公开(公告)号：US09003029B2

公开(公告)日：2015-04-07

申请号：US13721530

申请日：2012-12-20

Applicant: Huawei Technologies Co., Ltd.

Inventor： Hongbo Zhang ,  Guangyu Shi ,  Liufei Wen ,  Xiangyang Xu

IPC: H04L12/26 ,  H04L12/24 ,  H04L12/701 ,  H04L29/08

CPC classification number: H04L43/0876 ,  H04L41/0896 ,  H04L41/12 ,  H04L45/00 ,  H04L67/327

Abstract: Embodiments of the present invention disclose a method, an apparatus and a system for joint optimization. The method for joint optimization includes: decomposing the joint optimization of an entire network into the joint optimization performed in each sub-network, regarding a bandwidth requirement for a server off the sub-network as a bandwidth requirement for a virtual server on a port, iteratively performing the joint optimization in each sub-network, and applying results of the joint optimization in the network. In the embodiments of the present invention, the bandwidth requirement for the server off the sub-network is regarded as the bandwidth requirement for the virtual server on the port, the joint optimization is iteratively performed in each sub-network, and the results of the joint optimization are applied in the network, so that the joint optimization of the entire network is performed in parallel.

Abstract translation: 本发明的实施例公开了一种用于联合优化的方法，装置和系统。 联合优化方法包括：将关于子网的服务器的带宽要求作为端口上的虚拟服务器的带宽要求，将整个网络的联合优化分解为在每个子网中执行的联合优化， 迭代地执行每个子网中的联合优化，并将联合优化的结果应用于网络中。 在本发明的实施例中，将离开子网的服务器的带宽要求视为端口上虚拟服务器的带宽需求，在每个子网中迭代执行联合优化， 联合优化在网络中应用，从而并行执行整个网络的联合优化。

公开(公告)号：US09329783B2

公开(公告)日：2016-05-03

申请号：US14704735

申请日：2015-05-05

Applicant: Huawei Technologies Co., Ltd.

Inventor： Jian He ,  Guangyu Shi ,  Xiaoke Ni ,  Norbert Egi ,  Xiancai Li ,  Yu Liu ,  Huawei Liu

IPC: G06F13/12 ,  G06F13/38 ,  G06F3/06 ,  G06F13/10 ,  G06F13/40 ,  G06F13/42

CPC classification number: G06F13/4282 ,  G06F3/061 ,  G06F3/0611 ,  G06F3/0655 ,  G06F3/0673 ,  G06F12/10 ,  G06F13/10 ,  G06F13/28 ,  G06F13/4022 ,  G06F13/4221 ,  G06F2212/1016

Abstract: Embodiments of the present invention provide a data processing system and a data processing method. An MMIO address of a data request is obtained, where data stored in a PCIe storage device can be directly obtained from the MMIO address, and according to the MMIO address, a network adapter can directly read the data from the PCIe storage device of the data processing system, and transmit the data to a second data processing system, or directly write the data received from the second data processing system into the PCIe storage device. Therefore, the processing system can implement transmitting data directly from the PCIe storage device to the network adapter without using memory. During data transmission between the two data processing systems, a usage ratio of a resource, such as memory and a CPU, is reduced, and efficiency of data transmission is improved.

公开(公告)号：US09459798B2

公开(公告)日：2016-10-04

申请号：US14567656

申请日：2014-12-11

Applicant: Huawei Technologies Co., Ltd.

Inventor： Jian He ,  Guangyu Shi ,  Xiaoke Ni ,  Norbert Egi ,  Xiancai Li ,  Yu Liu ,  Huawei Liu

IPC: G06F13/12 ,  G06F13/38 ,  G06F3/06 ,  G06F13/10 ,  G06F12/10 ,  G06F13/28 ,  G06F13/40 ,  G06F13/42

CPC classification number: G06F13/4282 ,  G06F3/061 ,  G06F3/0611 ,  G06F3/0655 ,  G06F3/0673 ,  G06F12/10 ,  G06F13/10 ,  G06F13/28 ,  G06F13/4022 ,  G06F13/4221 ,  G06F2212/1016

Abstract: Embodiments of the present invention provide a data processing system and a data processing method. An MMIO address of a data request is obtained, where data stored in a PCIe storage device can be directly obtained from the MMIO address, and according to the MMIO address, a network adapter can directly read the data from the PCIe storage device of the data processing system, and transmit the data to a second data processing system, or directly write the data received from the second data processing system into the PCIe storage device. Therefore, the processing system can implement transmitting data directly from the PCIe storage device to the network adapter without using memory. During data transmission between the two data processing systems, a usage ratio of a resource, such as memory and a CPU, is reduced, and efficiency of data transmission is improved.

