公开(公告)号：US09807009B2

公开(公告)日：2017-10-31

申请号：US14676053

申请日：2015-04-01

Applicant: Google Inc.

Inventor： Junlan Zhou ,  Zhengrong Ji ,  Jiangbo Li

IPC: H04L12/801 ,  H04L12/825 ,  H04L12/721

CPC classification number: H04L47/11 ,  H04L45/66 ,  H04L47/13 ,  H04L47/263 ,  H04L47/33

Abstract: A system and method is provided for sending congestion notification messages through L3 networks. For example, a data packet is received at a first switch in a first fabric block of an L3 network, and the first switch performs source MAC tagging of the data packet. The data packet is then forwarded to a second switch in a second fabric block of the L3 network, and the source MAC tag is maintained by the second switch and any intermediate switches. The second switch determines, in response to receiving the data packet, whether it is congested, and generates a notification message if it is congested. The notification message is L2 forwarded to the first fabric block, and further forwarded from the first switch to a source of the data packet using ACL matching.

公开(公告)号：US09407565B1

公开(公告)日：2016-08-02

申请号：US14307713

申请日：2014-06-18

Applicant: Google Inc.

Inventor： Jiangbo Li ,  Junlan Zhou ,  Ke Xiao ,  Victor Lin ,  Ashish Vibhakar Naik ,  Shidong Zhang

IPC: H04L12/835

CPC classification number: H04L47/266

Abstract: Aspects of the present disclosure relate to detecting and repairing permanently pauses on a flow controlled fabric. In one aspect, one or more computing devices, such as a switch or a centralized controller, may detect whether a port of a network device receives one or more pause messages. The pause messages may instruct the network device to pause data transmission. Further, the one or more computing devices may determine a period of time during which the port receives the one or more pause messages and identify the port as a permanently paused port based on the determined period of time. The one or more computing devices may then reconfigure the permanently paused port to stop complying with the one or more pause messages.

Abstract translation: 本公开的方面涉及在流动控制的织物上检测和修复永久暂停。 在一个方面，诸如交换机或集中控制器的一个或多个计算设备可以检测网络设备的端口是否接收到一个或多个暂停消息。 暂停消息可以指示网络设备暂停数据传输。 此外，一个或多个计算设备可以基于所确定的时间段来确定端口接收到一个或多个暂停消息并且将端口识别为永久暂停的端口的时间段。 一个或多个计算设备然后可以重新配置永久暂停的端口以停止遵守一个或多个暂停消息。

公开(公告)号：US09247326B2

公开(公告)日：2016-01-26

申请号：US14169734

申请日：2014-01-31

Applicant: Google Inc.

Inventor： Ryohei Urata ,  Hong Liu ,  Amin Vahdat ,  Christopher Lyle Johnson ,  Junlan Zhou ,  Alexander Shraer

IPC: H04Q11/00 ,  G06F13/40

CPC classification number: H04Q11/0005 ,  G06F13/4022 ,  H04Q11/0062 ,  H04Q2011/0015 ,  H04Q2011/0052 ,  Y02D10/14 ,  Y02D10/151

Abstract: Systems and methods for increasing bandwidth in a computer network are provided. A computer network can include a first lower level switch having a first port and a second port. The computer network can include a second lower level switch having a first port and a second port. The computer network can include an upper level switch having respective ports directly coupled to ports of the first and second lower level switches. A third port of the upper level switch can couple to a first port of a passive optical splitter. The passive optical splitter can have second and third ports coupled to respective ports of the first and second lower level switches. The passive optical splitter can be configured to transmit signals received at its first port as output signals on both of its second and third ports.

Abstract translation: 提供了一种用于增加计算机网络带宽的系统和方法。 计算机网络可以包括具有第一端口和第二端口的第一下级开关。 计算机网络可以包括具有第一端口和第二端口的第二下层开关。 计算机网络可以包括具有直接耦合到第一和第二下级交换机的端口的相应端口的上级交换机。 上级开关的第三端口可以耦合到无源分光器的第一端口。 无源光分路器可以具有耦合到第一和第二下层开关的相应端口的第二和第三端口。 无源光分路器可以被配置为将在其第一端口处接收的信号作为其第二和第三端口两端的输出信号进行发送。

公开(公告)号：US20150180778A1

公开(公告)日：2015-06-25

申请号：US14139150

申请日：2013-12-23

Applicant: Google Inc.

Inventor： Alexander Shraer ,  Junlan Zhou ,  Amin Vahdat

IPC: H04L12/801

CPC classification number: H04L49/205 ,  H04L45/125 ,  H04L47/10 ,  H04L47/125 ,  H04L49/356

Abstract: The present disclosure provides for the determination of bandwidth allocation of inter-block traffic in a data center network. It employs a number of optimization objectives and a heuristic water-filling strategy to avoid producing unnecessary paths and to avoid determining paths that would be unavailable when actually needed. Allocation may be adjusted incrementally upon node and link failure, for instance to perform only the minimal allocation changes necessary. If demand between a source and a destination cannot be satisfied, a decomposition process may be used to allocate remaining demand. One aspect constructs a graph for route computation based on inter-block topology. Here, the graph initially starts with a highest level of abstraction with each node representing a middle block, and gradually reduces the abstraction level to identify paths of mixed abstraction level to satisfy additional demand.

