios can be adjusted according to application requirements, and the embodiments of this application do not specifically limit them. Furthermore, when the method for creating a cloud sandbox experiment provided in the embodiments of this application is applied to other scenarios, the executable program of the method can also be presented in the form of an application installation package or in other ways, and the embodiments of this application do not list them all.

[0113] The following uses the method for creating a cloud sandbox experiment provided in the embodiments of this application as an example applied to a cloud management platform to illustrate the implementation process of this method. Figure 4 This is a flowchart illustrating a method for creating a cloud sandbox experiment provided in an embodiment of this application. For example... Figure 4 As shown, the method for creating a cloud sandbox experiment includes the following steps:

[0114] Step 401: The cloud management platform receives the experiment creation request sent by the tenant. The experiment creation request is used to instruct the cloud management platform to create a cloud sandbox experiment for the tenant. The cloud sandbox experiment is used for the tenant to experience cloud services.

[0115] When a tenant needs to create a cloud sandbox experiment using the cloud management platform, they can perform a specified operation on their client to trigger an experiment creation request. The cloud management platform then creates the cloud sandbox experiment for the tenant based on this request. In one possible implementation, the cloud management platform can provide an interactive interface to the tenant, which the tenant can use to trigger the experiment creation request. After the tenant triggers the request, the cloud management platform can obtain the request through this interface. Optionally, the interactive interface can include one or more of the following implementations: an application programming interface (API) and a user interface (UI). When the interactive interface is implemented through a user interface, the tenant can operate within the user interface to indicate the functions they need to implement. The cloud management platform can obtain the information sent by the tenant from the user interface and determine the functions the tenant needs to implement based on that information.

[0116] For example, such as Figure 5 As shown, when a tenant views the description of a cloud service on the official website of a cloud service provider, the description page also displays an icon for a "Cloud Sandbox Experiment Assistant." This icon is used to prompt the tenant to use the cloud service's cloud sandbox experiment. When the tenant needs to use the cloud service's cloud sandbox experiment, they can click on it. Figure 5 The "Cloud Sandbox Experiment Assistant" icon in the system triggers an experiment creation request when clicked. The cloud management platform receives this request and determines, based on it, whether the tenant needs to use the current cloud service's cloud sandbox experiment to experience the cloud service. The interaction is implemented through a user interface such as an explanatory page.

[0117] Step 402: Based on the experiment creation request, the cloud management platform provides the tenant with an input interface for basic experiment information. The basic experiment information indicates one or more of the following: the name of the cloud sandbox experiment, the experiment objective, a summary of the experiment content, the experiment type, or the cloud services involved in the cloud sandbox experiment.

[0118] After receiving an experiment creation request from a tenant, the cloud management platform provides an input interface for basic experiment information. This allows the tenant to input the basic information of the cloud sandbox experiment to be created into the cloud management platform. In one possible implementation, the input interface can be implemented through an application programming interface (API) and a user interface. For example, when the input interface is implemented through a user interface, the cloud management platform, upon receiving the experiment creation request, displays the request to the tenant. Figure 6 The user interface shown prompts the user to input basic experimental information such as experimental objectives, a summary of experimental content, experimental type, and cloud services required for the cloud sandbox experiment.

[0119] The experiment name indicates the name of the cloud sandbox experiment. The experiment objective indicates the purpose that the tenant can achieve by operating the cloud sandbox experiment. For example, Figure 7 This is an example of an experimental manual for a cloud sandbox experiment provided in this application embodiment. For example... Figure 7 As shown, the objective of this cloud sandbox experiment is to help students quickly get started with the Object Storage Service (OBS). The experiment summary, also known as the experiment introduction or overview, provides a general description of the cloud sandbox experiment's content. For example... Figure 7 As shown, the summary of the cloud sandbox experiment outlines the experiment content as follows: creating an OBS bucket, uploading and downloading test data. The experiment type, also known as the experiment direction, indicates the type of cloud sandbox experiment. For example, the experiment type indicates whether the cloud sandbox experiment is an experiential experiment, a hands-on experiment, or a training experiment. The cloud services required for the cloud sandbox experiment indicate the cloud services needed to operate the cloud sandbox experiment. For example, Figure 7 The cloud sandbox experiment shown requires OBS as the cloud service. It should be noted that the content indicated in the experiment manual may be appropriately increased or decreased depending on the application requirements, and the content indicated in the experiment manual is not limited to the examples above. This application embodiment does not provide examples of each of these.

[0120] Step 403: The cloud management platform obtains the basic experimental information input by the tenant from the input interface.

[0121] After a tenant inputs basic experimental information through the input interface, the cloud management platform can retrieve this information from the interface. For example, if the tenant... Figure 6After entering basic experiment information in the user interface shown, the tenant can submit the information to the cloud management platform by clicking the "Create Experiment" button. The cloud management platform can then retrieve this information. The functions requested by the tenant's input can be one or more functions provided by the cloud management platform. For example, when a tenant triggers an experiment creation request while viewing a cloud service's description on the cloud service provider's website, the tenant's input can indicate some or all of these functions, since the same cloud service may offer multiple functionalities. For instance, if OBS can create OBS buckets, upload data to OBS buckets, and download data from OBS buckets, the tenant's input can indicate experiencing one or more of these functions.

[0122] Step 404: Based on the basic experimental information, the cloud management platform obtains the experimental content of the cloud sandbox experiment, uses the infrastructure to create a cloud sandbox experiment to implement the experimental content, and provides the tenant with the operation interface of the cloud sandbox experiment. The experimental content includes at least the experimental tasks of the cloud sandbox experiment and their implementation steps. The experimental tasks and their implementation steps are used to guide the tenant on how to operate the cloud sandbox experiment.

[0123] After obtaining basic experimental information, the cloud management platform can first determine the experimental content of the cloud sandbox experiment based on this information, and then create the cloud sandbox experiment to implement the experimental content using the infrastructure. After creating the cloud sandbox experiment, the cloud management platform needs to provide the tenant with an operation interface for the cloud sandbox experiment. Through this operation interface, the tenant can trigger an entry instruction to indicate that they need to enter the cloud sandbox experiment for operation. The cloud management platform can receive the entry instruction sent by the tenant through this operation interface and, based on the entry instruction, display the experimental platform and experimental content of the cloud sandbox experiment to the tenant. After the cloud management platform displays the experimental platform and experimental content to the tenant, the tenant can perform operations within the experimental platform based on the experimental content to trigger instructions to perform the operations required to complete the experimental tasks in the cloud sandbox experiment. Correspondingly, the cloud management platform can receive the operation instructions sent by the tenant, perform operations on the cloud sandbox experiment according to the operation instructions, and generate operation responses.

[0124] For example, tenants in Figure 6 After clicking the "Create Experiment" button in the user interface shown, the cloud management platform can obtain the experiment content of the cloud sandbox experiment based on the acquired basic experiment information, and use the infrastructure to create the cloud sandbox experiment to implement the experiment content. During this process... Figure 6 The dashed box displays the message "Creating, please wait." After the cloud sandbox experiment is created, Figure 6The system displays the message "Experiment created, click to enter." This message is the operation interface for the cloud sandbox experiment provided by the cloud management platform to the tenant. After clicking this message, the tenant can enter the cloud sandbox experiment platform. The experiment platform displays the experimental content of the cloud sandbox experiment. After the tenant performs the operations according to the instructions of the experimental content, the cloud management platform needs to respond to the operations performed by the tenant. Specifically, the tenant's operation on the cloud sandbox experiment refers to the tenant sending instructions from the client it uses to the cloud host created for it, so that the cloud host performs the corresponding operation according to the instructions. The cloud management platform displays the experimental content to the tenant, which means that the cloud management platform controls the client used by the tenant to display the content that needs to be shown to the tenant.

[0125] The cloud management platform utilizes infrastructure to create cloud sandbox experiments for implementing experimental content. Essentially, this involves the cloud management platform creating cloud hosts for tenants and configuring these hosts to enable them to perform operations necessary for the experimental content. For example, suppose the experimental task requires virtual machines to access a specified cloud service, and completing the task requires the virtual machines to have computing, storage, and network resources with specified specifications. The process of the cloud management platform creating a cloud sandbox experiment for this purpose includes: creating a virtual machine that meets the specified specifications by calling host resources according to the required computing, storage, and network resources and their corresponding specifications; and configuring APIs for the virtual machine to access the specified cloud service.

[0126] In this application, the experimental content includes at least the experimental tasks and implementation steps of a cloud sandbox experiment. Creating a cloud sandbox experiment using infrastructure to implement the experimental content includes: the cloud management platform determining the type of cloud host, required resources and their specifications, and the cloud services to be accessed required to implement the experimental content based on the experimental tasks and implementation steps included in the experimental content; creating the cloud host according to the determined cloud host type, required resources, and specifications; and configuring the cloud host with APIs for accessing cloud services, etc. Optionally, the experimental content also includes detection conditions for the experimental tasks, which indicate the conditions that must be met for the successful completion of the experimental tasks. During the execution of the cloud sandbox experiment, the cloud management platform can detect whether the tenant has successfully completed the experimental tasks based on the detection conditions. Creating a cloud sandbox experiment using infrastructure to implement the experimental content also includes: configuring an application in the cloud host to perform detection and obtain detection results based on the detection conditions. Similarly, the experimental content also indicates the cloud resources required for the experimental tasks. Creating a cloud sandbox experiment using infrastructure to implement the experimental content also includes: configuring the cloud host with the ability to access cloud resources, such as configuring the applications required to access cloud resources, etc. When the experiment content also indicates whether the required cloud resources need to be pre-configured for the experiment task, the cloud sandbox experiment creation process also requires pre-configuring or not pre-configuring the required cloud resources according to this instruction. By pre-configuring the cloud resources required for the cloud sandbox experiment, the cloud resources required for the cloud sandbox experiment can be prepared for the tenant in advance. The tenant does not need to configure the target cloud resources when operating the cloud sandbox experiment, avoiding the tenant spending too much effort and time on the environment preparation stage. Optionally, after the tenant enters the cloud sandbox experiment through the operation interface, the cloud management platform can also display a cloud service topology map to the tenant. Therefore, during the cloud sandbox experiment creation process, the cloud management platform also needs to determine the cloud service topology map of the cloud sandbox experiment and reserve an area in the cloud experiment platform for displaying the cloud service topology map to the tenant. The cloud service topology map is used to indicate the topological relationship between the cloud services required by the cloud sandbox experiment. Similarly, since the tenant needs to operate the cloud sandbox experiment on the experimental desktop, the cloud management platform also needs to determine the desktop image of the cloud sandbox experiment and configure the desktop image for the tenant during the cloud sandbox experiment creation process. The desktop image is used to indicate the configuration required for the tenant's client to display the experimental platform of the cloud sandbox experiment; that is, it indicates the experimental environment on which the experimental desktop of the cloud sandbox experiment depends. For example, the desktop image is the operating system (such as Linux) used by the tenant when operating the cloud sandbox experiment.

[0127] When the experiment content also includes detection conditions for the experimental task, after the tenant performs an operation on the cloud sandbox experiment and the cloud management platform generates a response to the tenant's operation, the method provided in this application further includes: using detection conditions to detect whether the experimental task has been successfully completed in response. When the response meets the conditions indicated by the detection conditions, it is determined that the experimental task has been successfully completed; when the response does not meet the conditions indicated by the detection conditions, it is determined that the experimental task has not been successfully completed.

[0128] Optionally, when the experiment content also includes detection conditions for the experiment task, the cloud management platform, after generating a response to the tenant's operation, may automatically detect whether the experiment task has been successfully completed based on the detection conditions. Alternatively, the cloud management platform may detect whether the experiment task has been successfully completed based on the detection conditions when the tenant indicates that it needs to be detected. In one implementation, the cloud management platform can provide an interactive interface to the tenant, allowing the tenant to indicate to the cloud management platform whether it is necessary to detect whether the experiment task has been successfully completed. This interactive interface can be implemented through APIs and UIs, etc. When the tenant determines that it is necessary to detect whether the experiment task has been successfully completed, it can perform a specified operation on its client, causing the client to send a detection command to the cloud management platform, which indicates that it is necessary to detect whether the experiment task has been successfully completed.