Abstract translation: 本发明的实施例提供了一种数据处理系统和数据处理方法。 获得数据请求的MMIO地址，其中可以从MMIO地址直接获取存储在PCIe存储设备中的数据，并且根据MMIO地址，网络适配器可以直接从PCIe存储设备读取数据 处理系统，并将数据发送到第二数据处理系统，或直接将从第二数据处理系统接收的数据写入PCIe存储设备。 因此，处理系统可以在不使用存储器的情况下实现从PCIe存储设备直接发送数据到网络适配器。 在两个数据处理系统之间的数据传输期间，诸如存储器和CPU之类的资源的使用率被降低，并且提高了数据传输的效率。

公开(公告)号：US20130263194A1

公开(公告)日：2013-10-03

申请号：US13907503

申请日：2013-05-31

Applicant: Huawei Technologies Co., Ltd.

Inventor： Hongbo Zhang ,  Guangyu Shi ,  Zhefeng Yan ,  Shueng-han Gary Chan

IPC: H04N21/231

CPC classification number: H04N21/23106 ,  H04L67/2842 ,  H04N21/2225 ,  H04N21/23116 ,  H04N21/237 ,  H04N21/24

Abstract: Embodiments of the present invention disclose a cooperative caching method and apparatus, relating to the field of network technologies, to improve the local hit ratio without increasing the local server costs. The technical solution provided in the present invention includes: obtaining, according to cache information, end-to-end delay between a local server and a neighbor server, and popularity in a cache list, a consolidated gain value of a cached video segment and a consolidated gain value of a candidate video segment in the local server; and replacing the cached video segment with the candidate video segment when the consolidated gain value of the cached video segment and the consolidated gain value of the candidate video segment in the local server meet a replacement condition.

Abstract translation: 本发明的实施例公开了与网络技术领域相关的协作缓存方法和装置，以提高本地命中率而不增加本地服务器成本。 本发明提供的技术方案包括：根据缓存信息获取本地服务器和邻居服务器之间的端到端延迟，以及缓存列表中的流行度，缓存视频段的合并增益值和 本地服务器中候选视频段的合并增益值; 并且当本地服务器中的缓存的视频段的合并增益值和候选视频段的合并增益值满足替换条件时，用候选视频段替换高速缓存的视频段。

公开(公告)号：US08527448B2

公开(公告)日：2013-09-03

申请号：US13722078

申请日：2012-12-20

Applicant: Huawei Technologies Co., Ltd.

Inventor： Kopylov Vladislav ,  Liufei Wen ,  Guangyu Shi

IPC: G06F9/44 ,  G06N7/02 ,  G06N7/06 ,  G06F15/18 ,  G06F7/00 ,  G06F17/30

CPC classification number: G06N5/02 ,  G06F17/2785 ,  G06F17/3071 ,  G06N5/003

Abstract: Embodiments of the present invention disclose a data processing method including: sending global initial statistical information to each slave node; merging received local statistical information of each slave node, to obtain new global statistical information; if Gibbs sampling performed by a slave node has ended, calculating a probability distribution between a document and topic and a probability distribution between the topic and a word according to the new global statistical information; according to the probability distributions obtained through calculation, establishing a likelihood function of a text set, and maximizing the likelihood function, to obtain a new hLDA hyper-parameter; and if iteration of solving for an hLDA hyper-parameter has converged, and according to the new hLDA hyper-parameter, calculating and outputting the probability distribution between the document and topic and the probability distribution between the topic and word.

Abstract translation: 本发明的实施例公开了一种数据处理方法，包括：向每个从节点发送全局初始统计信息; 合并接收每个从节点的本地统计信息，获得新的全局统计信息; 如果从节点执行的吉布斯抽样已经结束，则根据新的全局统计信息计算文档和主题之间的概率分布以及主题与单词之间的概率分布; 根据通过计算获得的概率分布，建立文本集的似然函数，并最大化似然函数，获得新的hLDA超参数; 并且如果求解hLDA超参数的迭代已经收敛，并且根据新的hLDA超参数，计算和输出文档和主题之间的概率分布以及主题与词之间的概率分布。

