Abstract: A storage device operation method is applied to the physical server. A storage card is inserted into the physical server. When a virtual machine on the physical server receives a data read/write instruction, the virtual machine generates an I/O request for a physical storage device, and sends the I/O request to a client in the storage card; the client transmits the I/O request to a server, so that the server performs an I/O operation on the physical storage device based on the I/O request. Since the virtual machine directly sends the I/O request to the client in the storage card without using a hypervisor in the physical server, computing resources of the physical server are conserved.

Abstract: Methods and systems of managing a resource in a distributed resource management system can include: receiving a resource request by at least one processor in the distributed resource management system, the resource request identifying a requested resource type corresponding to at least one of: a class identifier identifying a resource class assigned to a composite resource, and a class identifier identifying at least one additional resource associated with the composite resource; determining availability of the requested resource type; and scheduling a workload associated with the resource request for execution based on the determination.

Abstract: An apparatus is disclosed. The apparatus may be implemented in a database node or a storage node and includes one or more processors and memory storing instructions for causing the processor to perform a number of operations. Responsive to a page write request, the apparatus determines an identifier corresponding to the requested page, sends component blocks corresponding to the page to a storage node for appending to an append-only storage log of the storage node, receives the physical location of the stored component blocks, and associates the physical storage location of the stored component blocks with the logical identifier corresponding to the page.

Abstract: A redundancy method includes that a first disaster management function (DMF) device on a first site side receives a first request including identification information of a first virtual machine (VM) and a recovery point objective (RPO), allocates a maximum allowable delay time to each node that input/output (IO) data of the first VM passes through in a redundancy process, and sends a second request to a second DMF device on a second site side. The second request includes a maximum allowable delay time of a second replication gateway function (RGF) device on the second site side, and a maximum allowable delay time of an IO writer function (IOWF) device on the second site side and requests the second site side to perform redundancy on the first VM. Hence, the RPO requirements of the tenants can be satisfied in an entire redundancy process.

Abstract: A communication mode selection method includes transmitting, in a first communication mode, a vehicle-to-everything (V2X) service between a leading vehicle and a following vehicle that are in a vehicle platoon, obtaining, signal quality of the first communication mode, when the signal quality of the first communication mode meets a first preset condition, sending, a configuration instruction to the following vehicle, where the configuration instruction is used to instruct the following vehicle to communicate with the leading vehicle in a second communication mode, where the second communication mode is a communication mode in which a network device performs forwarding.

Abstract: Embodiments of this application provide a data processing method in a database. A plurality of aggregation intervals are selected, wherein the plurality of aggregation intervals are different from each other. And at least one piece of data in the database is aggregated separately based on each aggregation interval, so that data clusters stored in the database have data ranges with different attribute value spans.

Abstract: Systems and methods for controlling management operations and shared memory space are disclosed. A cloud cache management controller may receive multiple sets of service attributes. Each set of the multiple sets of service attributes may be related to a cloud cache service instance. The cloud cache management controller may receive a first cloud cache management request. The cloud cache management request may comprise a cloud cache management operation. The cloud cache management controller may retrieve a set of service attributes from the multiple sets of service attributes based on an evaluation of the cloud cache management operation. The cloud cache management controller may send the first cloud cache management request to a corresponding CCSI based on a priority value for the first cloud cache management request calculated based on the retrieved set of service attributes.

Abstract: A vehicle warning method, to avoid misjudgment caused by factors such as map precision and lane infringement of a vehicle, thereby improving warning accuracy. The method includes: A vehicle-to-everything V2X device receives a V2X message corresponding to a first vehicle. The V2X device determines, based on the V2X message, whether the first vehicle is in a warning area. The warning area is determined based on vehicle information of a second vehicle. The vehicle information includes at least position information of the second vehicle and a vehicle size of the second vehicle. The V2X device outputs warning information corresponding to the first vehicle to the second vehicle.

Abstract: A source mobile edge computing (MEC) platform sends first indication information to an application server (AS) when determining that a target application is to stop processing service data of user equipment (UE), wherein the first indication information indicates that the service data of the UE is to be processed in the AS. The target application is deployed on the source MEC platform to process the service data of the UE. The AS is configured to respond to the first indication information and send a first response message to the source MEC platform, wherein the first response message indicates that the AS is ready to process the service data of the UE.

Abstract: A method comprising receiving a platform as a service (PaaS) operation request sent by a first real tenant, where the PaaS operation request is used to indicate a PaaS operation intended for a PaaS resource, determining information about a first virtual tenant corresponding to the first real tenant, and controlling, based on the PaaS operation request and the information about the first virtual tenant, an infrastructure as a service (IaaS) management system to process an IaaS operation corresponding to the PaaS operation. According to the method for providing a PaaS service in this application, one corresponding virtual tenant is introduced for each real tenant, and an infrastructure resource is managed using the virtual tenant.

