公开(公告)号：US09015521B2

公开(公告)日：2015-04-21

申请号：US13712588

申请日：2012-12-12

Applicant: Huawei Technologies Co., Ltd.

Inventor： Haibo Tan ,  Zhenguo Wang ,  Baifeng Yu ,  Ping Huang ,  Junfeng Zhao

IPC: G06F11/00 ,  G06F11/20 ,  H04L12/26

CPC classification number: G06F11/202 ,  G06F11/2002 ,  G06F11/2017 ,  H04L43/0882

Abstract: Embodiments of the present invention disclose a method for switching an NC link, a processor system, and a node, where the processor system includes more than two nodes capable of communicating with each other, each node includes a node controller NC chip, a host bus adapter HBA apparatus, and at least one CPU, the NC chip is connected to each CPU in a node where the NC chip is located, and the HBA apparatus is connected to each CPU in a node where the HBA apparatus is located; an NC link borne by the NC chip is corresponding to an HBA link borne by the HBA apparatus. By using an HBA apparatus to deploy a redundant link, the cost of deploying the redundant link is reduced effectively under a premise of ensuring the reliability of the processor system.

Abstract translation: 本发明的实施例公开了一种用于切换NC链路，处理器系统和节点的方法，其中处理器系统包括能够彼此通信的多于两个的节点，每个节点包括节点控制器NC芯片，主机总线 适配器HBA装置和至少一个CPU，NC芯片连接到NC芯片所在的节点中的每个CPU，并且HBA装置连接到HBA装置所在的节点中的每个CPU; 由NC芯片承载的NC连接对应于由HBA设备承载的HBA链路。 通过使用HBA设备来部署冗余链路，在确保处理器系统的可靠性的前提下，有效地降低部署冗余链路的成本。

公开(公告)号：US11789619B2

公开(公告)日：2023-10-17

申请号：US17876861

申请日：2022-07-29

Applicant: Huawei Technologies Co., Ltd.

Inventor： Baifeng Yu ,  Zhou Yu ,  Jiongjiong Gu

IPC: G06F3/06 ,  G06F9/4401 ,  H04L67/1097 ,  G06F1/3221 ,  G06F1/3225 ,  G06F1/3234 ,  G06F1/3287 ,  H04L47/70 ,  G06F1/3203

CPC classification number: G06F3/0625 ,  G06F1/3203 ,  G06F1/3221 ,  G06F1/3225 ,  G06F1/3234 ,  G06F1/3268 ,  G06F1/3275 ,  G06F1/3287 ,  G06F3/067 ,  G06F3/0619 ,  G06F3/0631 ,  G06F3/0662 ,  G06F3/0665 ,  G06F3/0683 ,  G06F3/0689 ,  G06F9/4401 ,  G06F9/4418 ,  H04L47/70 ,  H04L67/1097 ,  Y02D10/00 ,  Y02D30/00

Abstract: A node interconnection apparatus includes a computing node, a resource control node, and a device interconnection interface connecting the computing node and the resource control node. Each of the computing node and the resource control node includes a processing unit and a storage unit, and the resource control node further includes a resource interface for connecting with a network storage device. The resource control node manages a storage resource of the network storage device, and when the computing node needs to start up, the resource control node obtains operating system startup information from the network storage device and provides the operating system startup information to the computing node. The computing node can start up without the need for storing startup information locally.

公开(公告)号：US11379265B2

公开(公告)日：2022-07-05

申请号：US16434403

申请日：2019-06-07

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Baifeng Yu ,  Jiongjiong Gu ,  Muhui Lin ,  Zhou Yu ,  Lingzhi Mao

IPC: G06F9/50 ,  G06F11/34 ,  G06F11/30 ,  G06F9/455

Abstract: A resource management technology is disclosed that may be applied to a host, where the host includes a Central Processing Unit (CPU), an endpoint connected to the CPU, and an Input/Output (I/O) device connected to the endpoint. A method includes: allocating, by the CPU, a target endpoint to a target process, where a virtual device is disposed on the target endpoint; obtaining, by the target endpoint, a performance specification of the target process, and adjusting a performance parameter of the virtual device according to the performance specification, where the adjusted virtual device satisfies requirements of performance specifications of all processes that use the target endpoint; and when the target process needs to access a resource, obtaining, from the I/O device, a resource that satisfies the performance specification of the target process, and providing the obtained resource to the target process for use.

