Abstract: A cloud service providing system is described. The cloud service providing system includes a first host that runs an active virtual instance, where the active virtual instance is used for providing a cloud service for a user. The cloud service providing system also include a second host that runs a first standby virtual instance. The first host stores, in a first shared memory during running of the active virtual instance, memory data needed for running the active virtual instance, and the second host runs, when the first host fails, the first standby virtual instance based on the memory data obtained from the first shared memory, to cause the first standby virtual instance to provide a cloud service for the user. This application improves timeliness and efficiency of hot standby of memory data, provides a real-time redundancy backup capability, and helps ensure performance of services provided based on backed-up memory data.

Abstract: A data processing method applied to the data warehouse system includes that a coordinator node allocates a data write request of a data table to a first computing node, and the first computing node checks whether target data corresponding to the data write request meets a constraint condition. If the check succeeds, a metadata cluster checks the target data again. If the check performed by the metadata cluster also succeeds, a cloud storage cluster writes the target data into the data table. It can be learned that the data warehouse system uses a two-layer indexing mechanism and the metadata cluster of a cloud native architecture to check whether data to be written into the data table meets the constraint condition. Therefore, when a data constraint capability is provided, data reliability is ensured, and that system running performance is not affected is also ensured.

Abstract: A cloud service system includes a first server and a physical device. The first server and the physical device are coupled through a physical channel. The first server includes a cloud instance. The physical device includes a trusted execution environment corresponding to the cloud instance. The physical device is configured to perform confidential calculation on data from the cloud instance through the trusted execution environment, and send a calculation result to the cloud instance.

Abstract: This application discloses a method for providing a container service, a container system, and a related apparatus, relating to the field of container technologies. The method includes: The container system provides base images and a public image for users, creates a container by using the base image, and mounts the public image. The public image does not need to be downloaded, thereby reducing occupation on storage space of a node on which the container is located. In addition, the shared public image exists independently, thereby facilitating management and quick update or upgrade for the public image. Because the public image is mounted, simply updating the public image can quickly update the container created for the users. There is no need to download the public image, and the mount-based approach also makes container creation more efficiently.

Abstract: In the method, a data exchange platform obtains status data that is of a first object and that is sent by a first device. The first object is an environmental object in a real scene or an item included in the real scene. The data exchange platform sends the status data of the first object to a second device. The second device obtains the status data of the first object, and adjusts, based on the status data of the first object, a second object that is included in a virtual scene and that corresponds to the first object.

Abstract: A method, wherein automatic execution of a series of complex operations including at least one AI operation and at least one data processing operation may be triggered by using a table operation instruction. A user only needs to input a simple table operation instruction, and does not need to input a series of complex operation instructions.

Abstract: A method for training a semantic vector-based information retrieval model in the field of artificial intelligence includes: obtaining a first training data set and a second training data set, the first training data includes a first query text, a first retrieval text corresponding to the first query text and having a positive label, and a second retrieval text corresponding to the first query text and having a negative label; updating a first information retrieval model based on the first training data set, to obtain a second information retrieval model; and updating the second information retrieval model based on the second training data set, to obtain a third information retrieval model.

Abstract: This disclosure provides an object storage system including a service layer, a persistence layer, and a metadata layer. The service layer receives a batch upload request sent by a client, where the batch upload request indicates a plurality of to-be-uploaded first objects; sends data of the plurality of first objects to the persistence layer; and generates metadata of the plurality of first objects, and sends the metadata of the plurality of first objects to the metadata layer. The persistence layer stores the data of the plurality of first objects. The metadata layer stores the metadata of the plurality of first objects into the persistence layer; and determines that metadata of at least one first object fails to be stored into the persistence layer, and sends first indication information to the service layer, where the first indication information indicates to perform garbage collection on the plurality of first objects.

Abstract: A check rule processing method includes: obtaining to-be-checked code and/or related information about the to-be-checked code, where the related information about the to-be-checked code includes at least one of the following: a programming language of the to-be-checked code, a development framework of the to-be-checked code, a deployment form of the to-be-checked code, a service scope of the to-be-checked code, a third-party library used in the to-be-checked code, a coding style of the to-be-checked code, or description information of a target project to which the to-be-checked code belongs; determining a target rule set from a static analysis rule library based on the to-be-checked code and/or the related information about the to-be-checked code, where the target rule set includes one or more candidate check rules, and the one or more candidate check rules are used to perform static analysis on the to-be-checked code; and performing control to display the target rule set.

Abstract: A code execution method includes: a pre-preparation service module obtaining a pre-preparation policy set by a user; the pre-preparation service module downloading, based on the pre-preparation policy, an updated code file included in an updated code version from a code hosting platform to a storage cluster; the pre-preparation service module receiving a code execution request; the pre-preparation service module determining a storage address of code content corresponding to the code execution request in the storage cluster and sending the storage address to an executor; the executor obtaining the code content based on the received storage address and executing the obtained code content.

