公开(公告)号：US20210342694A1

公开(公告)日：2021-11-04

申请号：US17371590

申请日：2021-07-09

Applicant: Huawei Technologies Co., Ltd.

Inventor： Jiafeng Zhu ,  Wei Wei ,  Jianle Chen ,  Wei Wang ,  Jie Shen

IPC: G06N3/08 ,  G06N3/04

Abstract: A first aspect relates to a computer-implemented method for performing model compression. The method includes compressing a machine learning (ML) network model comprising a multiple layer structure to produce a compressed ML network model. The compressed ML network model maintains the multiple layer structure of the ML network model. The method generates a model file for the compressed ML network model. The model file includes the compressed ML network model and decoding information for enabling the ML network model to be decompressed and executed layer-by-layer.

公开(公告)号：US10972817B2

公开(公告)日：2021-04-06

申请号：US16801529

申请日：2020-02-26

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Biao Qi ,  Jun Zhang ,  Wei Wang ,  Wei Xiong

IPC: H04Q11/00 ,  G02B6/44

Abstract: A topology processing method, apparatus, and system are provided. The topology processing method includes: obtaining, by a topology processing apparatus, a first onsite image collected from an optical distribution network ODN, where the first onsite image includes at least an imaging of a first port of a first ODN device, the first port is connected to a first cable, a first identification area used to identify the first cable is disposed on the first cable, and the first onsite image further includes at least an imaging of the first identification area on the first cable; and identifying, by the topology processing apparatus, the first cable based on the first identification area on the first onsite image, and identifying, based on the first onsite image, the first port connected to the first cable; and generating a first correspondence between the first ODN device, the first port, and the first cable.

公开(公告)号：US20190384775A1

公开(公告)日：2019-12-19

申请号：US16552833

申请日：2019-08-27

Applicant: Huawei Technologies Co., Ltd.

Inventor： Jijun Wen ,  Bo Shang ,  Wei Wang

IPC: G06F16/27 ,  G06F16/23

Abstract: An active transaction list synchronization method and apparatus comprising a first node that records, in a transaction list incremental log buffer, a transaction list incremental log that is obtained after last active transaction list synchronization, where the transaction list incremental log indicates a change of a transaction recorded in an active transaction list of the first node and includes an added-transaction log and a committed-transaction log, wherein the added-transaction log indicates a transaction is added to the active transaction list, wherein the committed-transaction log indicates a transaction is deleted from the active transaction list. When performing group commit on transactions recorded in the committed-transaction log, the first node sends the transaction list incremental log to at least one second node.

公开(公告)号：US10305823B2

公开(公告)日：2019-05-28

申请号：US15581619

申请日：2017-04-28

Applicant: Huawei Technologies Co., Ltd.

Inventor： Chuande Zhang ,  Wenliang Wu ,  Wei Wang

IPC: G06F15/173 ,  H04L12/861 ,  G06F13/16 ,  H04L12/26 ,  G06F12/084 ,  H04L12/801 ,  H04L12/911 ,  G06F13/364 ,  H04L12/28 ,  H04L12/24 ,  G06F9/4401 ,  G06F9/455 ,  G06F13/10 ,  G06F13/40

Abstract: A network interface card configuration method and a resource management center are provided. According to the method, after obtaining a network interface card allocation request of an operating system that runs in a first CPU core, a resource management center selects, from M physical network interface cards and based on a network parameter of a network service required by the operating system, a target physical network interface card that conforms to the network parameter. Further, the resource management center selects at least one target hardware queue from each target physical network interface card and sends a command message to a network interface card controller. After receiving queue information of the target hardware queue from the network interface card controller, the resource management center send an instruction message to a CPU controller on a CPU board to instruct the CPU controller to construct a virtual network interface card.

公开(公告)号：US20180300116A1

公开(公告)日：2018-10-18

申请号：US16012397

申请日：2018-06-19

Applicant: Huawei Technologies Co., Ltd.