Abstract: In a service deployment method performed by a function management platform that includes a function scheduler and a function executor manager, the function scheduler receives a function call request that carries a function identifier, selects an available host based on resource pool information of a tenant, and sends a function instance creation request to the function executor manager. The function executor manager receives the function instance creation request, selects the host corresponding to the host identifier from the resource pool of the tenant, and creates a function executor corresponding to the function identifier on the host.

Abstract: Methods, systems and media for deep neural network interpretation via rule extraction. The interpretation of the deep neural network is based on extracting one or more rules approximating classification behavior of the network. Rules are defined by identifying a set of hyperplanes through the data space that collectively define a convex polytope that separates a target class of input samples from input samples of different classes. Each rule corresponds to a set of decision boundaries between two different decision outcomes. Human-understandable representations of rules may be generated. One or more rules may be used to generate a classifier. The representations and interpretations exhibit faithfulness, robustness, and comprehensiveness relative to other known approaches.

Abstract: A method for processing a cloud service in a cloud system, wherein the cloud system includes a hosted cloud, an allied cloud, and an alliance management apparatus, wherein the hosted cloud includes a first region, the allied cloud includes a second region, the hosted cloud further provides a virtual region for a user, and the virtual region is mapped by the second region on the hosted cloud. The alliance management apparatus enables a user of the hosted cloud to obtain unified experience when the user accesses the first region and the virtual region in the cloud system, for example, unified experience in control interface access, and unified experience in application programming interface (API) access.

Abstract: A virtual machine migration method and apparatus are provided, which pertain to the field of computer technologies. The method includes: obtaining a first mapping relationship, pre-stored on a source cloud platform, between a source virtual machine (VM) and a specified user when data in the source VM is migrated to a destination VM; configuring the destination VM and the specified user to form a second mapping relationship according to the obtained first mapping relationship; and storing the second mapping relationship on a destination cloud platform. Therefore, a user that has registered on the source VM can log in to a virtual desktop corresponding to the destination VM and does not need to register on the destination VM again.

Abstract: In an object detection model training method, a classifier that has been trained in a first phase is duplicated to at least two copies, and in a training in a second phase, each classifier obtained through duplication is configured to detect to-be-detected objects with different sizes, and train an object detection model based on a detection result.

Abstract: A method, an apparatus, and a server for managing an image across cloud servers relate to the field of computer technologies. The method includes: obtaining layered image information of a layered image file of a service virtual machine on a second cloud server; creating a boot partition and a rootfs partition based on a pre-created installer virtual machine and the layered image information; and creating, based on the boot partition and the rootfs partition, an image corresponding to the service virtual machine. In this method, no manual operation is required, and not only reducing management costs, but also improving management efficiency.

Abstract: A data transmitting method includes obtaining a plurality of first objects, where each of the first objects includes an amount of data that is less than or equal to a first data amount threshold, encapsulating the first objects to generate a second object, where the second object includes the first objects, and sending the second object to a server in an object storage system.

Abstract: A border gateway receives a first overlay packet sent by a remote acceleration gateway, where a first service request packet is encapsulated in the first overlay packet, a source Internet Protocol (IP) address of the first service request packet is an IP address of a client, and a destination IP address is a public IP address associated with a virtual machine; the border gateway decapsulates the first overlay packet to obtain the first service request packet, encapsulates the first service request packet to generate a second overlay packet, and sends the second overlay packet to a virtual forwarding device, implementing transparent transmission of a source address of a service request packet.

Abstract: The present disclosure provides a system for detecting and preventing the intrusion of malicious data flows in a software defined network (SDN). The system comprises at least one data storage or memory, configured to store flow states of data flows, and to share and update the flow states across the system, at least one shared-state forwarding element (FE) configured to block, forward, or replicate a received data flow based on a flow state of the data flow and/or a comparison of the data flow with predetermined patterns, and at least one inspection element (IE), configured to receive a replicated data flow, and to classify, whether the data flow is malicious or allowed. The IE is configured to alter the flow state of the data flow according to a classification result. The present disclosure provides a corresponding method for detecting and preventing intrusion of malicious data flows in a SDN.

Abstract: The present disclosure provides a domain name resolution method and apparatus based on a plurality of availability zones (AZ). After receiving, from a client, a domain name resolution request including a domain name of a to-be-accessed resource, a domain name system (DNS) determines, from a plurality of AZs corresponding to the domain name, a first AZ in a service available state, where each of the plurality of AZs is capable of providing the to-be-accessed resource, and then sends an internet protocol (IP) address of the first AZ to the client. In this way, the DNS processes the domain name resolution request sent by the client. After receiving the domain name resolution request sent by the client, the DNS feeds back, to the client, an IP address that is of an AZ in the service available state and that corresponds to the domain name resolution request.