Abstract: A function node migration method includes obtaining access related information of at least one call chain, where the access related information includes at least one of the following information: a delay of each of N intervals included in the at least one call chain, and a quantity of interaction times between function nodes in each of the N intervals, where N is a positive integer greater than or equal to 1; determining, based on the access related information, P key nodes from a plurality of function nodes that form the N intervals; and migrating the P key nodes. Accordingly, a to-be-migrated function node is determined based on actual data of the call chain. Therefore, the to-be-migrated function node meets a migration requirement.

Abstract: The cloud management system is configured to manage an infrastructure on which a first service of a first tenant and a second service of a second tenant are configured. The cloud management system obtains an authorization request of the first tenant which includes a tenant identifier of the second tenant, an authorization name, and a permission for the second service to invoke the first service. The cloud management system creates the permission based on the authorization request; obtains a permission obtaining request that is inputted by the second tenant and that includes the authorization name and a tenant identifier of the first tenant, configures, based on the permission obtaining request, the second tenant to obtain the permission; and allows, based on the permission when the second tenant configures the second service to invoke the first service, the second service to invoke the first service.

Abstract: A cache replacement method includes a compute node in a distributed system that sends a first read operation request to a memory node. Then, the compute node separately calculates popularities of the plurality of data objects based on the access information of the plurality of data objects using a first cache replacement algorithm in a plurality of cache replacement algorithms. Finally, the compute node sends a cache replacement request to the memory node based on the popularities of the plurality of data objects, to indicate to replace cold data in the plurality of data objects with a data object to be inserted into the memory node. In this way, the compute node executes the cache replacement algorithm based on global access information.

Abstract: A network model training method includes providing a parameter configuration user interface; obtaining a parameter count and a type of a target network model from the parameter configuration user interface; determining training configuration information of the target network model based on the parameter count and the type of the target network model, where a training process of the target network model is divided into a plurality of training phases, different training phases correspond to different model parameter counts, the model parameter counts of the plurality of training phases increase sequentially with the order of training, and training the target network model based on the configuration information of the plurality of training phases.

Abstract: A permission check method is disclosed. A node receives an operation request, the operation request includes a target identifier. The node determines an operation permission of the operation request based on a file ACL of the target identifier and a file user identifier associated with the operation request, and determines a permission check result based on the operation permission of the operation request and a target operation type included in the operation request.

Abstract: This application provides a mesh model generation method and apparatus, and a device. The method includes: obtaining visual data that is of an object and that is uploaded by a user, where the visual data includes at least one of an image, a point cloud, and a voxel model; performing geometric feature extraction on the visual data of the object, to obtain a first geometric feature of the object; obtaining, from a prior knowledge base, prior knowledge corresponding to an object category to which the object belongs, where the prior knowledge is used to guide establishment of a geometric generation network; establishing a first geometric generation network based on the first geometric feature and the prior knowledge corresponding to the object category to which the object belongs; and generating a first mesh model of the object based on the first geometric generation network.

Abstract: A software module running method includes obtaining a running request for a first software module, where the first software module is associated with a first user, the running request for the first software module is to run the first software module that has been deployed in a first computing instance set, or deploy the first software module in a provisioned first computing instance set for running, or provision a first computing instance set for the first software module and run the first software module by using the first computing instance set, and the first computing instance set includes at least one computing instance; and obtaining a first cloud resource authorization specification corresponding to the first software module.

Abstract: This application discloses a method and an apparatus for hybrid deployment across runtime instances based on a cloud computing technology, and belongs to the field of cloud technologies. In this method, a resource scheduling system manages resources of a plurality of hybrid nodes. Because the hybrid node supports deployment of a plurality of types of runtime instances, the resource scheduling system can deploy different types of runtime instances on the hybrid node. Specifications of different types of runtime instances differ greatly, and fragmented resources of the hybrid node can be allocated to a runtime instance with a small specification, so that the fragmented resources of the hybrid node can be fully utilized.

Abstract: A communication method includes a first dynamic host configuration protocol (DHCP) server that allocates a plurality of internet protocol (IP) addresses to an elastic network interface of a first elastic cloud server (ECS), and is configured based on first configuration information of a tenant.

Abstract: This application discloses a point cloud data processing method. The method includes: obtaining an original point cloud data set of a scene, where the scene includes a target object; augmenting point cloud data of a target object, to generate an augmented point cloud data set; training a first model by using the augmented point cloud data set, to obtain a trained first model; training a second model by using the at least one original point cloud data set and the trained first model; and obtaining a trained second model. An original point cloud is augmented to obtain a dense point cloud, the dense point cloud and the original point cloud are separately used for training to obtain a model, and inter-model training is performed, allowing a target model to have a capability of extracting high-quality information from sparse point clouds and executing point cloud processing tasks.