Abstract: A data processing method includes sending, by a network interface card of a first device, a request packet to a second device. The request packet is used to request to read data in a destination storage area of the second device. The network interface card receives a response packet that is sent by the second device in response to the request packet. The response packet includes the data. The network interface card initiates, based on the response packet, direct memory access to a storage address to write the data into a memory area to which the storage address points. The first data does not need to be cached in a memory of the network interface card. Bandwidth resource usage and storage space usage of the memory of the network interface card can be reduced.

Abstract: One example method includes receiving network topology information delivered by a topology manager, where the data center includes a plurality of servers, a plurality of electrical switches, and at least one optical cross-connect device. A data flow can be obtained. A routing policy can be configured for the data flow based on the network topology information, where the routing policy includes any one or a combination of the following routing policies: a first routing policy, where the first routing policy indicates to forward the data flow through an optical channel in the at least one optical cross-connect device; a second routing policy, where the second routing policy indicates to split the data flow into at least two sub-data flows for forwarding; or a third routing policy, where the third routing policy indicates to forward the data flow through an electrical switch of the plurality of electrical switches.

Abstract: An example network adapter includes a processor and a hardware processing engine. The processor is configured to obtain a first input/output (I/O) command that indicates to write first data stored in a local server into at least one peer server, and to send first indication information to the hardware processing engine. The first indication information indicates the hardware processing engine to obtain the first data from the local server, and the indication information includes address information and length information that are of the first data and that are stored in the local server. The hardware processing engine is configured to directly process the first data when the first data is obtained from the local server based on the address information and the length information, and send the processed first data to the at least one peer server, where the first data is not stored in a memory of the network adapter.

Abstract: A data processing method for a network adapter includes the network adapter that obtains a first input/output (I/O) command. The first I/O command instructs to write data stored in a local server to at least one remote server, and the first I/O command includes address information and length information that are of the data and that are stored in the local server. The network adapter splits the data based on the address information and the length information to obtain a plurality of groups of address information and length information. The network adapter obtains, from the local server based on the groups of address information and length information, data corresponding to the groups of address information and length information, and sends the data to the at least one remote server.

Abstract: This application discloses a method for forwarding a packet in a data center network. A first device obtains an original packet, and adds a first source label to the original packet to obtain a first packet. The first source label includes a forwarding type, an indication field, and an interface sequence. The forwarding type indicates that the first packet supports source label forwarding, the interface sequence indicates a first source label forwarding path of the original packet, and the indication field indicates information that is about an outbound interface and that should be read from the interface sequence. The first device sends the first packet to a next-hop switch through the outbound interface corresponding to the first source label forwarding path. The next-hop switch receives the first packet, and forwards the first packet based on the first source label.

Abstract: In a data access method, a processor of a host converts a first descriptor recognized by a virtual machine interface card of the virtual machine into a second descriptor recognized by a physical interface card of the host. The first descriptor includes a virtual machine physical memory address and a data length of accessible data. The physical interface card of the host obtains, based on the second descriptor, a physical address that is in a memory and to which the virtual machine physical memory address is mapped, and accesses the accessible data according to the physical address in the memory.

Abstract: In a data center, a network interface card is utilized to assist processing of data packets. The server comprising a host and the network interface card connected to the host. The host is configured to generate processing information when receiving data packet from the network interface card, and send the processing information to the network interface card. The network interface card is configured to generate an action field and a match field of a session entry to process the data packet, based on the processing information sent by the host and the matching information of the data packet.

Abstract: A data processing method includes sending, by a network interface card of a first device, a request packet to a second device. The request packet is used to request to read data in a destination storage area of the second device. The network interface card receives a response packet that is sent by the second device in response to the request packet. The response packet includes the data. The network interface card initiates, based on the response packet, direct memory access to a storage address to write the data into a memory area to which the storage address points. The first data does not need to be cached in a memory of the network interface card. Bandwidth resource usage and storage space usage of the memory of the network interface card can be reduced.

Abstract: A virtual machine live migration method and a server, to resolve a problem that live migration of a virtual machine cannot be implemented in an SR-IOV technology scenario. According to the virtual machine live migration method and the server in the embodiments of the present invention, by setting a dirty write flag, a virtual PCIE device of a to-be-migrated virtual machine performs at least one read and write operation on received data; a virtual machine manager of the to-be-migrated virtual machine can identify a change in the data and migrate changed data to a destination virtual machine, thereby resolving a problem in the prior art that data cannot be migrated during a virtual machine live migration process because the data passes through the virtual PCIE device but the virtual machine manager cannot perceive the data.