Inventor： Dmitry Meytin ,  Zhuoyue Liu ,  Wei Wang ,  Shay Horovitz

IPC: G06F8/60 ,  G06F9/455

Abstract: A method for automatic application deployment, and a cloud management node, where the method includes receiving an application orchestration template input by a user, parsing out an application model characteristic of an application from the application orchestration template, determining, according to the application model characteristic and a preset rule, a first instance creation manner corresponding to the application, determining a first computer device corresponding to the first instance creation manner in multiple computer devices, and sending a first instance creation command to the first computer device, the first instance creation command instructing the first computer device to create a first instance for the application according to the first instance creation manner. Therefore, application can be efficiently deployed automatically.

公开(公告)号：US20170075717A1

公开(公告)日：2017-03-16

申请号：US15360012

申请日：2016-11-23

Applicant: Huawei Technologies Co., Ltd.

Inventor： Dan Chen ,  Wei Wang ,  Kangkang Shen

IPC: G06F9/455

CPC classification number: G06F9/45558 ,  G06F9/455 ,  G06F2009/4557

Abstract: A virtual trusted platform module function implementation method is provided, the method is executed at an exception level EL3 of a processor that uses an ARM V8 architecture, and the method includes: generating, according to requirements of one or more VMs, one or more vTPM instances corresponding to each VM, and storing the generated one or more vTPM instances in preset secure space, where each vTPM instance has a dedicated instance communication queue for a VM corresponding to itself to use, and a physical address is allocated to each instance communication queue; and interacting with a VMM and the VM, so that the VM acquires a VM communication queue virtual address, in VM virtual address space, corresponding to a communication queue physical address of the vTPM instance, and the VM communicates with a vTPM instance communication queue by using the VM communication queue virtual address.

Abstract translation: 提供了一种虚拟可信平台模块功能实现方法，该方法在使用ARM V8架构的处理器的异常级EL3上执行，该方法包括：根据一个或多个VM的要求，生成一个或多个vTPM 与每个VM对应的实例，并将所生成的一个或多个vTPM实例存储在预设的安全空间中，其中每个vTPM实例具有用于与其自身对应的VM使用的专用实例通信队列，并且物理地址被分配给每个实例通信队列 ; 并且与VMM和VM进行交互，使得VM在VM虚拟地址空间中获取对应于vTPM实例的通信队列物理地址的VM通信队列虚拟地址，并且VM通过vTPM实例通信队列通信 使用VM通信队列虚拟地址。

公开(公告)号：US20170052979A1

公开(公告)日：2017-02-23

申请号：US15346114

申请日：2016-11-08

Applicant: Huawei Technologies Co., Ltd.

Inventor： Kai Qi ,  Wei Wang ,  Keping Chen

IPC: G06F17/30 ,  G06F9/48

CPC classification number: G06F16/188 ,  G06F9/4881 ,  G06F16/144 ,  G06F16/172

Abstract: An input/output (IO) request processing method and a file server, where the method includes adding, according to different service levels carried in IO requests of users, the IO requests of the users to corresponding cache queues for processing at a virtual file system layer, a block IO layer and a device driver layer separately, thereby meeting different service level requirements for the IO requests of the users.

Abstract translation: 一种输入/输出（IO）请求处理方法和文件服务器，其中该方法包括根据用户IO请求中携带的不同服务级别将用户的IO请求添加到相应的缓存队列中以在虚拟文件系统处理 层，块IO层和设备驱动层，从而满足用户的IO请求的不同服务级别要求。

公开(公告)号：US09195543B2

公开(公告)日：2015-11-24

申请号：US13732017

申请日：2012-12-31

Applicant: Huawei Technologies Co., Ltd.

Inventor： Yiyang Liu ,  Wei Wang ,  Haizhen Zheng

IPC: G06F11/00 ,  G06F11/14 ,  G06F11/10 ,  G06F11/16

CPC classification number: G11B5/48 ,  G06F11/1004 ,  G06F11/1666

Abstract: Embodiments of the present invention disclose a method and a device for detecting data reliability, which belong to the field of computer technologies. The method includes: dividing source data into multiple source data blocks; establishing a corresponding duplicate data block for each source data block, where the duplicate data block and the source data block are stored in different physical memory blocks respectively; establishing a corresponding reliability maintenance data structure for the source data block; and registering the reliability maintenance data structure with a first reliability maintenance thread of a central processing unit, the source data block corresponding to the reliability maintenance data structure registered with the first reliability maintenance thread, comparing a reliability check value of the source data block with a reliability check value recorded in the reliability maintenance data structure, and replacing the source data block having the error with the duplicate data block.