公开(公告)号：US09678918B2

公开(公告)日：2017-06-13

申请号：US15280700

申请日：2016-09-29

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Jian He ,  Guangyu Shi ,  Xiaoke Ni ,  Norbert Egi ,  Xiancai Li ,  Yu Liu ,  Huawei Liu

IPC: G06F13/12 ,  G06F13/38 ,  G06F13/42 ,  G06F3/06 ,  G06F13/10 ,  G06F12/10 ,  G06F13/28 ,  G06F13/40

CPC classification number: G06F13/4282 ,  G06F3/061 ,  G06F3/0611 ,  G06F3/0655 ,  G06F3/0673 ,  G06F12/10 ,  G06F13/10 ,  G06F13/28 ,  G06F13/4022 ,  G06F13/4221 ,  G06F2212/1016

Abstract: Embodiments of the present invention provide a data processing system and a data processing method. An MMIO address of a data request is obtained, where data stored in a PCIe storage device can be directly obtained from the MMIO address, and according to the MMIO address, a network adapter can directly read the data from the PCIe storage device of the data processing system, and transmit the data to a second data processing system, or directly write the data received from the second data processing system into the PCIe storage device. Therefore, the processing system can implement transmitting data directly from the PCIe storage device to the network adapter without using memory. During data transmission between the two data processing systems, a usage ratio of a resource, such as memory and a CPU, is reduced, and efficiency of data transmission is improved.

公开(公告)号：US20170017607A1

公开(公告)日：2017-01-19

申请号：US15280700

申请日：2016-09-29

Applicant: HUAWEI TECHNOLOGIES CO.,LTD.

Inventor： Jian He ,  Guangyu Shi ,  Xiaoke Ni ,  Norbert Egi ,  Xiancai Li ,  Yu Liu ,  Huawei Liu

IPC: G06F13/42 ,  G06F12/10 ,  G06F13/28 ,  G06F3/06

CPC classification number: G06F13/4282 ,  G06F3/061 ,  G06F3/0611 ,  G06F3/0655 ,  G06F3/0673 ,  G06F12/10 ,  G06F13/10 ,  G06F13/28 ,  G06F13/4022 ,  G06F13/4221 ,  G06F2212/1016

Abstract: Embodiments of the present invention provide a data processing system and a data processing method. An MMIO address of a data request is obtained, where data stored in a PCIe storage device can be directly obtained from the MMIO address, and according to the MMIO address, a network adapter can directly read the data from the PCIe storage device of the data processing system, and transmit the data to a second data processing system, or directly write the data received from the second data processing system into the PCIe storage device. Therefore, the processing system can implement transmitting data directly from the PCIe storage device to the network adapter without using memory. During data transmission between the two data processing systems, a usage ratio of a resource, such as memory and a CPU, is reduced, and efficiency of data transmission is improved.

Abstract translation: 本发明的实施例提供了一种数据处理系统和数据处理方法。 获得数据请求的MMIO地址，其中可以从MMIO地址直接获取存储在PCIe存储设备中的数据，并且根据MMIO地址，网络适配器可以直接从PCIe存储设备读取数据 处理系统，并将数据发送到第二数据处理系统，或直接将从第二数据处理系统接收的数据写入PCIe存储设备。 因此，处理系统可以在不使用存储器的情况下实现从PCIe存储设备直接发送数据到网络适配器。 在两个数据处理系统之间的数据传输期间，诸如存储器和CPU之类的资源的使用率被降低，并且提高了数据传输的效率。

公开(公告)号：US09213943B2

公开(公告)日：2015-12-15

申请号：US14153257

申请日：2014-01-13

Applicant: Huawei Technologies Co., Ltd.

Inventor： Kopylov Vladislav ,  Liufei Wen ,  Guangyu Shi

IPC: G06N5/04 ,  G06F17/30 ,  G06K9/62

CPC classification number: G06N5/048 ,  G06F17/30011 ,  G06K9/6218

Abstract: A parameter inference method to solve a problem that precision of a Latent Dirichlet Allocation model is poor is provided. The method includes: calculating a Latent Dirichlet Allocation model according to a preset initial first hyperparameter, a preset initial second hyperparameter, a preset initial number of topics, a preset initial count matrix of documents and topics, and a preset initial count matrix of topics and words to obtain probability distributions; obtaining the number of topics, a first hyperparameter, and a second hyperparameter that maximize log likelihood functions of the probability distributions; and determining whether the number of topics, the first hyperparameter, and the second hyperparameter converge, and if not, putting the number of topics, the first hyperparameter, and the second hyperparameter into the Latent Dirichlet Allocation model until the optimal number of topics, an optimal first hyperparameter, and an optimal second hyperparameter that maximize the log likelihood functions of the probability distributions.