Abstract translation: 本公开提供了确定数据中心网络中的块间业务的带宽分配。 它采用了一些优化目标和启发式注水策略，以避免产生不必要的路径，并避免确定在实际需要时不可用的路径。 在节点和链路故障时，分配可以逐渐调整，例如仅执行必要的最小分配更改。 如果源和目的地之间的需求不能满足，则可以使用分解处理来分配剩余的需求。 一方面构建了基于块间拓扑的路由计算图。 这里，图形最初以最高级别的抽象开始，每个节点表示一个中间块，并逐渐降低抽象级别，以识别混合抽象级别的路径以满足额外的需求。

公开(公告)号：US09660914B1

公开(公告)日：2017-05-23

申请号：US14676068

申请日：2015-04-01

Applicant: Google Inc.

Inventor： Junlan Zhou ,  Zhengrong Ji

IPC: H04L1/00 ,  H04L12/26 ,  H04L12/28 ,  H04L12/801 ,  H04L12/723

CPC classification number: H04L43/062 ,  H04L47/11 ,  H04L47/26 ,  H04L47/35

Abstract: A system and method is provided for sending congestion notification messages through L3 networks and implementation of QCN in L3 switches. For example, according to this system and method, an L3 switch receives one or more data packets, and determines, based on the received one or more data packets, whether the L3 switch is congested. If the L3 switch is congested, it generates a congestion notification message, the congestion notification message including an Internet Protocol (IP) header, the IP header identifying a source of the one or more received data packets as its destination. The L3 switch sends the congestion notification message to the source of the one or more received data packets using information in the IP header.

公开(公告)号：US20150372918A1

公开(公告)日：2015-12-24

申请号：US14676053

申请日：2015-04-01

Applicant: Google Inc.

Inventor： Junlan Zhou ,  Zhengrong Ji ,  Jiangbo Li

IPC: H04L12/801

CPC classification number: H04L47/11 ,  H04L45/66 ,  H04L47/13 ,  H04L47/263 ,  H04L47/33

Abstract: A system and method is provided for sending congestion notification messages through L3 networks. For example, a data packet is received at a first switch in a first fabric block of an L3 network, and the first switch performs source MAC tagging of the data packet. The data packet is then forwarded to a second switch in a second fabric block of the L3 network, and the source MAC tag is maintained by the second switch and any intermediate switches. The second switch determines, in response to receiving the data packet, whether it is congested, and generates a notification message if it is congested. The notification message is L2 forwarded to the first fabric block, and further forwarded from the first switch to a source of the data packet using ACL matching.

Abstract translation: 提供一种通过L3网络发送拥塞通知消息的系统和方法。 例如，在L3网络的第一结构块中的第一交换机处接收数据分组，并且第一交换机执行数据分组的源MAC标记。 数据包然后转发到L3网络的第二个结构块中的第二个交换机，源MAC标签由第二个交换机和任何中间交换机维护。 第二交换机响应于接收到数据分组而确定其是否拥塞，并且如果其拥塞，则生成通知消息。 通知消息是L2转发到第一个结构块，并使用ACL匹配从第一个交换机进一步转发到数据包的源。

公开(公告)号：US20140105023A1

公开(公告)日：2014-04-17

申请号：US14106672

申请日：2013-12-13

Applicant: Google Inc.

Inventor： Yuguang Wu ,  Junlan Zhou

IPC: H04L12/891

CPC classification number: H04L47/41 ,  H04L45/125 ,  H04L45/38 ,  H04L47/2441

Abstract: A method includes, receiving a plurality of data flows. A respective data flow includes a respective source address and a respective destination address. The method further includes generating, without regard to priorities associated with the plurality of data flows, an ordering of the plurality of data flows; and iteratively modifying, without regard to the priorities, the ordering of the plurality of data flows by applying a randomization algorithm to the plurality of data flows, until a cost associated with path assignments for the ordering of the plurality of data flows satisfies a predetermined condition. A respective path assignment for a respective data flow specifies a respective path from a respective source address to a respective destination address. The method also includes executing the data flows based on the path assignments for the ordering of the plurality of data flows having the cost that satisfies the predetermined condition.

Abstract translation: 一种方法包括：接收多个数据流。 相应的数据流包括相应的源地址和相应的目的地地址。 该方法还包括：生成与多个数据流相关联的优先级，多个数据流的排序; 并且通过对所述多个数据流应用随机化算法来迭代地修改所述多个数据流的顺序，直到与所述多个数据流的顺序的路径分配相关联的成本满足预定条件 。 相应数据流的相应路径分配指定从相应源地址到相应目的地地址的相应路径。 该方法还包括基于具有满足预定条件的成本的多个数据流的顺序的路径分配来执行数据流。

公开(公告)号：US09491102B1

公开(公告)日：2016-11-08

申请号：US14176518

申请日：2014-02-10

Applicant: Google Inc.