[0129] The following describes the process by which the cloud management platform obtains the experimental content of the cloud sandbox experiment based on the basic experimental information. Depending on the different experimental content instructions, the implementation process for obtaining different experimental content instructions from the cloud management platform is explained below.

[0130] I. For example Figure 8 As shown, the experimental content indicates the experimental tasks and implementation steps of the cloud sandbox experiment.

[0131] In one possible implementation, the cloud management platform obtains the experimental content of the cloud sandbox experiment based on the basic experimental information, including: the cloud management platform generates one or more experimental tasks of the cloud sandbox experiment based on the basic experimental information, and generates the implementation steps required to complete each experimental task based on the one or more experimental tasks.

[0132] Based on the basic information of the cloud sandbox experiment input by the tenant, the cloud management platform can obtain the functions that the tenant can experience by operating the cloud sandbox experiment. According to cloud knowledge data, the cloud management platform can determine one or more operations required to achieve these functions and the execution order of these operations. Based on the function of each operation, these operations can be divided into multiple operation categories, each operation category is used to implement an experimental task, thus obtaining the experimental tasks of the cloud sandbox experiment. Based on the execution order of the operations, the implementation order between different experimental tasks can be obtained, as well as the implementation steps required to complete an experimental task. The cloud knowledge data is used to indicate the cloud services that the cloud management platform can provide. For example, the cloud knowledge data is used to indicate the characteristics, functions, and usage instructions of the cloud services that the cloud management platform can provide.

[0133] Since the experimental content needs to be displayed to tenants, it can be represented in text format. When a cloud sandbox experiment includes one or more experimental tasks, the text-represented experimental content may include one or more chapters corresponding to the experimental tasks. Each chapter indicates the corresponding experimental task and its implementation steps. Each chapter includes at least two parts: a chapter title and chapter content. The chapter title indicates the experimental task. The chapter content below the chapter title indicates the implementation steps required to complete the experimental task indicated by the chapter title. For example... Figure 9 As shown, generating one or more experimental tasks for a cloud sandbox experiment is equivalent to generating one or more chapter titles, and generating the implementation steps required to complete each experimental task is equivalent to generating the chapter content under the corresponding chapter title.

[0134] Example, corresponding to Figure 7 The experimental manual shown includes basic experimental information such as the experimental objectives and a summary of the experimental content. The experimental objective is to help students quickly get started with the Object Storage Service (OBS). The experimental content summary includes creating an OBS instance and testing data upload and download. Based on this basic experimental information, the cloud management platform can obtain information about tenant operations... Figure 7 The cloud sandbox experiment shown allows users to experience functions such as creating OBS, uploading and downloading test data. Based on cloud knowledge data, the cloud management platform can determine that the following operations are required to achieve these functions:

[0135] 1. First, go to the [Experimental Operation Desktop], open your browser, log in to your account, and enter the public cloud console page;

[0136] 2. Then log in to the public cloud website, click "Console" in the upper right corner of the website page, then click "Service List" in the pop-up interface, then click "Object Storage Service OBS" in the storage column of the pop-up interface to enter the Object Storage Service page, and click "Create Bucket" on the bucket list page;

[0137] 3. Then, right-click in the blank area of ​​the main operation interface, select "Create File", and name the created file "sandbox_test";

[0138] 4. Then, on the bucket list page, click the bucket created in step 2 to enter the bucket management interface, and then click the "Object" button in the management interface;

[0139] 5. Then drag the created sandbox_test into the object box and click upload;

[0140] 6. Then click on the uploaded sandbox_test, copy its link to the tab and load it. In the pop-up page after loading, select to download sandbox_test to the / home / user / Downloads / path.

[0141] Based on the functions of the above operations, these operations can be divided into five categories. The experimental tasks implemented by these five categories are: logging into a cloud account, creating an OBS bucket, creating test data, uploading files, and downloading files. Specifically, step 1 is used to log into the cloud account. Step 2 is used to create an OBS bucket. Step 4 is used to create test data. Steps 4 and 5 are used to upload files. Step 6 is used to download files. Therefore, the experimental tasks for the cloud sandbox experiment are: logging into a cloud account, creating an OBS bucket, creating test data, uploading files, and downloading files. Based on the execution order, the order of these five tasks is: first log into the cloud account, then create an OBS bucket, then create test data, then upload files, and finally download files.

[0142] Accordingly, the experimental content represented by the text may include five chapters corresponding to the five experimental tasks. The five chapters are: Logging into a Cloud Account, Creating an OBS Bucket, Creating Test Data, Uploading Files, and Downloading Files. The first chapter is titled: Logging into a Cloud Account. Its content is: Enter the [Experimental Operation Desktop], open a browser, log in to your account, and enter the public cloud console page. The second chapter is titled: Creating an OBS Bucket. Its content is: Log in to the public cloud website, click "Console" in the upper right corner of the website page, then click "Service List" in the pop-up interface, then click "Object Storage Service (OBS)" in the storage section of the pop-up interface to enter the Object Storage Service page, and click "Create Bucket" on the bucket list page. The third chapter is titled: Creating Test Data. Its content is: Right-click in the blank area of ​​the main operation interface, select "Create File," and name the created file "sandbox_test." The fourth chapter is titled: Uploading Files. The content of the first section is as follows: On the bucket list page, click the bucket created in step 2 to enter the bucket management interface, and then click the "Object" button in the management interface; then drag the created sandbox_test into the object box and click upload. The title of the fifth section is: Download File. The content of the fifth section is as follows: Click the uploaded sandbox_test, copy its link to the tab and load it, and in the pop-up page after loading, select to download sandbox_test to the path / home / user / Downloads / .

[0143] Optionally, the cloud management platform can also configure diagrams for experimental tasks, allowing tenants to more clearly understand how the experimental tasks are implemented based on the diagrams. For example... Figure 10As shown, the cloud management platform obtains the experimental content of the cloud sandbox experiment based on the basic experimental information. This also includes: the cloud management platform configuring target illustrations for the experimental tasks based on the experimental tasks. These target illustrations are used to visually represent the implementation method of the experimental tasks. In one possible implementation, the cloud management platform searches for target illustrations matching the experimental tasks in a pre-defined illustration index library. If a matching target illustration exists in the library, it configures that illustration for the experimental task. For example, each illustration in the illustration index library carries an illustration description indicating the content expressed by the illustration. When searching for a target illustration matching the experimental task in the illustration index library, the illustration descriptions of all illustrations in the library can be matched with the textual description of the experimental task. If the illustration description expresses the task that the experimental task needs to implement, that illustration is identified as the target illustration matching the experimental task, and its illustration description is identified as the illustration description of the target illustration. The target illustration configured for the experimental task can be inserted into the chapters in the experimental manual that use text representation to describe the experimental tasks. For example, the target illustration for the experimental task is inserted between the description of the implementation steps of the experimental task and the next chapter.

[0144] II. Figure 8 As shown, the experimental content also indicates the detection conditions of the experimental task. The detection conditions are used to indicate the conditions that should be met for the successful execution of the experimental task.

[0145] The detection conditions for the experimental task are determined by the experimental task itself. Therefore, the cloud management platform, based on the basic experimental information, obtains the experimental content of the cloud sandbox experiment, which also includes: the cloud management platform acquiring the experimental task of the cloud sandbox experiment; and the cloud management platform generating detection conditions that match the experimental task. For example, such as... Figure 11 As shown, after the cloud management platform obtains one or more experimental tasks and their experimental steps from the cloud sandbox experiment, it can choose to first decompose one or more experimental tasks, and then determine the target detection conditions for each experimental task.

[0146] In one possible implementation, such as Figure 11 As shown, the implementation process includes: the cloud management platform obtains the expected state that should be achieved when the experimental task is successfully completed based on the experimental task; the cloud management platform obtains the target detection type to which the target detection condition used to detect whether the expected state has been achieved based on the expected state; the cloud management platform obtains the target detection condition template based on the target detection type, and the target detection condition template is a template that all detection conditions belonging to the target detection type must satisfy; the cloud management platform instantiates the parameters in the target detection condition template based on the expected state to obtain the target detection condition, and uses the target detection condition as the detection condition of the experimental task.

[0147] Although cloud sandbox experiments involve a variety of tasks, they can be categorized into several types. Tasks belonging to the same type can use detection condition templates of the same detection type. However, due to individual differences among tasks, the execution results of different tasks within the same type may vary. These differences can be reflected in the parameters used to achieve the desired state during the execution of the task, and consequently, the parameters of the corresponding detection conditions will also differ. Therefore, when obtaining target detection conditions, the cloud management platform can first obtain the target detection type to which the target detection conditions used to detect whether the desired state has been achieved belong. Then, based on the target detection type, it can obtain the target detection condition template for the target detection conditions. Finally, based on the differentiated parameters in the desired state, it can instantiate the parameters in the target detection condition template to obtain the target detection conditions for the experimental task.

[0148] In one possible implementation, the cloud management platform can pre-configure detection condition templates for various detection types. After determining the target detection type of the target detection condition to which the desired state needs to be achieved after successfully completing the experimental task belongs, the cloud management platform can match this target detection type with the detection condition templates for various detection types. Once a detection condition template for detecting the same target detection type as the desired state is matched, this matched detection condition template is determined as the target detection condition template for the experimental task. Then, based on the differentiated parameters in the desired state, the parameters in the target detection condition template are instantiated to obtain the target detection conditions.

[0149] For example, taking the previous file download experiment as an example, since the implementation steps of this task instruct that sandbox_test be downloaded to the path / home / user / Downloads / , the expected state that the experiment should achieve upon successful completion is: sandbox_test has been downloaded to the path / home / user / Downloads / . The target detection condition for this expected state belongs to the target detection type of "detecting file existence". After matching this target detection type with the detection condition templates of various detection types, we can obtain "detecting file existence in X" as the target detection condition template used for this type of target detection condition. Here, X represents the file's storage path. Furthermore, since the expected state is: sandbox_test has been downloaded to the / home / user / Downloads / path, after obtaining the target detection condition template "check if the file exists in X", the parameters in the target detection condition template can be instantiated based on the storage path of the file indicated by the expected state. X, which identifies the storage path, is instantiated as home / user / Downloads / sandbox_test, resulting in the target detection condition "check if the file exists in home / user / Downloads / sandbox_test".

[0150] It should be noted that, to facilitate the recognition of target detection conditions by computing devices, target detection conditions can be represented by strings with a specified template format. For example, the string corresponding to "detect whether a file exists in X" is "0|0|1|X". Accordingly, the target detection condition above is "0|0|1|home / user / Downloads / sandbox_test". In this way, generating detection conditions that match the experimental task is essentially the cloud management platform generating corresponding strings with a special format based on the experimental task. It is worth noting that all detection conditions can be unified into this string format. The numbers represent different detection types, and the string following the numbers represents the corresponding key information. Furthermore, the format of the detection conditions is not limited to the format in the example; any format that can contain the detection type and corresponding key information is acceptable, and this application embodiment does not impose specific limitations on it.

[0151] Here are a few more examples of strings and the detection conditions they represent:

[0152] The string "1|1|0|filename|matching content" indicates the detection condition: to determine whether the file indicated by the filename in the string contains the matching content indicated by the string.

[0153] The string "1|0|3|number of ECS" indicates the detection condition: determine the number of ECS instances.

[0154] The string "2|1|0|ECS name" indicates the detection condition: to determine whether the ECS indicated by the ECS name in the string exists.

[0155] The string "2|1|1|ECS name|port number" indicates the detection condition: check whether the port indicated by the ECS name in the string is enabled.

[0156] The string "2|1|3|ECS name|number of hard drives" indicates the detection condition: check the total number of hard drives mounted on the ECS instance as indicated by the ECS name in the string.