公开(公告)号：US10331487B2

公开(公告)日：2019-06-25

申请号：US15471202

申请日：2017-03-28

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Baifeng Yu ,  Jiongjiong Gu ,  Muhui Lin ,  Zhou Yu ,  Lingzhi Mao

IPC: G06F9/50 ,  G06F9/455

Abstract: Embodiments provide a resource management technology that may be applied to a host, where the host includes a CPU, an endpoint connected to the CPU, and an I/O device connected to the endpoint. A method includes: allocating, by the CPU, a target endpoint to a target process, where a virtual device is disposed on the target endpoint; obtaining, by the target endpoint, a performance specification of the target process, and adjusting a performance parameter of the virtual device according to the performance specification, where the adjusted virtual device satisfies a total requirement of performance specifications of all processes that use the target endpoint; and when the target process needs to access a resource, obtaining, from the I/O device, a resource that satisfies the performance specification of the target process, and providing the obtained resource to the target process for use.

公开(公告)号：US10310756B2

公开(公告)日：2019-06-04

申请号：US15296054

申请日：2016-10-18

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Baifeng Yu ,  Zhou Yu ,  Jiongjiong Gu

IPC: G06F3/06 ,  H04L29/08 ,  G06F1/3203 ,  G06F1/3234 ,  G06F1/3287 ,  G06F9/4401 ,  H04L12/911

Abstract: Embodiments of the application provide a node interconnection apparatus, a resource control node, and a server system. The node interconnection apparatus includes a computing node and a resource control node, and a device interconnection interface connecting the two. Each of the computing node and the resource control node comprises a processing unit and a memory, and the resource control node further comprises a resource interface for connecting with one or more storage devices. The resource control node manages storage resources of the storage devices, and when a storage resource is required by the computing node in performing a computing task, the resource control node allocates the storage resource in the storage devices for the computing node. The computing node can be powered off to save the energy, while the resource control node remains powered on, so that the access to the storage devices is not affected.

公开(公告)号：US20170031620A1

公开(公告)日：2017-02-02

申请号：US15296054

申请日：2016-10-18

Applicant: HUAWEI TECHNOLOGIES CO.,LTD.

Inventor： Baifeng Yu ,  Zhou Yu ,  Jiongjiong Gu

IPC: G06F3/06 ,  G06F9/44 ,  G06F1/32 ,  H04L29/08 ,  H04L12/911

CPC classification number: G06F3/0625 ,  G06F1/3203 ,  G06F1/3234 ,  G06F1/3287 ,  G06F3/0619 ,  G06F3/0631 ,  G06F3/0665 ,  G06F3/067 ,  G06F3/0689 ,  G06F9/4401 ,  G06F9/4418 ,  H04L47/70 ,  H04L67/1097 ,  Y02D10/154 ,  Y02D10/44 ,  Y02D30/40

Abstract: Embodiments of the application provide a node interconnection apparatus, a resource control node, and a server system. The node interconnection apparatus includes a computing node and a resource control node, and a device interconnection interface connecting the two. Each of the computing node and the resource control node comprises a processing unit and a memory, and the resource control node further comprises a resource interface for connecting with one or more storage devices. The resource control node manages storage resources of the storage devices, and when a storage resource is required by the computing node in performing a computing task, the resource control node allocates the storage resource in the storage devices for the computing node. The computing node can be powered off to save the energy, while the resource control node remains powered on, so that the access to the storage devices is not affected.

Abstract translation: 应用的实施例提供节点互连设备，资源控制节点和服务器系统。 节点互连装置包括计算节点和资源控制节点，以及连接两者的设备互连接口。 计算节点和资源控制节点中的每一个包括处理单元和存储器，并且资源控制节点还包括用于与一个或多个存储设备连接的资源接口。 资源控制节点管理存储装置的存储资源，当计算节点在执行计算任务时需要存储资源时，资源控制节点将存储资源分配给计算节点的存储装置。 可以关闭计算节点以节约能源，同时资源控制节点保持通电，从而不影响对存储设备的访问。

公开(公告)号：US09026835B2

公开(公告)日：2015-05-05

申请号：US13717205

申请日：2012-12-17

Applicant: Huawei Technologies Co., Ltd.