Inventor： Junlan Zhou ,  Min Zhu ,  Amin Vahdat

IPC: H04L12/26 ,  H04L12/803 ,  H04L12/66 ,  H04L12/46 ,  H04L12/715

CPC classification number: H04L47/125 ,  H04L12/4625 ,  H04L12/4633 ,  H04L12/66 ,  H04L43/0882 ,  H04L45/64

Abstract: A technique for load balancing in a multi-topology network selects a network path having a favorable bandwidth*hop count product. Accordingly, shorter paths are prioritized while longer paths are used for traffic load balancing when appropriate. For example, a telecommunications network may employ a Clos-style topology in which hosts are connected to each other through multiple hierarchical levels of forwarding devices. One or more high level switches may be removed, creating isolated blocks of forwarding devices. The blocks may be connected by the remaining original high level Clos-style topology switch(es) and a second type of network topology. In an exemplary embodiment, traffic may be scheduled over routes with the following decreasing priorities: (1) the shortest path, using either or both network topologies; (2) a path using only a single network topology type; (3) non-shortest paths between directly-connected blocks; and (4) non-shortest paths between indirectly-connected blocks.

Abstract translation: 在多拓扑网络中进行负载平衡的技术选择具有有利带宽跳跃计数乘积的网络路径。 因此，在适当的情况下，较长的路径用于流量负载平衡，优先考虑较短的路径。 例如，电信网络可以采用Clos式拓扑，其中主机通过多个层次级的转发设备相互连接。 可以去除一个或多个高级交换机，从而创建隔离的转发设备。 块可以由剩余的原始高级Clos风格的拓扑交换机和第二类型的网络拓扑连接。 在示例性实施例中，可以通过具有以下降低优先级的路由来调度业务：（1）使用网络拓扑中的一个或两个的最短路径; （2）仅使用单个网络拓扑类型的路径; （3）直接连接块之间的非最短路径; 和（4）间接连接的块之间的非最短路径。

公开(公告)号：US09397957B2

公开(公告)日：2016-07-19

申请号：US14139150

申请日：2013-12-23

Applicant: Google Inc.

Inventor： Alexander Shraer ,  Junlan Zhou ,  Amin Vahdat

IPC: H04L12/801 ,  H04L12/931 ,  H04L12/803 ,  H04L12/729

CPC classification number: H04L49/205 ,  H04L45/125 ,  H04L47/10 ,  H04L47/125 ,  H04L49/356

Abstract: The present disclosure provides for the determination of bandwidth allocation of inter-block traffic in a data center network. It employs a number of optimization objectives and a heuristic water-filling strategy to avoid producing unnecessary paths and to avoid determining paths that would be unavailable when actually needed. Allocation may be adjusted incrementally upon node and link failure, for instance to perform only the minimal allocation changes necessary. If demand between a source and a destination cannot be satisfied, a decomposition process may be used to allocate remaining demand. One aspect constructs a graph for route computation based on inter-block topology. Here, the graph initially starts with a highest level of abstraction with each node representing a middle block, and gradually reduces the abstraction level to identify paths of mixed abstraction level to satisfy additional demand.

Abstract translation: 本公开提供了确定数据中心网络中的块间业务的带宽分配。 它采用了一些优化目标和启发式注水策略，以避免产生不必要的路径，并避免确定在实际需要时不可用的路径。 在节点和链路故障时，分配可以逐渐调整，例如仅执行必要的最小分配更改。 如果源和目的地之间的需求不能满足，则可以使用分解处理来分配剩余的需求。 一方面构建了基于块间拓扑的路由计算图。 这里，图形最初以最高级别的抽象开始，每个节点表示一个中间块，并逐渐降低抽象级别，以识别混合抽象级别的路径以满足额外的需求。

公开(公告)号：US09363162B1

公开(公告)日：2016-06-07

申请号：US14059015

申请日：2013-10-21

Applicant: Google Inc.

Inventor： Junlan Zhou ,  Arjun Singh ,  Amin Vahdat

IPC: H04J1/16 ,  H04L12/721

CPC classification number: H04L45/123

Abstract: Exemplary embodiments determine respective capacities of network links in a multi-stage network. Specifically, the capacities of the upstream links for a given network element are determined based on the capacities of the downstream links for that network element. According to the various embodiments, a network element determines the amount of data, i.e. traffic, which may be assigned to downstream links based on the determined capacities. A network element may not receive more traffic than the total direct downstream capacity of the network element. Thus, optimum throughput may be attained for the entire network.

Abstract translation: 示例性实施例确定多级网络中的网络链路的相应容量。 具体地，基于用于该网络元件的下游链路的容量来确定给定网元的上行链路的容量。 根据各种实施例，网络元件基于所确定的容量确定可以分配给下游链路的数据量，即流量。 网络元件可能不会比网络元件的总直接下行容量接收更多的流量。 因此，可以为整个网络获得最佳的吞吐量。