[0157] Depending on the type of detection, experimental detection is divided into local system detection and external cloud API detection. Local system detection refers to a series of checks performed on the local system, such as the experimental desktop environment, for example, checking if local files exist. External cloud API detection targets the cloud services used in the cloud sandbox experiment. Figure 8 As shown, the experimental platform connects to the cloud service via an API. The platform can request the required cloud service API and then perform detection based on the cloud service's response to the request. External cloud API detection can be used, for example, to detect the number of ECS instances or the existence of a specific OBS bucket.

[0158] The cloud management platform configures detection conditions for experimental tasks, enabling the cloud sandbox experimental platform to intelligently detect the completion status of experimental tasks based on the detection conditions, thereby monitoring whether the corresponding experimental tasks have been successfully executed.

[0159] III. Figure 8 As shown, the experiment also indicates the cloud resources required for the experimental task.

[0160] The cloud resources required for the experimental task are determined by the experimental task itself. Based on the basic experimental information, the cloud management platform obtains the experimental content of the cloud sandbox experiment, including: the cloud management platform acquiring the experimental task of the cloud sandbox experiment; and the cloud management platform determining the cloud resources required for the experimental task based on the experimental task. The experimental content indicates the cloud resources required for the experimental task, including the type of cloud resources required and the specification constraints that the cloud resources must meet. For example, the experimental content indicates that the types of cloud resources required for the experimental task are EVS and ECS. The specification constraints for EVS are: 1 EVS, 20GB disk size for each EVS; and the specification constraints for ECS are: 1 ECS, 16GB memory for each ECS, 4 CPU cores for each ECS, and 40GB system disk size for each ECS. Correspondingly, the cloud management platform determines the cloud resources required for the experimental task based on the experimental task, including: the cloud management platform acquiring the type of cloud resources required to complete the experimental task and the specification constraints that the cloud resources must meet.

[0161] In one possible implementation, the cloud management platform determines the type of cloud resources required to complete each experimental task in the cloud sandbox experiment, as well as the maximum demand for each type of cloud resource. When a certain cloud resource is used only to complete one experimental task in the cloud sandbox experiment, the maximum demand for that cloud resource to complete that experimental task is the maximum specification of that cloud resource allowed to be configured to complete that experimental task. When a certain cloud resource is used to complete multiple experimental tasks in the cloud sandbox experiment, the maximum value among the maximum demands for that cloud resource to complete these multiple experimental tasks is the maximum specification of that cloud resource allowed to be configured to complete each of these multiple experimental tasks. The limitation on the maximum specification of cloud resources allowed to be configured to complete experimental tasks here is the specification constraint condition that the cloud resources need to meet. Optionally, the cloud management platform can pre-record the type of cloud resources required to complete each experimental task that the cloud management platform can provide, as well as the maximum demand for each type of cloud resource, and this maximum demand can be obtained through simulation experiments. In this implementation, since it is necessary to determine the maximum demand for cloud resources to complete each experimental task when determining the cloud resources required for the experimental task, therefore, as Figure 12 As shown, after the cloud management platform obtains one or more experimental tasks and their experimental steps for the cloud sandbox experiment, it can choose to first break down one or more experimental tasks, and then determine the maximum amount of cloud resources required to complete each experimental task, thereby obtaining the specifications of the cloud resources required for the experimental task.

[0162] It should be noted that the granularity of cloud resource demand can be changed according to application requirements. For example, assuming the cloud resource is a cloud server, in some application scenarios, the granularity of the demand for cloud servers can be the number of cloud servers, while in other application scenarios, the granularity of the demand for cloud servers can be finer, such as limiting not only the number of cloud servers, but also the specifications of the virtual processors and virtual memory used by the cloud servers. This application does not impose specific limitations on this.

[0163] The cloud management platform obtains the specifications and restrictions that the cloud resources required to complete the cloud sandbox experiment must meet. This is equivalent to the cloud management platform determining the rules that the cloud resources required to complete the experiment must meet, so as to restrict the cloud resources used by the cloud sandbox experiment, avoid waste of cloud resources, and thus improve the utilization rate of cloud resources.

[0164] Optional, such as Figure 13 As shown, the cloud management platform determines the cloud resources required for the experimental task based on the experimental task. This also includes: determining whether the cloud management platform needs to pre-configure the required cloud resources for the experimental task; and if so, determining the cloud resources and their specifications required for pre-configuration. In one possible implementation, the cloud management platform determines the experimental type of the experimental task to be completed in the cloud sandbox experiment, and then determines whether cloud resources need to be pre-configured based on the experimental type. For example, when the experimental task indicates that the experimental task type is an experiential experiment, it can be determined that the tenant mainly wants to experience the functions of the cloud service, without caring about the configuration method of the cloud service. In this case, the cloud management platform determines that the cloud resources need to be pre-configured for the experimental task. Determining the target cloud resources that need to be pre-configured for the experimental task includes: determining the type and specifications of the target cloud resources that need to be pre-configured for the experimental task. Accordingly, pre-configuring target cloud resources can be achieved by loading the type and specifications of the target cloud resources into a configuration template. The cloud management platform then restricts or pre-configures the experimental resources of the cloud sandbox experiment according to this configuration data in the configuration template. Figure 8As shown, the experimental platform connects to the cloud service via a cloud API. The experimental platform configures cloud resources by requesting the required cloud service API. When the experimental task indicates that it is a hands-on experiment, it can be determined that the tenant not only wants to experience the functionality of the cloud service but also wants to understand how to configure it. Therefore, the cloud management platform determines that it does not need to pre-configure the required cloud resources for the experimental task. In this case, the cloud management platform does not need to pre-configure the required cloud resources in the cloud sandbox experiment; the tenant configures them when operating the cloud sandbox experiment. By determining the target cloud resources that need to be pre-configured for the experimental task and pre-configuring them when creating the cloud sandbox experiment, the cloud management platform eliminates the need for the tenant to configure these target cloud resources when operating the cloud sandbox experiment, allowing them to directly perform the experiment and avoiding the tenant spending too much effort and time on the environment preparation stage.

[0165] In another implementation, the cloud management platform can also provide an interactive interface to tenants, allowing them to indicate whether the necessary cloud resources need to be pre-configured for the experimental task. This interface can be implemented through APIs and UIs. When a tenant determines whether or not the cloud management platform needs to pre-configure the required cloud resources for its experimental task, it can perform a specified operation on its client, causing the client to send a configuration command to the cloud management platform. This configuration command indicates whether the necessary cloud resources need to be pre-configured for the experimental task. For example, during the process of obtaining the experimental content of the cloud sandbox experiment based on the basic experimental information, the cloud management platform can display a query interface to the tenant, asking whether the tenant needs to pre-configure the required cloud resources for the experimental task. After the tenant inputs information on this query interface, the configuration command can be transmitted to the cloud management platform. Alternatively, the basic information of the cloud sandbox experiment may also include configuration requirement information, which indicates whether the necessary cloud resources need to be pre-configured for the experimental task. When entering the basic experimental information, the tenant can input a configuration command in the configuration requirement information indicating whether the cloud management platform needs to pre-configure the required cloud resources for the experimental task, according to its own needs. When a tenant's configuration instructions indicate that cloud resources need to be pre-configured for an experimental task, the cloud management platform determines that pre-configuration of the required cloud resources is necessary. Therefore, during the creation of a cloud sandbox experiment, the cloud management platform also needs to pre-configure the required cloud resources for the experimental task according to the type of cloud resources required and the specification limitations that the cloud resources must meet, as indicated by the experiment content. Conversely, if a tenant's configuration instructions indicate that cloud resources need to be pre-configured for an experimental task, the cloud management platform determines that pre-configuration is not required. Therefore, during the creation of a cloud sandbox experiment, the cloud management platform does not need to pre-configure the required cloud resources for the experimental task.

[0166] When a cloud sandbox experiment uses multiple cloud services, the cloud management platform can also obtain a cloud service topology map and display it to the tenant when the tenant operates the cloud sandbox experiment. This allows the tenant to easily understand the cloud services required for the cloud sandbox experiment and the relationships between them based on the cloud service topology map. The cloud service topology map is used to indicate the topological relationships between the cloud services required for the cloud sandbox experiment. Therefore, the method for creating a cloud sandbox experiment provided in this application embodiment further includes: the cloud management platform matches the topological relationships between cloud services recorded in the cloud management platform based on the multiple cloud services required for the cloud sandbox experiment to obtain the cloud service topology map of the cloud sandbox experiment. In one possible implementation, when multiple cloud services included in a certain topological relationship recorded by the cloud management platform correspond one-to-one with the multiple cloud services required for the cloud sandbox experiment, the cloud management platform determines the cloud service topology map indicating that topological relationship as the cloud service topology map of the cloud sandbox experiment. The cloud service topology map of the topological relationships recorded by the cloud management platform can be pre-built offline.

[0167] When a cloud management platform creates a cloud sandbox experiment, it also needs to determine the desktop image of the cloud sandbox experiment and configure the desktop image for the tenant. Therefore, the method for creating a cloud sandbox experiment provided in this application embodiment further includes: the cloud management platform matches the desktop images recorded on the cloud management platform based on the experiment type of the cloud sandbox experiment to obtain the desktop image of the cloud sandbox experiment. In one possible implementation, the cloud management platform has pre-built correspondences between various experiment types and their corresponding desktop images. When determining the desktop image of the cloud sandbox experiment, the cloud management platform first determines the experiment type of the cloud sandbox experiment, and then queries the correspondence between the experiment type and the desktop image based on that experiment type to obtain the desktop image of the cloud sandbox experiment. The desktop image is used to instruct the tenant's client to display the configuration required for the experiment platform of the cloud sandbox experiment. That is, the desktop image is used to indicate the experimental environment on which the experimental desktop of the cloud sandbox experiment depends. When the tenant operates the cloud sandbox experiment, the operation is performed on the client used by the tenant, and the operation interacts with the cloud services provided by the cloud management platform to achieve the purpose of operating the cloud sandbox experiment. When a tenant performs an operation on the client, the operation is actually performed on the client's experimental desktop. For example, the desktop image is the desktop operating system used by the experimental desktop, such as Windows or Linux.

[0168] It should be noted that, as Figure 8 As shown, in addition to displaying the experimental tasks and implementation steps, testing conditions, cloud resources required for the experimental tasks, cloud service topology diagrams, and desktop images to tenants, the cloud sandbox experiment platform also has other functions such as account management.

[0169] In some implementation scenarios, the process of obtaining the experimental content described above can optionally be implemented by a cloud management platform using a machine learning (ML) model. For example, such as... Figure 14 As shown, the machine learning model can be selected as a large model. A large model (also called a foundation model) refers to a machine learning model with a large number of parameters and a complex structure. It has powerful semantic understanding and logical reasoning capabilities, and can process massive amounts of data and complete various complex tasks, such as natural language processing, computer vision, and speech recognition.

[0170] At this point, based on the basic experimental information, the cloud management platform obtains the experimental content of the cloud sandbox experiment, including: the cloud management platform inputs the basic experimental information into the target large model and receives the experimental content of the cloud sandbox experiment output by the target large model. For example, when the experimental content indicates the experimental task and its implementation steps of the cloud sandbox experiment, after the cloud management platform inputs the basic experimental information into the target large model, the target large model can output the experimental task and its implementation steps of the cloud sandbox experiment, and thus the cloud management platform can obtain the experimental task and its implementation steps of the cloud sandbox experiment. When the experimental content indicates the experimental task and its implementation steps of the cloud sandbox experiment, the detection conditions of the experimental task, and the cloud resources required for the experimental task, after the cloud management platform inputs the basic experimental information into the target large model, the target large model can output the experimental task and its implementation steps of the cloud sandbox experiment, the detection conditions of the experimental task, and the cloud resources required for the experimental task, and thus the cloud management platform can obtain the experimental content output by the target large model. In one possible implementation, the target large model includes a first sub-model, a second sub-model, and a third sub-model. After the cloud management platform inputs the basic experimental information into the first sub-model, the first sub-model can output the experimental task and its implementation steps for the cloud sandbox experiment. Then, after the cloud management platform inputs the experimental task and its implementation steps into the second sub-model, the second sub-model can output the detection conditions for the experimental task. After the cloud management platform inputs the experimental task and its implementation steps into the third sub-model, the third sub-model can output the cloud resources required for the experimental task. Alternatively, the first sub-model can be cascaded with the second and third sub-models respectively. This cascading relationship allows the first sub-model to output to the cloud management platform, the second sub-model, and the third sub-model respectively, eliminating the need for the cloud management platform to input the output of the first sub-model into the second and third sub-models.