Inventor： Dengben Wu ,  Yu Zhang ,  Baifeng Yu

IPC: G06F1/04 ,  G06F1/12 ,  G06F11/16 ,  G06F11/20

CPC classification number: G06F1/04 ,  G06F1/12 ,  G06F11/1604 ,  G06F11/20

Abstract: The present invention relates to a computer system and a clock configuring method. The computer system comprises at least two nodes, wherein each of the at least two nodes includes a selecting module and a CPU, inputs to the selecting module of any node comprise a clock of the node and a clock output from other node, and an output terminal of the selecting module is connected to the CPU and an input terminal of the selecting module of other node; the computer system further comprises a clock controlling module, whose output terminal is connected to a control terminal of the selecting module to control the clocks of the at least two nodes to be the same clock. When clocks of plural nodes are abnormal, the computer system can still normally operate as long as there is a normal clock in the computer system.

Abstract translation: 计算机系统和时钟配置方法技术领域本发明涉及计算机系统和时钟配置方法。 计算机系统包括至少两个节点，其中至少两个节点中的每一个包括选择模块和CPU，任何节点的选择模块的输入包括节点的时钟和来自其他节点的时钟输出，以及输出 选择模块的终端连接到CPU和其他节点的选择模块的输入端; 计算机系统还包括时钟控制模块，其输出端连接到选择模块的控制端，以控制至少两个节点的时钟为相同的时钟。 当多个节点的时钟异常时，只要计算机系统中有正常的时钟，计算机系统仍然可以正常运行。

公开(公告)号：US20130238871A1

公开(公告)日：2013-09-12

申请号：US13871596

申请日：2013-04-26

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Fan Fang ,  Baifeng Yu

IPC: G06F3/06

CPC classification number: G06F3/0604 ,  G06F13/1668 ,  G06F2213/0026

Abstract: A data processing method and apparatus, a PCI-E bus system, and a server are provided. The method includes: configuring address information of a PCI-E memory of a PCI-E device, so that the PCI-E device stores data received by the PCI-E device in the PCI-E memory; and controlling a CPU to access the data stored in the PCI-E memory, so that the CPU processes the data. thus a problem that in the prior art, because the PCI-E device stores the data received by the PCI-E device in a memory of the CPU, when the data stored in the memory of the CPU is accessed by another CPU, a part of a bandwidth of a bus between the another CPU and the CPU is occupied and a bus through which the CPU accesses the memory that corresponds the CPU is occupied can be avoided, thereby improving a utilization rate of the CPU.

Abstract translation: 提供数据处理方法和装置，PCI-E总线系统和服务器。 该方法包括：配置PCI-E设备的PCI-E存储器的地址信息，使得PCI-E设备将PCI-E设备接收的数据存储在PCI-E存储器中; 并控制CPU访问存储在PCI-E存储器中的数据，以便CPU处理数据。 因此在现有技术中，由于PCI-E设备将由PCI-E设备接收到的数据存储在CPU的存储器中，所以当存储在CPU的存储器中的数据被另一个CPU访问时， 占用另一CPU和CPU之间的总线带宽，并且可以避免CPU访问对应于CPU的存储器的总线，从而提高CPU的利用率。

公开(公告)号：US20130103975A1

公开(公告)日：2013-04-25

申请号：US13712588

申请日：2012-12-12

Applicant: Huawei Technologies Co., Ltd.

Inventor： Haibo Tan ,  Zhenguo Wang ,  Baifeng Yu ,  Ping Huang ,  Junfeng Zhao

IPC: G06F11/20

CPC classification number: G06F11/202 ,  G06F11/2002 ,  G06F11/2017 ,  H04L43/0882

Abstract: Embodiments of the present invention disclose a method for switching an NC link, a processor system, and a node, where the processor system includes more than two nodes capable of communicating with each other, each node includes a node controller NC chip, a host bus adapter HBA apparatus, and at least one CPU, the NC chip is connected to each CPU in a node where the NC chip is located, and the HBA apparatus is connected to each CPU in a node where the HBA apparatus is located; an NC link borne by the NC chip is corresponding to an HBA link borne by the HBA apparatus. By using an HBA apparatus to deploy a redundant link, the cost of deploying the redundant link is reduced effectively under a premise of ensuring the reliability of the processor system.

Abstract translation: 本发明的实施例公开了一种用于切换NC链路，处理器系统和节点的方法，其中处理器系统包括能够彼此通信的多于两个的节点，每个节点包括节点控制器NC芯片，主机总线 适配器HBA装置和至少一个CPU，NC芯片连接到NC芯片所在的节点中的每个CPU，并且HBA装置连接到HBA装置所在的节点中的每个CPU; 由NC芯片承载的NC连接对应于由HBA设备承载的HBA链路。 通过使用HBA设备来部署冗余链路，在确保处理器系统的可靠性的前提下，有效地降低部署冗余链路的成本。