Abstract translation: 提供了一种参数推理方法来解决潜在的Dirichlet分配模型的精度差的问题。 该方法包括：根据预设的初始第一超参数，预设的初始第二超参数，预设的初始数量的主题，预设的文档和主题的初始计数矩阵以及主题的预设初始计数矩阵来计算潜在Dirichlet分配模型，以及 词获得概率分布; 获得最大化概率分布的对数似然函数的主题数量，第一超参数和第二超参数; 以及确定主题数量，第一超参数和第二超参数是否收敛，如果不是，将主题数量，第一超参数和第二超参数放入潜在Dirichlet分配模型中，直到最优数量的主题， 最优的第一超参数，以及最大化概率分布的对数似然函数的最优的第二超参数。

公开(公告)号：US09137045B2

公开(公告)日：2015-09-15

申请号：US13676924

申请日：2012-11-14

Applicant: Huawei Technologies Co., Ltd.

Inventor： Xueliang Chen ,  Guangyu Shi ,  Liufei Wen

IPC: H04L12/28 ,  H04L12/54 ,  H04L12/413 ,  H04L12/40 ,  H04L12/18 ,  H04L12/761

CPC classification number: H04L12/56 ,  H04L12/185 ,  H04L12/40013 ,  H04L12/413 ,  H04L45/16

Abstract: An Ethernet multicast method and device are provided, which relate to the communications technology field, and improve the capability of distributing multicast data in the Ethernet. The method includes: receiving a request from a host/multicast source, wherein the request carries a multicast Media Access Control (MAC) address of a destination multicast group; selecting, according to the request, a switch satisfying a particular optimization condition as a multicast root node corresponding to the multicast MAC address of the destination multicast group when it is determined that the multicast MAC address of the destination multicast group does not have the corresponding multicast root node; transmitting an identification of the multicast root node to the host/multicast source. The embodiments of the present invention are mainly applied to the process of the multicast data distribution in the Ethernet.

Abstract translation: 提供了与通信技术领域相关的以太网组播方法和设备，提高了以太网中组播数据的分发能力。 该方法包括：从主机/组播源接收请求，其中所述请求携带目的地多播组的组播媒体访问控制（MAC）地址; 当确定目的组播组的组播MAC地址不具有对应的组播时，根据该请求，选择满足特定优化条件的交换机作为与目的组播组的组播MAC地址对应的组播根节点 根节点 向主机/组播源发送组播根节点的标识。 本发明的实施例主要应用于以太网中的组播数据分发过程。

公开(公告)号：US20130159236A1

公开(公告)日：2013-06-20

申请号：US13722078

申请日：2012-12-20

Applicant: Huawei Technologies Co., Ltd.

Inventor： Kopylov Vladislav ,  Liufei Wen ,  Guangyu Shi

IPC: G06N5/02

CPC classification number: G06N5/02 ,  G06F17/2785 ,  G06F17/3071 ,  G06N5/003

Abstract: Embodiments of the present invention disclose a data processing method including: sending global initial statistical information to each slave node; merging received local statistical information of each slave node, to obtain new global statistical information; if Gibbs sampling performed by a slave node has ended, calculating a probability distribution between a document and topic and a probability distribution between the topic and a word according to the new global statistical information; according to the probability distributions obtained through calculation, establishing a likelihood function of a text set, and maximizing the likelihood function, to obtain a new hLDA hyper-parameter; and if iteration of solving for an hLDA hyper-parameter has converged, and according to the new hLDA hyper-parameter, calculating and outputting the probability distribution between the document and topic and the probability distribution between the topic and word.

Abstract translation: 本发明的实施例公开了一种数据处理方法，包括：向每个从节点发送全局初始统计信息; 合并接收每个从节点的本地统计信息，获得新的全局统计信息; 如果从节点执行的吉布斯抽样已经结束，则根据新的全局统计信息计算文档和主题之间的概率分布以及主题与单词之间的概率分布; 根据通过计算获得的概率分布，建立文本集的似然函数，并最大化似然函数，获得新的hLDA超参数; 并且如果求解hLDA超参数的迭代已经收敛，并且根据新的hLDA超参数，计算和输出文档和主题之间的概率分布以及主题与词之间的概率分布。