[0171] When generating experimental content using a large model, the cloud management platform can instruct the large model to perform tasks by inputting task instructions. These instructions carry a task description and the input data required for the large model to execute the task. In one possible implementation, task instructions can be conveyed through prompt templates. These prompt templates are task instructions input to the large model according to a specific instruction format and description; different instruction formats or descriptions will generate different results.

[0172] For example, the prompt template that the cloud management platform inputs to the first sub-model is as follows:

[0173] You are a cloud sandbox experiment creation expert. Based on the following basic experimental information, please generate the cloud sandbox experiment task and its implementation steps. Please complete the task generation by following these steps:

[0174] Generate chapter titles that are clearly structured and relevant to the experimental topic;

[0175] Generate detailed chapter content for each chapter title;

[0176] Based on the images in the known image index library, insert the target image that matches the experimental task at the appropriate location;

[0177] The basic information of the experiment is as follows: <******>.

[0178] The "******" following the basic experimental information indicates the specific content of the basic experimental information.

[0179] The prompt template that the cloud management platform inputs to the second sub-model is as follows:

[0180] You are a cloud sandbox experiment creation expert. Please generate appropriate experimental testing conditions based on the following experimental tasks. If the task does not require generating experimental testing conditions or cannot obtain experimental testing conditions, output empty;

[0181] Experimental task: <******>.

[0182] The "******" following the experimental task indicates the description of the experimental task.

[0183] The prompt template that the cloud management platform inputs to the third sub-model is as follows:

[0184] You are a cloud sandbox experiment creation expert. From the following experimental tasks, extract the types and specifications of the core cloud resources required to implement them, and output the results in JSON format. If the task does not involve cloud resource specifications, output empty.

[0185] Experimental task: <******>.

[0186] The "******" following the experimental task indicates the description of the experimental task.

[0187] like Figure 14 As shown, the large model can be pre-trained before being applied to the experimental content of the cloud sandbox experiment. Its training process includes the following two training phases:

[0188] The first training phase involves the cloud management platform using cloud knowledge data and historical cloud sandbox experiment content as the initial training data to incrementally pre-train the initial large model. The cloud knowledge data indicates the cloud services that the cloud management platform can provide. The purpose of this first training phase is to enable the trained initial large model to recognize the cloud knowledge data and the experimental content of the cloud sandbox experiments. The experimental content of the historical cloud sandbox experiments is typically obtained manually.

[0189] The second training phase involves the cloud management platform extracting key information from specified experiments in historical cloud sandbox experiments as secondary training data. This data is then used to fine-tune the initial large model, which has undergone incremental pre-training, to obtain the target large model used to generate the specified experimental content. The purpose of this second training phase is to enable the trained target large model to generate the required specified experimental content more accurately.

[0190] The training process of the first and second training phases is basically the same as the current model training process, and will not be described in detail here. The following mainly describes the training data used in the second training phase.

[0191] In the first scenario, when the experimental tasks, implementation steps, detection conditions, and cloud resources required for the cloud sandbox experiment are generated by the same large model, the specified experimental content for training the historical cloud sandbox experiment includes the experimental tasks, implementation steps, detection conditions, and cloud resources required for the historical cloud sandbox experiment. In the second scenario, when the experimental tasks and implementation steps are generated by the first sub-model, the detection conditions are generated by the second sub-model, and the cloud resources required for the experimental task are generated by the third sub-model, the specified experimental content for training the first sub-model is the experimental tasks and implementation steps; the specified content for training the second sub-model is the detection conditions; and the specified content for training the third sub-model is the cloud resources required for the historical cloud sandbox experiment. The training data in this second scenario is also called the instruction fine-tuning dataset.

[0192] The first instruction fine-tuning dataset for the first sub-model was obtained by extracting key information from the experimental manual of the historical cloud sandbox experiment. For example... Figure 15 As shown, when constructing the first instruction fine-tuning dataset for the first sub-model, it is necessary to extract the basic experimental information, experimental tasks, and their implementation steps from the historical cloud sandbox experiment manual. If the historical cloud sandbox experiment manual also includes attached figures and their descriptions for the experimental tasks, it is also necessary to extract these figures and their descriptions from the historical cloud sandbox experiment manual. The basic experimental information can be obtained through information extraction. For example, relevant information extraction techniques can be used to extract content indicated by specified keywords from the experiment manual. The experimental tasks, their implementation steps, attached figures, and their descriptions can be obtained by deconstructing the content of the experiment manual. For example, the experimental tasks can be obtained by identifying the chapter titles in the experiment manual, the implementation steps of the experimental tasks can be obtained by identifying the chapter content under the chapter titles, the images inserted between the chapter titles and the next chapter titles can be identified as the attached figures for the experimental tasks indicated by the current chapter titles, and the descriptions added to these attached figures are the descriptions for the attached figures for the experimental tasks indicated by the current chapter titles. It should be noted that the attached figure index library can be composed of attached figures obtained during the training process.

[0193] The first instruction fine-tuning dataset obtained in this way is actually a series of first input-output data pairs. Each first input-output data pair includes basic experimental information from historical cloud sandbox experiments, experimental tasks and their implementation steps, and attached figures and their descriptions configured for the experimental tasks. Specifically, the basic experimental information from historical cloud sandbox experiments serves as the input for training the first sub-model, while the experimental tasks and their implementation steps, along with the attached figures and their descriptions, serve as the output for training the first sub-model. Using such data pairs to train the first sub-model enables it to generate experimental tasks and their implementation steps, as well as the attached figures and their descriptions configured for the cloud sandbox experiments, based on the basic experimental information from the cloud sandbox experiments.

[0194] For example, the following is an example. Figure 7 Taking the experimental manual shown as an example, the process of constructing the first instruction fine-tuning dataset for training the first sub-model based on it is illustrated.

[0195] First of all, Figure 7 The experimental manual shown extracts the basic experimental information from the historical cloud sandbox experiment, resulting in the following basic experimental information:

[0196] Experimental objective: To help students quickly get started and experience the Object Storage Service (OBS).

[0197] Experiment Summary: Create OBS, test data upload and download.

[0198] Experiment type: experiential.

[0199] The experiment involves the cloud service OBS.

[0200] Then, to Figure 7 The experimental manual shown is deconstructed, focusing on the experimental steps, resulting in the experimental tasks and implementation steps of the historical cloud sandbox experiment, along with accompanying diagrams and explanations. The content is as follows:

[0201] Experiment Task 1: Log in to your cloud account.

[0202] Steps to implement Experiment Task 1: Go to the [Experiment Operation Desktop], open your browser, log in to your account, and enter the public cloud console page.

[0203] Please see the attached diagram for Experiment Task 1. Figure 16 .

[0204] The attached image for Experiment Task 1 shows the login interface for the cloud account.

[0205] Experiment Task 2: Create an OBS bucket.

[0206] The steps to implement Experiment Task 2 are as follows: Log in to the public cloud website, click "Console" in the upper right corner of the website page, then click "Service List" in the pop-up interface, then click "Object Storage Service" in the storage column of the pop-up interface to enter the object storage service page, and select "Create Bucket" on the bucket list page.

[0207] Please see the attached diagram for Experiment Task 2. Figure 17 .

[0208] The accompanying diagram for Experiment Task 2 shows: Access the object storage service page and create a bucket.

[0209] Experiment Task 3: Create test data.

[0210] The steps to implement Experiment Task 3 are as follows: Right-click on the blank area of ​​the main operation interface, select Create Document, and name the created document sandbox_test.

[0211] For the attached image in Experiment Task 3, please see [link to Experiment Task 3]. Figure 18 Please see the attached image for the document named sandbox_test. Figure 19 .

[0212] The accompanying diagram for Experiment Task 3 shows the creation of test data (sandbox_test) on the sandbox desktop.

[0213] Experiment Task 4: Upload files.

[0214] The steps to implement Experiment Task 4 are as follows: On the bucket list page, click the bucket created in step 2 to enter the bucket management interface, and then click the "Upload Object" button in the management interface; then drag the created sandbox_test into the object box and click upload.

[0215] Please see the attached diagram for Experiment Task 4. Figure 20 .

[0216] The accompanying diagram for Experiment Task 4 shows: Enter the object storage bucket and click "Upload Object" to upload the file.

[0217] Experiment Task 5: Download files.

[0218] The steps to implement Experiment Task 5 are as follows: Click on the uploaded sandbox_test, copy its link to the tab and load it. In the pop-up page after loading, select to download sandbox_test to the path / home / user / Downloads / .

[0219] Please see the attached diagram for Experiment Task 5. Figure 21 .

[0220] The accompanying diagram for Experiment Task 5 shows that clicking the object link in the object bucket will allow the test file to be downloaded normally.

[0221] Based on this, a complete first input-output data pair for training the first sub-model is constructed, wherein the input used for training the first sub-model in the first input-output data pair is:

[0222] Experimental objective: To help students quickly get started and experience the Object Storage Service (OBS).

[0223] Experiment Summary: Create OBS, test data upload and download.

[0224] Experiment type: experiential.

[0225] The experiment involves the cloud service OBS.

[0226] The output used in the first input-output data pair for training the first sub-model is:

[0227] Experiment Task 1: Log in to your cloud account.

[0228] Steps to implement Experiment Task 1: Go to the [Experiment Operation Desktop], open your browser, log in to your account, and enter the public cloud console page.

[0229]

【 Figure 16 [Image of the login screen for your cloud account]

[0230] Experiment Task 2: Create an OBS bucket.

[0231] The steps to implement Experiment Task 2 are as follows: Log in to the public cloud website, click "Console" in the upper right corner of the website page, then click "Service List" in the pop-up window, then click "Object Storage Service" in the storage column of the pop-up window to enter the object storage service page, and click "Create Bucket" on the bucket list page.

[0232]

【 Figure 17 [Go to the object storage service page and create a bucket]

[0233] Experiment Task 3: Create test data.

[0234] The steps to implement Experiment Task 3 are as follows: Right-click on the blank area of ​​the main operation interface, select Create Document, and name the created document sandbox_test.

[0235]

【 Figure 18 [Select to create document] Figure 19 Name the created document sandbox_test

[0236] Experiment Task 4: Upload files.

[0237] The steps to implement Experiment Task 4 are as follows: On the bucket list page, click the bucket created in step 2 to enter the bucket management interface, and then click the "Object" button in the management interface; then drag the created sandbox_test into the object box and click upload.

[0238]

【 Figure 20 [Go to the object bucket and click "Upload Object" to upload the file]

[0239] Experiment Task 5: Download files.

[0240] The steps to implement Experiment Task 5 are as follows: Click on the uploaded sandbox_test, copy its link to the tab and load it. In the pop-up page after loading, select to download sandbox_test to the path / home / user / Downloads / .

[0241]

【 Figure 21 [Clicking the object link in the object bucket allows the test file to download normally.]

[0242] The second instruction fine-tuning dataset for the second sub-model was obtained by extracting key information from the experimental manual of the historical cloud sandbox experiment. For example... Figure 22As shown, when constructing the second instruction fine-tuning dataset for the second sub-model, it is necessary to deconstruct the content of the historical cloud sandbox experiment manual. Based on the deconstructed content, the experimental tasks included in the manual are broken down, and each experimental task, its experimental steps, and detection conditions are extracted from the manual. According to pre-configured detection condition representation rules, the detection conditions are represented as strings with a special format. The experimental tasks, their experimental steps, and detection conditions can be obtained through information extraction. For example, relevant information extraction techniques can be used to extract content indicated by specified keywords from the manual. Furthermore, the extracted detection conditions can be selected as system detection conditions and / or cloud API detection conditions. System detection conditions, for example, are for detecting local files. Cloud API detection conditions, for example, are for detecting cloud services such as ECS, OBS, and VPC.