Abstract translation: 本发明的实施例公开了一种属于计算机技术领域的用于检测数据可靠性的方法和装置。 该方法包括：将源数据划分为多个源数据块; 为每个源数据块建立相应的重复数据块，其中重复数据块和源数据块分别存储在不同的物理存储器块中; 为源数据块建立相应的可靠性维护数据结构; 以及将所述可靠性维护数据结构与中央处理单元的第一可靠性维护线程对应，所述源数据块与所述第一可靠性维护线程中登记的可靠性维护数据结构相对应，将所述源数据块的可靠性检查值与 记录在可靠性维护数据结构中的可靠性检查值，以及用复制数据块替换具有错误的源数据块。

公开(公告)号：US09140858B2

公开(公告)日：2015-09-22

申请号：US13757293

申请日：2013-02-01

Applicant: Huawei Technologies Co., Ltd.

Inventor： Tengfang Xiong ,  Wei Wang ,  Baoliang Sun ,  Jianguo Wang

IPC: G02B6/36 ,  G02B6/38 ,  G02B6/42

CPC classification number: G02B6/36 ,  G02B6/3816 ,  G02B6/3887 ,  G02B6/3893 ,  G02B6/3894 ,  G02B6/3897 ,  G02B6/4292

Abstract: A waterproof optical fiber connector includes a flange and a terminal connector. The terminal connector includes an outer casing, an inner casing, an optical fiber and a connecting component coupled with the fiber. The inner casing is tubular for housing the fiber and the connecting component. The outer casing is tubular for housing the inner casing. The inner casing includes a body and a waterproof clip connected to a back end of the body. The fiber passes through the clip. The tail of the inner wall of the outer casing tapers from the front to the back, pressing the clip inwards such that the clip and the fiber are fixed together. At least one first protrusion is provided on the inner surface of the front end of the sleeve of the flange, and the front end of the body of the inner casing bears against the first protrusion.

Abstract translation: 防水光纤连接器包括凸缘和端子连接器。 端子连接器包括外壳，内壳，光纤和与光纤耦合的连接部件。 内壳是管状的，用于容纳纤维和连接部件。 外壳是用于容纳内壳的管状的。 内壳包括主体和连接到主体的后端的防水夹。 纤维穿过夹子。 外套的内壁的尾部从前到后逐渐变细，向内按压夹子，使夹子和纤维固定在一起。 至少一个第一突起设置在凸缘的套筒的前端的内表面上，并且内壳的主体的前端抵靠第一突起。

公开(公告)号：US20150257147A1

公开(公告)日：2015-09-10

申请号：US14721799

申请日：2015-05-26

Applicant: Huawei Technologies Co., Ltd.

Inventor： Xiaojun Ji ,  Wei Wang ,  Haobing Zhu

IPC: H04W72/04 ,  H04W76/02

CPC classification number: H04W72/0426 ,  H04W24/02 ,  H04W88/08 ,  H04W88/10 ,  H04W92/02 ,  H04W92/20

Abstract: The present invention discloses an inter-network collaboration method, a collaboration node, and a network side device. The method includes: obtaining status information of cells of different standards, where the status information indicates at least one of a radio resource status, terminal distribution, and a service condition of the cells of the different standards; and sending, according to the status information, coordination indication information to base stations or base station controllers of the cells of the different standards, so that the base stations or the base station controllers perform an inter-network collaboration operation according to the coordination indication information.

Abstract translation: 本发明公开了一种网络间协作方法，协作节点和网络侧设备。 该方法包括：获取不同标准的小区的状态信息，其中状态信息指示不同标准的小区的无线电资源状态，终端分配和服务条件中的至少一个; 根据状态信息将协调指示信息发送到不同标准的小区的基站或基站控制器，使得基站或基站控制器根据协调指示信息进行网间协作操作 。