Abstract: In a data access method, a processor of a host converts a first descriptor recognized by a virtual machine interface card of the virtual machine into a second descriptor recognized by a physical interface card of the host. The first descriptor includes a virtual machine physical memory address and a data length of accessible data. The physical interface card of the host obtains, based on the second descriptor, a physical address that is in a memory and to which the virtual machine physical memory address is mapped, and accesses the accessible data according to the physical address in the memory.

Abstract: In a data center, a network interface card is utilized to assist processing of data packets. The server comprising a host and the network interface card connected to the host. The host is configured to generate processing information when receiving data packet from the network interface card, and send the processing information to the network interface card. The network interface card is configured to generate an action field and a match field of a session entry to process the data packet, based on the processing information sent by the host and the matching information of the data packet.

Abstract: A method for managing a memory of a virtual machine, a physical host, a PCIE device and a configuration method thereof are provided. The method executed by a virtual machine includes: allocating a memory to a service carried on a PCIE device, where the memory includes multiple memory blocks, the multiple memory blocks is used to save working information of the service; generating a base address table BAT and a chip logic address table CLAT, where the BAT includes a CLAT entry base address corresponding to the service, and the CLAT includes a first address of each memory block; and sending an address of the BAT and a function number corresponding to the virtual machine to the PCIE device. Therefore, the PCIE device can obtain, according to the address of the BAT and the function number, working information of a service from the virtual machine.

Abstract: A switch board that includes a port configuring unit and a plurality of ports, where the port configuring unit is configured to divide the switch board into more than one virtual sub-switch-board, allocate the ports to the virtual sub-switch-boards, and configure each port of each virtual sub-switch-board to a first-type port or a second-type port, where the first-type port and the second-type port have different bandwidth; and each of the ports is configured to connect a server blade according to a configuration on the port configuring unit. The technical solutions of the present disclosure can meet a requirement for flexible port bandwidth configuration.

Abstract: A method for managing a memory of a virtual machine, a physical host, a PCIE device and a configuration method thereof are provided. The method executed by a virtual machine includes: allocating a memory to a service carried on a PCIE device, where the memory includes multiple memory blocks, the multiple memory blocks is used to save working information of the service; generating a base address table BAT and a chip logic address table CLAT, where the BAT includes a CLAT entry base address corresponding to the service, and the CLAT includes a first address of each memory block; and sending an address of the BAT and a function number corresponding to the virtual machine to the PCIE device. Therefore, the PCIE device can obtain, according to the address of the BAT and the function number, working information of a service from the virtual machine.

Abstract: A switch board that includes a port configuring unit and a plurality of ports, where the port configuring unit is configured to divide the switch board into more than one virtual sub-switch-board, allocate the ports to the virtual sub-switch-boards, and configure each port of each virtual sub-switch-board to a first-type port or a second-type port, where the first-type port and the second-type port have different bandwidth; and each of the ports is configured to connect a server blade according to a configuration on the port configuring unit. The technical solutions of the present disclosure can meet a requirement for flexible port bandwidth configuration.

Abstract: A switch board of a blade server, a port configuring method thereof, and a blade server. The switch board includes a port configuring unit and a plurality of ports, where the port configuring unit is configured to divide the switch board into more than one virtual sub-switch-board, allocate the ports to the virtual sub-switch-boards, and configure each port of each virtual sub-switch-board to a first-type port or a second-type port, where the first-type port and the second-type port have different bandwidth; and each of the ports is configured to connect a server blade according to a configuration on the port configuring unit. The technical solutions of the present invention can meet a requirement for flexible port bandwidth configuration.

Abstract: A virtual machine live migration method and a server, to resolve a problem that live migration of a virtual machine cannot be implemented in an SR-IOV technology scenario. According to the virtual machine live migration method and the server in the embodiments of the present invention, by setting a dirty write flag, a virtual PCIE device of a to-be-migrated virtual machine performs at least one read and write operation on received data; a virtual machine manager of the to-be-migrated virtual machine can identify a change in the data and migrate changed data to a destination virtual machine, thereby resolving a problem in the prior art that data cannot be migrated during a virtual machine live migration process because the data passes through the virtual PCIE device but the virtual machine manager cannot perceive the data.

Abstract: A switch board of a blade server, a port configuring method thereof, and a blade server. The switch board includes a port configuring unit and a plurality of ports, where the port configuring unit is configured to divide the switch board into more than one virtual sub-switch-board, allocate the ports to the virtual sub-switch-boards, and configure each port of each virtual sub-switch-board to a first-type port or a second-type port, where the first-type port and the second-type port have different bandwidth; and each of the ports is configured to connect a server blade according to a configuration on the port configuring unit. The technical solutions of the present invention can meet a requirement for flexible port bandwidth configuration.