[0243] The second instruction fine-tuning dataset obtained in this way is actually a series of second input-output data pairs. Each second input-output data pair includes the experimental task and its steps from the historical cloud sandbox experiment, and a string representing the detection conditions of the experimental task. The experimental task and its steps from the historical cloud sandbox experiment serve as the input for training the second sub-model, and the string representing the detection conditions of the experimental task serves as the output for training the second sub-model. Using such data pairs to train the second sub-model enables it to generate the string representing the detection conditions based on the experimental task and its steps from the cloud sandbox experiment.

[0244] For example, the following is an example. Figure 7 Taking the experimental manual shown as an example, the process of constructing the second instruction fine-tuning dataset for training the second sub-model is illustrated.

[0245] First of all, Figure 7 The experimental manual shown is deconstructed. Based on the deconstructed content, the experimental tasks included in the manual are broken down to obtain the "Download File" experimental task, and its implementation steps are: "Click on the uploaded sandbox_test, copy its link to the tab and load it, and in the pop-up page after loading, select to download sandbox_test to the path / home / user / Downloads / ". Then, based on the string below this experimental task: detection(0|0|1| / home / user / Downloads / sandbox_test), this string represents the detection condition for the "Download File" task, and this detection condition is used to detect whether the file sandbox_test has been successfully downloaded to the specified path home / user / Downloads / .

[0246] Based on this, a complete second input-output data pair for training the second sub-model is constructed. The input for training the second sub-model is: download file, click on the uploaded sandbox_test, copy its link to the tab and load it, and in the pop-up window, select to download sandbox_test to the path / home / user / Downloads / . The output for training the second sub-model is: 0|0|1| / home / user / Downloads / sandbox_test.

[0247] Here are a few more examples of second input-output data pairs used to train the second sub-model:

[0248] The input used in the second input-output data pair 1 for training the second sub-model is: writing "husky" into dog.txt. The output used in the second input-output data pair 1 for training the second sub-model is: 1|1|0|dog.txt|husky.

[0249] The input used in the second input-output data pair 2 for training the second sub-model is: creating 6 ECSs. The output used in the second input-output data pair 2 for training the second sub-model is: 1|0|3|6.

[0250] The input used in the second input-output data pair 3 for training the second sub-model is: setting the name of ECS to floppy. The output used in the second input-output data pair 3 for training the second sub-model is: 2|1|0|floppy.

[0251] The input used in the second input-output data pair 4 for training the second sub-model is: setting the port number of the ECS named floppy to 8888. The output used in the second input-output data pair 4 for training the second sub-model is: 2|1|1|floppy|8888.

[0252] The input used in the second input-output data pair 5 for training the second sub-model is: an ECS instance named floppy with 3 hard drives mounted. The output used in the second input-output data pair 5 for training the second sub-model is: 2|1|3|floppy|3.

[0253] The input used in the second input-output data pair 6 for training the second sub-model is: create one ECS instance, name it floppy, set the port number to 8888, and mount three hard drives. The output used in the second input-output data pair 6 for training the second sub-model is: 1|0|3|1, 2|1|0|floppy, 2|1|1|floppy|8888, 2|1|3|floppy|3.

[0254] The third instruction fine-tuning dataset for the third sub-model was obtained by extracting key information from the experimental manual of the historical cloud sandbox experiment. For example... Figure 23 As shown, when constructing the third instruction fine-tuning dataset for the third sub-model, it is necessary to deconstruct the content of the historical cloud sandbox experiment manual. Based on the deconstructed content, the experimental tasks included in the manual are broken down, and key information such as the implementation steps, cloud services used, and cloud service specifications are extracted from the manual for each experimental task. The experimental tasks, implementation steps, cloud services used, and cloud service specifications can be obtained through key information extraction. For example, relevant information extraction techniques can be used to extract content indicated by specified keywords from the experimental manual.

[0255] The third-instruction fine-tuning dataset obtained in this way is actually a series of third-input-output data pairs. Each third-input-output data pair includes the experimental task and its implementation steps from the historical cloud sandbox experiment, the cloud service used, and the specifications of the cloud service. The experimental task and its implementation steps from the historical cloud sandbox experiment serve as the input for training the third sub-model, while the cloud service used and its specifications serve as the output for training the third sub-model. Using such data pairs to train the third sub-model enables it to determine the cloud resources required for the experimental task based on the experimental task and its implementation steps from the cloud sandbox experiment.

[0256] For example, the following is an example. Figure 7 Taking the experimental manual shown as an example, the process of constructing the third instruction fine-tuning dataset for training the third sub-model is illustrated.

[0257] First of all, Figure 7The experimental manual shown is deconstructed, and the experimental tasks included in the manual are broken down based on the deconstructed content. The experimental task "Creating an OBS Bucket" is obtained, and its implementation steps are: "Log in to the public cloud website, click 'Console' in the upper right corner of the website page, then click 'Service List' in the pop-up interface, then click 'Object Storage Service OBS' in the storage section of the pop-up interface to enter the Object Storage Service page, and click 'Create Bucket' on the bucket list page." Then, from the description of the experimental task "Creating an OBS Bucket," the following key information is obtained: "Resources involved in the experiment: OBS" and "Resource specifications: {"Quantity": 1}."

[0258] Accordingly, a complete third input-output data pair for training the third sub-model is constructed. The input used for training the third sub-model in this third input-output data pair is: Creating an OBS bucket: Log in to the public cloud website, click "Console" in the upper right corner of the website page, then click "Service List" in the pop-up window, and then click "Object Storage Service (OBS)" in the storage section of the pop-up window to enter the Object Storage Service page. Click "Create Bucket" on the bucket list page. The output used for training the third sub-model in the third input-output data pair is: {"Resource Name": "OBS", "Quantity": 1}.

[0259] Based on the above instructions and fine-tuning of the dataset, a large model can be trained to better complete the relevant tasks of experiment creation, namely, generating cloud sandbox experiments, their implementation steps, detection conditions, and cloud resources required for the experiments. The following is a specific example illustrating the process of creating a sandbox experiment end-to-end using a large model.

[0260] Assume the tenant inputs the following basic experimental information:

[0261] Experimental objective: To become familiar with the ECS and VPC creation process, and to become familiar with the process of adding EVS to ECS.

[0262] Experiment content: Create a VPC and an ECS instance, and add EVS to the ECS instance.

[0263] Experiment type: experiential.

[0264] The experiment involved cloud services: ECS, VPC, and EVS.

[0265] I. Based on the above experimental information, the first sub-model generates the experimental tasks for the cloud sandbox experiment. The generated content is as follows:

[0266] 1. Log in to your cloud account.

[0267] 2. Create a Virtual Private Cloud (VPC).

[0268] 2.1 Access the Virtual Private Cloud (VPC) interface.

[0269] 2.2 Create a Virtual Private Cloud (VPC).

[0270] 3. Create an Elastic Cloud Server (ECS).

[0271] 3.1 Access the Elastic Cloud Server (ECS) interface.

[0272] 3.2 Create an Elastic Cloud Server (ECS).

[0273] 4. Add EVS (Elastic Cloud Server) to Elastic Cloud Server (ECS).

[0274] 4.1 Purchase EVS cloud disks.

[0275] 4.2 Mounting the cloud disk EVS.

[0276] II. The first sub-model generates the implementation steps for each experimental task and configures corresponding diagrams and descriptions for each task. The generated content is as follows:

[0277] 1. Log in to your cloud account

[0278] Enter the [Experimental Operation Desktop], open your browser, log in to your account, and access the public cloud console page.

[0279]

【 Figure 16 [Image of the login screen for your cloud account]

[0280] 2. Create a Virtual Private Cloud (VPC)

[0281] Virtual Private Cloud (VPC) creates an isolated virtual private network environment for Elastic Cloud Server (ECS) resources. Purchasing an ECS requires binding to a VPC.

[0282] 2.1 Accessing the Virtual Private Cloud (VPC) Interface

[0283] Once you're in the cloud console, move your mouse to the left-hand menu and go to the service list. First, click "Network," and then in the pop-up window, click "Virtual Private Cloud (VPC)."

[0284] 2.2 Creating a Virtual Private Cloud (VPC)

[0285] Click the "Create Virtual Private Cloud" button in the upper right corner to enter the VPC creation interface.

[0286]

【 Figure 24 Creating a VPC Interface

[0287] Basic information configuration:

[0288] Region: North China - Beijing

[0289] Name: vpc-test

[0290] IPv4 network segment: 192.168.0.0 / 00

[0291] Default subnet configuration:

[0292] Availability Zone: Availability Zone 1

[0293] Name: subnet-test

[0294] Subnet IPv4 field: 192.168.0.0 / 24

[0295] Other settings will remain unchanged using default values.

[0296]

【 Figure 25 VPC Basic Information Configuration Interface

[0297] Then click the "Create Now" button to complete the VPC creation.

[0298] 3. Create an Elastic Cloud Server (ECS)

[0299] The database service is deployed on an Elastic Cloud Server (ECS). To build a single-machine database, you need to purchase an Elastic Cloud Server (ECS). Recommended specifications: 4 CPU cores and 16GB of memory.

[0300] 3.1 Accessing the Elastic Cloud Server (ECS) Interface

[0301] Access the Huawei Cloud console, move your mouse to the left-hand menu, and navigate to the service list: "Compute" -> "Elastic Cloud Server (ECS)". Click the "Purchase Elastic Cloud Server" button in the upper right corner.

[0302]

【 Figure 26 [Easy to purchase Elastic Cloud Server (ECS) interface]

[0303] 3.2 Creating an Elastic Cloud Server (ECS)

[0304] Basic configuration:

[0305] ① Billing model: Pay-as-you-go

[0306] ②Region: North China - Beijing

[0307] ③Availability Zone: Availability Zone 1

[0308] ④ CPU Architecture: Kunpeng Computing

[0309] ⑤ Specifications: Kunpeng General Purpose Enhanced Computing | kc1.xlarge.4 | 4 vCPUs | 16GB

[0310] ⑥ Mirror: Public Mirror

[0311] ⑦ Operating System: openEuler – openEuler 20.03 64bit with ARM (40GB)

[0312] ⑧ System disk: General purpose SSD 40G

[0313] 9. Purchase quantity: 1 unit

[0314] Click Next to proceed with network configuration.

[0315] Network configuration:

[0316] Network: VPC: vpc-test (consistent with the steps in creating the VPC), Subnet: subnet-test, automatically assign IP addresses

[0317] ② Security group: Use the default sg-test

[0318] ③ Elastic public IP: Purchase now

[0319] ④ Route: Fully Dynamic BGP

[0320] ⑤ Public network bandwidth: billed based on bandwidth

[0321] ⑥ Bandwidth: 2 Mbit / s

[0322]

【 Figure 27 Elastic Cloud Server (ECS) network configuration interface

[0323] Click Next to proceed to advanced configuration.

[0324] Advanced configuration:

[0325] ① Cloud server name: ecs-test

[0326] ② Login credentials: Password

[0327] ③ Username: root

[0328] ④ Password:.........

[0329] ⑤ Confirm password:............

[0330] ⑥ Cloud Backup: Not for now

[0331]

【 Figure 28 Advanced configuration interface for Elastic Cloud Server (ECS)

[0332] Leave other options at their default settings, click Next to confirm the configuration, and then click "Buy Now". (ECS creation will take approximately 2 minutes.)

[0333] 4. Add EVS (Elastic Cloud Server) to Elastic Cloud Server (ECS)

[0334] 4.1. Purchase EVS cloud disks

[0335] Log in to the management console, move your mouse to the left-hand menu bar of the experimental operation desktop browser page, and select "Service List > Storage > EVS". On the EVS page, click "Purchase Disk".

[0336]

【 Figure 29 [EVS Cloud Disk Purchase Interface]

[0337] Configure the basic information of the cloud disk according to the on-screen prompts.

[0338] ① Billing model: Pay-as-you-go

[0339] ②Region: North China - Beijing

[0340] ③Availability Zone: Availability Zone 1

[0341] ④ Disk Specifications: General Purpose SSD

[0342] ⑤ Disk size: 20G

[0343] ⑥ Cloud Backup: Not for now

[0344] ⑦ More: Not configured yet

[0345] ⑧ Disk name: volume-test

[0346] 9. Purchase quantity: 1

[0347]

【 Figure 30 [Configure EVS basic information interface]

[0348] Click "Buy Now".

[0349] On the "Details" page, you can double-check the EVS (Elastic Virtual Disk) information. Once confirmed, click "Submit" to begin creating the EVS. If further modifications are needed, click "Previous" to edit the parameters. Return to the "Disk List". On the "EVS" main page, check the EVS status. When the EVS status changes to "available", the creation was successful.

[0350] 4.2. Mounting the EVS cloud disk

[0351] A separately purchased EVS (Elastic Cloud Disk) is a data disk. You can see the disk attribute as "Data Disk" and the disk status as "Available" in the EVS list. You need to mount this data disk to an Elastic Cloud Server (ECS) for use.

[0352] In the EVS cloud disk list, find the cloud disk we created, click on the "Mount" option on the right, and the disk mounting configuration window will pop up.

[0353]

【 Figure 31 [EVS Cloud Disk Mounting Interface]

[0354] Select the Elastic Cloud Server (ECS) to which the EVS (Elastic Cloud Server Deployment System) will be mounted. Here, we choose ecs-test. This ECS must be located on the same availability partition as the EVS. Select the "Mount Point" from the drop-down list, and choose the data disk as the mount point. Click "Confirm." Return to the EVS list page. At this point, the EVS status will be "Mounting," indicating that the EVS is in the process of being mounted to the ECS. When the EVS status is "in-use," the mounting to the ECS is successful.

[0355] III. The second sub-model generates the detection conditions for each experimental task. The generated content is as follows:

[0356] The detection condition for 【2.2 Creating a Virtual Private Cloud (VPC)】 is “3|0|5|1”. The template for this detection condition is “3|0|5|Expected Number” (the number of VPCs to be detected).

[0357] The detection condition for 【3.2 Creating an Elastic Cloud Server (ECS)】 is “1|0|3|1”. The template for this detection condition is “1|0|3|Number of ECSs” (detecting the number of ECSs).

[0358] The detection condition for 【4.1 Purchasing EVS Cloud Disk】 is “4|0|volume-test|available”. This detection condition template is “4|0|name|status” (checking if a cloud disk with the specified name / status exists).

[0359] The detection condition for 【4.2 Mounting Cloud Disk EVS】 is “4|0|volume-test|in-use”. This detection condition template is “4|0|name|status” (checking if a cloud disk with the specified name / status exists).

[0360] IV. The third sub-model generates cloud resource configurations. The generated content is as follows:

[0361] ECS: CPU - 4 cores, Memory - 16GB, System disk size - 40GB, Quantity - 1.

[0362] EVS: Disk size -20G, quantity -1.

[0363] To avoid wasting cloud resources during the experiment, the experimental platform needs to limit experimental resources based on these core resource specifications. After the third sub-model outputs these resource specifications, it loads these specifications into a specific configuration template:

[0364]

[0365] After receiving a configuration template containing cloud resource specifications, the experimental platform will limit the resources for the experiment based on these configurations. Furthermore, if it is determined that the required cloud resources need to be pre-configured for the experimental task, the experimental platform needs to use these configurations to prepare the specific cloud resources required for the current experiment for the tenant during the experimental phase, avoiding the tenant spending too much time and effort on the environment preparation stage.

[0366] It should be noted that this application can be provided to tenants in various forms. For example, this application can be provided to companies and other enterprises as a business-to-business (2B) product. In this case, the ability to create cloud sandbox experiments can be sold to 2B customers as an API, helping them to automatically build experiments. By calling the API, 2B customers can obtain complete experiment manuals, experiment testing conditions, and cloud resource configuration information based on the basic experiment information. Combined with their existing experiment platform, they can quickly build a complete cloud sandbox experiment. Alternatively, this application can be provided to customers as a customer-to-customer (2C) product. In this case, the ability to create cloud sandbox experiments can be provided to customers by cloud service providers as a service. After the customer inputs basic experiment information, the cloud service provider's cloud management platform can create a cloud sandbox experiment for the customer based on this information. Tenants can then directly enter the experiment environment after accessing the cloud sandbox experiment through the interface. Figure 5 and Figure 6 This document provides an example of how the ability to create cloud sandbox experiments for this application can be offered to customers as a service.

[0367] In summary, in this application, the cloud management platform can use the method provided in this application to automatically obtain the experimental content of the cloud sandbox experiment based on the basic experimental information of the cloud sandbox experiment provided by the tenant, and use the infrastructure to create cloud sandbox experiments for implementing the experimental content. This improves the automation and efficiency of the cloud management platform in creating cloud sandbox experiments, thereby enhancing the learning experience and efficiency of the tenant.

[0368] Furthermore, since cloud sandbox experiments are created based on the basic experimental information provided by the tenant, they can meet the tenant's personalized experimental needs. This enables the cloud management platform to provide cloud sandbox experiments that meet the tenant's personalized experimental needs. It helps the cloud management platform to use this capability to provide personalized cloud sandbox experiments to its many tenants, enriching the types of cloud sandbox experiments that the cloud management platform can provide and improving the flexibility of the cloud sandbox experiments provided by the cloud management platform.

[0369] It should be noted that the order of steps in the method for creating a cloud sandbox experiment provided in this application embodiment can be appropriately adjusted, and steps can also be added or removed as needed. Any variations that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the protection scope of this application, and therefore will not be elaborated further.

[0370] The following describes an example of a virtual device in an embodiment of this application.

[0371] The above describes the method for creating a cloud sandbox experiment according to the embodiments of this application. Corresponding to the above method, the embodiments of this application also provide an apparatus for creating a cloud sandbox experiment. Figure 32 This is a schematic diagram of a device for creating a cloud sandbox experiment provided in an embodiment of this application. Based on Figure 32 The following are several components shown, which Figure 32 The apparatus shown for creating a cloud sandbox experiment is capable of performing the above. Figure 4 All or part of the operations shown. It should be understood that the apparatus may include more additional components than those shown, or omit some of the shown components; this application embodiment does not limit this. Optionally, the apparatus for creating a cloud sandbox experiment may be deployed on a cloud management platform. The cloud management platform is used to manage infrastructure providing at least one cloud service. The infrastructure is used to deploy cloud instances implementing the cloud service. The infrastructure includes at least one cloud data center. Each cloud data center has multiple servers. At least one or any combination of at least one cloud service is deployed on at least one server in the infrastructure. Figure 32 As shown, the apparatus 320 for creating a cloud sandbox experiment may include:

[0372] The interaction module 3201 is used to receive the experiment creation request sent by the tenant. The experiment creation request is used to instruct the cloud management platform to create a cloud sandbox experiment for the tenant. The cloud sandbox experiment is used for the tenant to experience cloud services.

[0373] The interaction module 3201 is also used to provide tenants with an input interface for basic experimental information based on the experimental creation request.

[0374] The interaction module 3201 is also used to obtain the basic experimental information input by the tenant from the input interface. The basic experimental information indicates one or more of the following: the experimental name, experimental objective, experimental content summary, experimental type, or the cloud services involved in the cloud sandbox experiment.

[0375] The processing module 3202 is used to obtain the experimental content of the cloud sandbox experiment based on the basic experimental information. The experimental content includes at least the experimental tasks and implementation steps of the cloud sandbox experiment. The experimental tasks and implementation steps are used to guide tenants on how to operate the cloud sandbox experiment.

[0376] Deployment module 3203 is used to create cloud sandbox experiments that utilize infrastructure to implement experimental content.

[0377] Interaction module 3201 is also used to provide tenants with an operation interface for cloud sandbox experiments.

[0378] In one possible implementation, the experiment content further includes: detection conditions for the experimental task, which indicate the conditions that must be met for the successful completion of the experimental task. The interaction module 3201 is further configured to receive operation instructions sent by the tenant, which instruct the operations required to complete the experimental task in the cloud sandbox experiment; the processing module 3202 is further configured to execute the operations according to the operation instructions for the cloud sandbox experiment and generate a response to the operations; the processing module 3202 is further configured to, upon receiving a detection instruction from the tenant, use the detection conditions to detect whether the experimental task has been successfully completed in response, where the detection instruction instructs the detection of the completion status of the experimental task; the interaction module 3201 is further configured to provide feedback on the detection results to the tenant.

[0379] In one possible implementation, the experiment content also includes: the type of cloud resources required for the experiment task and the specification constraints that the cloud resources need to meet. Then, the interaction module 3201 is further configured to receive a configuration instruction sent by the tenant, which indicates whether the required cloud resources need to be pre-configured for the experiment task; the processing module 3202 is further configured to, if the configuration instruction indicates that the required cloud resources need to be pre-configured for the experiment task, configure the required cloud resources for the experiment task according to the type of cloud resources required for the experiment task and the specification constraints that the cloud resources need to meet as indicated in the experiment content.

[0380] In one possible implementation, the processing module 3202 is specifically used to: generate one or more experimental tasks for the cloud sandbox experiment based on the basic experimental information; generate the implementation steps required to complete each experimental task based on the one or more experimental tasks; and configure target illustrations for the experimental tasks based on the experimental tasks, wherein the target illustrations are used to display the implementation method of the experimental tasks in the form of images.

[0381] In one possible implementation, the experiment content also includes: detection conditions for the experimental task. Detection conditions are used to indicate the conditions that should be met for the successful completion of the experimental task. Then, processing module 3202 is specifically used for: based on the experimental task, obtaining the expected state that should be achieved for the successful completion of the experimental task; based on the expected state, obtaining the target detection type to which the target detection condition used to detect whether the expected state has been reached belongs; based on the target detection type, obtaining the target detection condition template, which is a template that all detection conditions belonging to the target detection type must satisfy; based on the expected state, instantiating the parameters in the target detection condition template to obtain the target detection condition, and using the target detection condition as the detection condition for the experimental task.

[0382] In one possible implementation, the experiment content also includes: the type of cloud resources required for the experiment task and the specification constraints that the cloud resources need to meet. The processing module 3202 is specifically used to: based on the experiment task, obtain the type of cloud resources required to complete the experiment task and the specification constraints that the cloud resources need to meet, and match the type of cloud resources and the specification constraints that the cloud resources need to meet with the experiment task.

[0383] In one possible implementation, the processing module 3202 is further configured to match the topological relationships between cloud services recorded in the cloud management platform based on the cloud services involved in the cloud sandbox experiment, to obtain a cloud service topology map of the cloud sandbox experiment. The cloud service topology map is used to indicate the topological relationships between cloud services involved in the cloud sandbox experiment. The interaction module 3201 is further configured to display the cloud service topology map to the tenant.

[0384] And / or, the processing module 3202 is also used to match the desktop image recorded by the cloud management platform based on the experiment type of the cloud sandbox experiment to obtain the desktop image of the cloud sandbox experiment. The desktop image is used to instruct the tenant's client to display the configuration required by the experiment platform of the cloud sandbox experiment. The interaction module 3201 is also used to provide the desktop image to the tenant.

[0385] In one possible implementation, the processing module 3202 is specifically used to: input basic experimental information into the target large model, receive the experimental content of the cloud sandbox experiment output by the target large model, wherein the target large model is an initial large model obtained through incremental pre-training and fine-tuning training, wherein incremental pre-training refers to using cloud knowledge data and the experimental content of historical cloud sandbox experiments as the first training data to train the initial large model, wherein the cloud knowledge data is used to indicate the cloud services that the cloud management platform can provide, and fine-tuning training refers to using key information extracted from the specified experimental content of historical cloud sandbox experiments as the second training data to train the initial large model that has undergone incremental pre-training.

[0386] In one possible implementation, the second training data includes: a first instruction fine-tuning dataset, which comprises multiple first input-output data pairs. Each first input-output data pair includes: basic information about historical experiments in the historical cloud sandbox experiment, historical experimental tasks and their implementation steps, and attached figures and their descriptions configured for the historical experimental tasks. The basic information about historical experiments serves as the input for training the initial large model sub-model, while the historical experimental tasks and their implementation steps, and the attached figures and their descriptions configured for the historical experimental tasks serve as the output for training the initial large model sub-model.

[0387] In one possible implementation, the second training data further includes: a second instruction fine-tuning dataset and / or a third instruction fine-tuning dataset. The second instruction fine-tuning dataset includes multiple second input-output data pairs, each pair including: historical experimental tasks from historical cloud sandbox experiments, their experimental steps, and detection conditions. The historical experimental tasks and their experimental steps serve as input for training the initial large model sub-model, and the detection conditions of the historical experimental tasks serve as output for training the initial large model sub-model. The third instruction fine-tuning dataset includes multiple third input-output data pairs, each pair including: historical experimental tasks from historical cloud sandbox experiments, their implementation steps, the cloud services used, and the specifications of the cloud services. The historical experimental tasks and their experimental steps serve as input for training the initial large model sub-model, and the cloud services used and their specifications serve as output for training the initial large model sub-model.

[0388] Here, the detailed working process of the interaction module 3201, processing module 3202, and deployment module 3203 is described in the preceding method embodiments. For example, the interaction module 3201 receives the experiment creation request sent by the tenant using the aforementioned step 401, provides the tenant with an input interface for basic experiment information based on the experiment creation request using the aforementioned step 402, obtains the basic experiment information input by the tenant from the input interface using the aforementioned step 403, and provides the tenant with an operation interface for the cloud sandbox experiment using the aforementioned step 404. The processing module 3202 obtains the experiment content of the cloud sandbox experiment based on the basic experiment information using the aforementioned step 404. The deployment module 3203 uses the aforementioned step 404 to create a cloud sandbox experiment for implementing the experiment content using the infrastructure. The implementation process will not be described again in this embodiment.

[0389] The interaction module 3201, processing module 3202, and deployment module 3203 can all be implemented in software or in hardware. For example, the implementation of the interaction module 3201 will be described below. Similarly, the implementation of the processing module 3202 and deployment module 3203 can refer to the implementation of the interaction module 3201.

[0390] As an example of a software functional unit, the interaction module 3201 may include code running on a computing instance. The computing instance may include at least one of a physical host (computing device), a virtual machine, or a container. Further, the aforementioned computing instance may be one or more. For example, the interaction module 3201 may include code running on multiple hosts / virtual machines / containers. It should be noted that the multiple hosts / virtual machines / containers used to run the code may be distributed in the same region or in different regions. Further, the multiple hosts / virtual machines / containers used to run the code may be distributed in the same availability zone (AZ) or in different AZs, each AZ including one cloud data center or multiple geographically proximate cloud data centers. Typically, a region may include multiple AZs.

[0391] Similarly, multiple hosts / virtual machines / containers used to run this code can be distributed within the same Virtual Private Cloud (VPC) or across multiple VPCs. Typically, a VPC is set up within a single region. Communication between two VPCs within the same region, and between VPCs in different regions, requires a communication gateway to be configured within each VPC. Interoperability between VPCs is achieved through this communication gateway.

[0392] As an example of a hardware functional unit, the interaction module 3201 may include at least one computing device, such as a server. Alternatively, the interaction module 3201 may also be a device implemented using an application-specific integrated circuit (ASIC) or a programmable logic device (PLD). The PLD may be implemented using a complex programmable logical device (CPLD), a field-programmable gate array (FPGA), generic array logic (GAL), or any combination thereof.

[0393] The multiple computing devices included in the interaction module 3201 can be distributed in the same region or in different regions. Similarly, the multiple computing devices included in the interaction module 3201 can be distributed in the same Availability Zone (AZ) or in different AZs. Likewise, the multiple computing devices included in the interaction module 3201 can be distributed in the same Virtual Private Cloud (VPC) or in multiple VPCs. These multiple computing devices can be any combination of computing devices such as servers, ASICs, PLDs, CPLDs, FPGAs, and GALs.

[0394] It should be noted that, in other embodiments, any one of the interaction module 3201, processing module 3202, and deployment module 3203 can be used to execute any step in the method for creating a cloud sandbox experiment. The steps implemented by the interaction module 3201, processing module 3202, and deployment module 3203 can be specified as needed. By implementing different steps in the method for creating a cloud sandbox experiment through the interaction module 3201, processing module 3202, and deployment module 3203, all functions of the device for creating a cloud sandbox experiment can be achieved.

[0395] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of each component described above can be referred to the corresponding content in the foregoing method embodiments, and will not be repeated here.

[0396] The following provides examples illustrating the basic hardware structures involved in the embodiments of this application.

[0397] This application also provides a computing device 3300. For example... Figure 33 As shown, the computing device 3300 includes a bus 3302, a processor 3304, a memory 3306, and a communication interface 3308. The processor 3304, the memory 3306, and the communication interface 3308 communicate with each other via the bus 3302. The computing device 3300 can be a server or a terminal device. It should be understood that this application does not limit the number of processors and memories in the computing device 3300.

[0398] The 3302 bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 33The bus 3302 may be represented by a single line, but this does not mean that there is only one bus or one type of bus. The bus 3302 may include a path for transmitting information between various components of the computing device 3300 (e.g., memory 3306, processor 3304, communication interface 3308).

[0399] The processor 3304 may include any one or more processors such as a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP), or a digital signal processor (DSP).

[0400] The memory 3306 may include volatile memory, such as random access memory (RAM). The processor 3304 may also include non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid state drive (SSD).

[0401] The memory 3306 stores executable program code, and the processor 3304 executes the executable program code to implement the functions of the aforementioned interaction module 3201, processing module 3202, and deployment module 3203, thereby realizing the method for creating a cloud sandbox experiment. That is, the memory 3306 stores instructions for executing the method for creating a cloud sandbox experiment.

[0402] The communication interface 3308 uses transceiver modules, such as, but not limited to, network interface cards and transceivers, to enable communication between the computing device 3300 and other devices or communication networks.

[0403] This application also provides a computing device cluster. The computing device cluster includes at least one computing device. The computing device can be a server, such as a central server, an edge server, or a local server in a local data center. In some embodiments, the computing device can also be a terminal device such as a desktop computer, a laptop computer, or a smartphone.

[0404] like Figure 34 As shown, the computing device cluster includes at least one computing device 3300. The memory 3306 in one or more computing devices 3300 in the computing device cluster may store the same instructions for executing methods to create cloud sandbox experiments.

[0405] In some possible implementations, the memory 3306 of one or more computing devices 3300 in the computing device cluster may also store partial instructions for executing the method of creating a cloud sandbox experiment. In other words, a combination of one or more computing devices 3300 can jointly execute the instructions for executing the method of creating a cloud sandbox experiment.

[0406] It should be noted that the memory 3306 in different computing devices 3300 within the computing device cluster can store different instructions, each used to execute a portion of the functions of the device for creating the cloud sandbox experiment. That is, the instructions stored in the memory 3306 of different computing devices 3300 can implement the functions of one or more modules among the interaction module 3201, processing module 3202, and deployment module 3203.

[0407] In some possible implementations, one or more computing devices in a computing device cluster can be connected via a network. This network can be a wide area network (WAN), a local area network (LAN), or similar. Figure 35 One possible implementation is shown. For example... Figure 35 As shown, the two computing devices 3300A and 3300B are connected via a network. Specifically, they are connected to the network through communication interfaces in each computing device. In this possible implementation, the memory 3306 in computing device 3300A stores instructions for executing the functions of the interaction module 3201 and the processing module 3202. Meanwhile, the memory 3306 in computing device 3300B stores instructions for executing the functions of the deployment module 3203.

[0408] It should be understood that Figure 35 The functions of the computing device 3300A shown can also be performed by multiple computing devices 3300. Similarly, the functions of the computing device 3300B can also be performed by multiple computing devices 3300.

[0409] This application also provides another computing device cluster. The connection relationships between the computing devices in this computing device cluster can be similarly referred to... Figure 34 and Figure 35 The connection method of the computing device cluster. The difference is that the memory 3306 of one or more computing devices 3300 in the computing device cluster can store the same instructions for executing the method of creating cloud sandbox experiments.

[0410] In some possible implementations, the memory 3306 of one or more computing devices 3300 in the computing device cluster may also store partial instructions for executing the method of creating a cloud sandbox experiment. In other words, a combination of one or more computing devices 3300 can jointly execute the instructions for executing the method of creating a cloud sandbox experiment.

[0411] This application also provides a computer program product containing instructions. The computer program product may be a software or program product containing instructions capable of running on a computing device or stored on any usable medium. When the computer program product is run on at least one computing device, it causes the at least one computing device to perform a method for creating a cloud sandbox experiment.

[0412] This application also provides a computer-readable storage medium. The computer-readable storage medium can be any available medium that a computing device can store, or a data storage device such as a data center containing one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive). The computer-readable storage medium includes instructions that instruct the computing device to perform a method for creating a cloud sandbox experiment, or instruct the computing device to perform a method for creating a cloud sandbox experiment.

[0413] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0414] It should be noted that all information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.), and signals involved in this application have been authorized by the user or fully authorized by all parties, and the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions. For example, the raw data and executable code involved in this application were obtained with full authorization.

[0415] In the embodiments of this application, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance. The term "at least one" refers to one or more, and the term "multiple" refers to two or more, unless otherwise expressly defined.

[0416] In this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0417] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the protection scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for creating a cloud sandbox experiment, characterized in that, The method is executed by a cloud management platform, which manages the infrastructure providing cloud services and deploys cloud instances that implement the cloud services. The method includes: The cloud management platform receives an experiment creation request sent by a tenant. The experiment creation request is used to instruct the cloud management platform to create a cloud sandbox experiment for the tenant. The cloud sandbox experiment is used for the tenant to experience the cloud service. Based on the experiment creation request, the cloud management platform provides the tenant with an input interface for basic experiment information; The cloud management platform obtains the basic experimental information input by the tenant from the input interface. The basic experimental information indicates one or more of the following: the experimental name, experimental objective, experimental content summary, experimental type, or the cloud services involved in the cloud sandbox experiment. Based on the basic experimental information, the cloud management platform obtains the experimental content of the cloud sandbox experiment. The experimental content includes at least the experimental tasks and implementation steps of the cloud sandbox experiment. The experimental tasks and implementation steps are used to guide the tenant in operating the cloud sandbox experiment. The cloud management platform utilizes the infrastructure to create the cloud sandbox experiment for implementing the experimental content, and provides the tenant with the operation interface of the cloud sandbox experiment.

2. The method as described in claim 1, characterized in that, The experiment also includes: detection conditions for the experimental task, wherein the detection conditions are used to indicate the conditions that should be met for the successful completion of the experimental task; the method further includes: The cloud management platform receives operation instructions sent by the tenant, which are used to instruct the cloud sandbox experiment to perform the operations required to complete the experimental task. The cloud management platform executes operations on the cloud sandbox experiment according to the operation instructions and generates a response to the operation; Upon receiving a detection instruction from the tenant, the cloud management platform uses the detection conditions to detect whether the experimental task has been successfully completed in response, and then feeds back the detection result to the tenant. The detection instruction is used to instruct the completion status of the experimental task to be detected.

3. The method as described in claim 1 or 2, characterized in that, The experiment also includes: the types of cloud resources required for the experimental task and the specification constraints that the cloud resources need to meet; the method further includes: The cloud management platform receives configuration instructions sent by the tenant, which are used to indicate whether the required cloud resources need to be pre-configured for the experimental task. When the configuration instruction indicates that the required cloud resources need to be pre-configured for the experimental task, the cloud management platform configures the required cloud resources for the experimental task according to the type of cloud resources required by the experimental task as indicated by the experimental content and the specification restrictions that the cloud resources need to meet.

4. The method according to any one of claims 1 to 3, characterized in that, Based on the basic experimental information, the cloud management platform obtains the experimental content of the cloud sandbox experiment, including: Based on the basic experimental information, the cloud management platform generates one or more experimental tasks for the cloud sandbox experiment. The cloud management platform generates the implementation steps required to complete each experimental task based on the one or more experimental tasks. Based on the experimental task, the cloud management platform configures a target diagram for the experimental task, which is used to illustrate the implementation method of the experimental task in the form of an image.

5. The method as described in claim 4, characterized in that, The experimental content also includes: the detection conditions of the experimental task, which are used to indicate the conditions that the experimental task should meet for successful completion; the cloud management platform obtains the experimental content of the cloud sandbox experiment based on the basic experimental information, and also includes: Based on the experimental task, the cloud management platform obtains the expected state that the experimental task should be achieved upon successful completion. Based on the desired state, the cloud management platform obtains the target detection type to which the target detection condition used to detect whether the desired state has been reached belongs; Based on the target detection type, the cloud management platform obtains the target detection condition template for the target detection conditions. The target detection condition template is a template that all detection conditions belonging to the target detection type must satisfy. Based on the desired state, the cloud management platform instantiates the parameters in the target detection condition template to obtain the target detection conditions, and uses the target detection conditions as the detection conditions for the experimental task.

6. The method as described in claim 4 or 5, characterized in that, The experimental content also includes: the types of cloud resources required for the experimental task and the specification constraints that the cloud resources need to meet. Based on the basic experimental information, the cloud management platform obtains the experimental content of the cloud sandbox experiment, which also includes: Based on the experimental task, the cloud management platform obtains the types of cloud resources required to complete the experimental task and the specification restrictions that the cloud resources need to meet, and the types of cloud resources and the specification restrictions that the cloud resources need to meet match the experimental task.

7. The method as described in any one of claims 1 to 6, characterized in that, The method further includes: Based on the cloud services involved in the cloud sandbox experiment, the cloud management platform matches the topological relationships between the cloud services recorded on the cloud management platform to obtain the cloud service topology map of the cloud sandbox experiment, and displays the cloud service topology map to the tenant. The cloud service topology map is used to indicate the topological relationships between the cloud services involved in the cloud sandbox experiment. And / or, the cloud management platform matches the desktop images recorded by the cloud management platform based on the experiment type of the cloud sandbox experiment to obtain the desktop image of the cloud sandbox experiment, and provides the desktop image to the tenant. The desktop image is used to instruct the tenant's client to display the configuration required by the experiment platform of the cloud sandbox experiment.

8. The method according to any one of claims 1 to 7, characterized in that, Based on the basic experimental information, the cloud management platform obtains the experimental content of the cloud sandbox experiment, including: The cloud management platform inputs the basic experimental information into the target large model and receives the experimental content of the cloud sandbox experiment output by the target large model. The target large model is an initial large model obtained through incremental pre-training and fine-tuning training. The incremental pre-training refers to using cloud knowledge data and the experimental content of historical cloud sandbox experiments as the first training data to train the initial large model. The cloud knowledge data is used to indicate the cloud services that the cloud management platform can provide. The fine-tuning training refers to using key information extracted from the specified experimental content of historical cloud sandbox experiments as the second training data to train the initial large model after incremental pre-training.

9. The method as described in claim 8, characterized in that, The second training data includes: a first instruction fine-tuning dataset, which includes multiple first input-output data pairs. Each first input-output data pair includes: basic information of historical experiments in the historical cloud sandbox experiment, historical experimental tasks and their implementation steps, and attached figures configured for the historical experimental tasks and their descriptions. The basic information of the historical experiments is used as input when training the initial large model sub-model, and the historical experiment tasks and their implementation steps, the attached drawings configured for the historical experiment tasks and their descriptions are used as output when training the initial large model sub-model.

10. The method as described in claim 9, characterized in that, The second training data also includes: a second instruction fine-tuning dataset and / or a third instruction fine-tuning dataset; The second instruction fine-tuning dataset includes multiple second input-output data pairs. Each second input-output data pair includes: historical experimental tasks and their experimental steps and detection conditions from historical cloud sandbox experiments. The historical experimental tasks and their experimental steps are used as inputs for training the initial large model sub-model, and the detection conditions of the historical experimental tasks are used as outputs for training the initial large model sub-model. The third instruction fine-tuning dataset includes multiple third input-output data pairs, each of which includes: historical experimental tasks and their implementation steps from historical cloud sandbox experiments, the cloud services used, and the specifications of the cloud services. The historical experimental tasks and their experimental steps are used as inputs for training the initial large model sub-model, and the cloud services and cloud service specifications used in the historical experimental tasks are used as outputs for training the initial large model sub-model.

11. An apparatus for creating a cloud sandbox experiment, characterized in that, The device is deployed on a cloud management platform, which manages the infrastructure providing cloud services. The infrastructure is used to deploy cloud instances that implement cloud services. The device includes: An interaction module is used to receive an experiment creation request sent by a tenant. The experiment creation request is used to instruct the cloud management platform to create a cloud sandbox experiment for the tenant. The cloud sandbox experiment is used for the tenant to experience the cloud service. The interaction module is also used to provide the tenant with an input interface for basic experimental information based on the experiment creation request; The interaction module is also used to obtain the basic experimental information input by the tenant from the input interface. The basic experimental information indicates one or more of the following: the experimental name, experimental objective, experimental content summary, experimental type, or the cloud service involved in the cloud sandbox experiment. The processing module is used to obtain the experimental content of the cloud sandbox experiment based on the basic experimental information. The experimental content includes at least the experimental tasks and implementation steps of the cloud sandbox experiment. The experimental tasks and implementation steps are used to guide the tenant on how to operate the cloud sandbox experiment. A deployment module is used to create the cloud sandbox experiment for implementing the experimental content using the infrastructure; The interaction module is also used to provide the tenant with the operation interface for the cloud sandbox experiment.

12. The apparatus as claimed in claim 11, characterized in that, The experiment also includes: detection conditions for the experimental task, which are used to indicate the conditions that the experimental task should meet for successful execution. The interaction module is also used to receive operation instructions sent by the tenant, the operation instructions being used to instruct the operations required to complete the experimental task for the cloud sandbox experiment; The processing module is also configured to perform operations on the cloud sandbox experiment according to the operation instructions, and generate a response to the operation; The processing module is further configured to, when the interaction module receives a detection instruction sent by the tenant, use the detection conditions to detect whether the experimental task has been successfully completed in response to the response, wherein the detection instruction is used to instruct the completion status of the experimental task to be detected; The interaction module is also used to provide feedback on the detection results to the tenant.

13. The apparatus as claimed in claim 11 or 12, characterized in that, The experiment also includes: the types of cloud resources required for the experimental task and the specification limitations that the cloud resources need to meet; The interaction module is also used to receive configuration instructions sent by the tenant, the configuration instructions being used to indicate whether the required cloud resources need to be pre-configured for the experimental task; The processing module is further configured to configure the required cloud resources for the experimental task according to the type of cloud resources required by the experimental task as indicated by the experimental content and the specification constraints that the cloud resources need to meet, when the configuration instruction indicates that the required cloud resources need to be pre-configured for the experimental task.

14. The apparatus as claimed in any one of claims 11 to 13, characterized in that, The processing module is specifically used for: Based on the basic experimental information, one or more experimental tasks for the cloud sandbox experiment are generated. Based on the one or more experimental tasks, generate the implementation steps required to complete each experimental task; Based on the experimental task, a target diagram is configured for the experimental task, which is used to illustrate the implementation of the experimental task in the form of an image.

15. The apparatus as claimed in claim 14, characterized in that, The experiment also includes: detection conditions for the experimental task, which indicate the conditions that should be met for the successful completion of the experimental task; and the processing module, specifically used for: Based on the experimental task, obtain the expected state that should be achieved when the experimental task is successfully completed. Based on the desired state, obtain the target detection type to which the target detection condition used to detect whether the desired state has been reached belongs; Based on the target detection type, obtain the target detection condition template for the target detection conditions. The target detection condition template is a template that all detection conditions belonging to the target detection type must satisfy. Based on the desired state, the parameters in the target detection condition template are instantiated to obtain the target detection conditions, which are then used as the detection conditions for the experimental task.

16. The apparatus as claimed in claim 14 or 15, characterized in that, The experiment also includes: the type of cloud resources required for the experimental task and the specification constraints that the cloud resources need to meet; the processing module is specifically used for: Based on the experimental task, the types of cloud resources required to complete the experimental task and the specification constraints that the cloud resources need to meet are obtained, and the types of cloud resources and the specification constraints that the cloud resources need to meet are matched with the experimental task.

17. The apparatus according to any one of claims 11 to 16, characterized in that, The processing module is further configured to match the topological relationships between cloud services recorded in the cloud management platform based on the cloud services involved in the cloud sandbox experiment, and obtain a cloud service topology map of the cloud sandbox experiment. The cloud service topology map is used to indicate the topological relationships between cloud services involved in the cloud sandbox experiment. The interaction module is also used to display the cloud service topology diagram to the tenant; And / or, the processing module is further configured to match the desktop image recorded by the cloud management platform based on the experiment type of the cloud sandbox experiment to obtain the desktop image of the cloud sandbox experiment, the desktop image being used to instruct the tenant's client to display the configuration required by the experiment platform of the cloud sandbox experiment; The interaction module is also used to provide the desktop image to the tenant.

18. The apparatus as claimed in any one of claims 11 to 17, characterized in that, The processing module is specifically used for: The basic experimental information is input into the target large model, and the experimental content of the cloud sandbox experiment output by the target large model is received. The target large model is an initial large model obtained through incremental pre-training and fine-tuning training. The incremental pre-training refers to using cloud knowledge data and the experimental content of historical cloud sandbox experiments as the first training data to train the initial large model. The cloud knowledge data is used to indicate the cloud services that the cloud management platform can provide. The fine-tuning training refers to using key information extracted from the specified experimental content of historical cloud sandbox experiments as the second training data to train the initial large model after incremental pre-training.

19. The apparatus as claimed in claim 18, characterized in that, The second training data includes: a first instruction fine-tuning dataset, which includes multiple first input-output data pairs. Each first input-output data pair includes: basic information of historical experiments in the historical cloud sandbox experiment, historical experimental tasks and their implementation steps, and attached figures configured for the historical experimental tasks and their descriptions. The basic information of the historical experiments is used as input when training the initial large model sub-model, and the historical experiment tasks and their implementation steps, the attached drawings configured for the historical experiment tasks and their descriptions are used as output when training the initial large model sub-model.

20. The apparatus as claimed in claim 19, characterized in that, The second training data also includes: a second instruction fine-tuning dataset and / or a third instruction fine-tuning dataset; The second instruction fine-tuning dataset includes multiple second input-output data pairs. Each second input-output data pair includes: historical experimental tasks and their experimental steps and detection conditions from historical cloud sandbox experiments. The historical experimental tasks and their experimental steps are used as inputs for training the initial large model sub-model, and the detection conditions of the historical experimental tasks are used as outputs for training the initial large model sub-model. The third instruction fine-tuning dataset includes multiple third input-output data pairs, each of which includes: historical experimental tasks and their implementation steps from historical cloud sandbox experiments, the cloud services used, and the specifications of the cloud services. The historical experimental tasks and their experimental steps are used as inputs for training the initial large model sub-model, and the cloud services and cloud service specifications used in the historical experimental tasks are used as outputs for training the initial large model sub-model.

21. A computing device cluster, characterized in that, The system includes multiple computing devices, each comprising multiple processors and multiple memories, the multiple memories storing program instructions, and the multiple processors executing the program instructions to cause the cluster of computing devices to perform the method as described in any one of claims 1 to 10.

22. A computer-readable storage medium, characterized in that, Includes program instructions that, when executed on a computing device, cause the computing device to perform the method as described in any one of claims 1 to 10.

23. A computer program product containing instructions, characterized in that, When the instruction is executed by the computing device cluster, the computing device cluster causes the computing device cluster to perform the method as described in any one of claims 1 to 10.

Images