公开(公告)号：US20130083796A1

公开(公告)日：2013-04-04

申请号：US13366292

申请日：2012-02-04

Applicant: Brad Matthews

Inventor： Brad Matthews

IPC: H04L12/56

CPC classification number: H04L49/103 ,  H04L49/201 ,  H04L49/351

Abstract: A system and method for improving multicast performance in banked shared memory architectures. Temporal localities created by multicast packets in a shared memory bank are addressed through caching. In one embodiment, multicast packets are stored in a cache memory that is associated with a bank of shared memory. In another embodiment, read requests for multicast packets are stored in a read request cache, wherein additional read requests are accumulated prior to an actual read event.

Abstract translation: 一种用于提高存储共享存储器架构中的组播性能的系统和方法。 通过组播数据包在共享存储器中创建的时间局部通过缓存来解决。 在一个实施例中，多播分组被存储在与一组共享存储器相关联的高速缓冲存储器中。 在另一个实施例中，对多播分组的读取请求被存储在读取请求高速缓存中，其中在实际读取事件之前累积附加的读取请求。

公开(公告)号：US08363649B2

公开(公告)日：2013-01-29

申请号：US12601274

申请日：2008-05-22

Applicant: Venkat Konda

Inventor： Venkat Konda

IPC: H04L12/50

CPC classification number: H04L49/201 ,  H04L49/1515

Abstract: A generalized multi-link multi-stage network comprising (2×logdN)−1 stages is operated in strictly nonblocking manner for unicast, also in rearrangeably nonblocking manner for arbitrary fan-out multicast when s≧2, and in strictly nonblocking manner for arbitrary fan-out multicast when s≧3, includes an input stage having N/d switches with each of them having d inlet links and s×d outgoing links connecting to second stage switches, an output stage having N/d switches with each of them having d outlet links and s×d incoming links connecting from switches in the penultimate stage. The network also has (2×logdN)−3 middle stages with each middle stage having N/d switches, and each switch in the middle stage has s×d incoming links connecting from the switches in its immediate preceding stage, and s×d outgoing links connecting to the switches in its immediate succeeding stage. Also each multicast connection is set up by use of at most two outgoing links from the input stage switch.

Abstract translation: 包括（2×logdN）-1级的广义多链路多级网络以严格的非阻塞方式单播操作，当s≥2时，也可以以任意扇出组播的可重排非阻塞方式进行操作，严格地不阻塞任意方式 在s≥3时出现扇出组播，包括具有N / d个开关的输入级，其中每一个具有d个入口链路和连接到第二级交换机的s×d个输出链路，具有每个N / d开关的输出级 在倒数第二阶段有开关连接的d出口连接和s×d进入链路。 网络还具有（2×logdN）-3个中间阶段，每个中间阶段有N / d个交换机，中间阶段的每个交换机都有从前一个交换机连接的s×d个入站链路，s×d 在其后续阶段连接到交换机的出站链路。 此外，每个组播连接通过使用来自输入级交换机的至多两个输出链路来建立。

公开(公告)号：US20130024899A1

公开(公告)日：2013-01-24

申请号：US13606578

申请日：2012-09-07

Applicant: Yong Ho Son ,  Nikhil Rajdev

Inventor： Yong Ho Son ,  Nikhil Rajdev

IPC: H04N21/266

CPC classification number: H04L12/2801 ,  H04L29/06027 ,  H04L49/103 ,  H04L49/201 ,  H04L65/4092 ,  H04N7/17309 ,  H04N21/2385 ,  H04N21/64 ,  H04N21/643

Abstract: A system for generating and sending a Logical Node identification signal as part of a data stream is disclosed. The system also includes subscriber stations capable of receiving and extracting Logical Node identification information from a data stream. The subscriber stations create new messages including the Logical Node identification signal and send the message to the transmission network system control, such that switching of data streams to an appropriate channel over the information transmission network is done consistent with the network topology.

Abstract translation: 公开了一种用于生成和发送逻辑节点识别信号作为数据流的一部分的系统。 该系统还包括能够从数据流接收和提取逻辑节点识别信息的用户站。 用户站创建包括逻辑节点识别信号的新消息，并将消息发送到传输网络系统控制，使得数据流在信息传输网络上的适当信道的切换与网络拓扑一致。

公开(公告)号：US08332693B2

公开(公告)日：2012-12-11

申请号：US12790603

申请日：2010-05-28

Applicant: Wei Cao ,  Guoyi Chen

Inventor： Wei Cao ,  Guoyi Chen

IPC: G06F11/00

CPC classification number: H04L45/22 ,  H04L41/06 ,  H04L45/16 ,  H04L45/28 ,  H04L45/50 ,  H04L49/201 ,  H04L49/557

Abstract: A method and an apparatus for failure notification are provided. The method includes: sending, when it is detected that a failure occurs in a multicast label switch path, failure information to a downstream node of a multicast tree where a failed point locates, where the multicast label switch path is established based on a multicast label distribution protocol; and forwarding the failure information through the downstream node, where the failure information arrives at a leaf node of the multicast tree where the failed point locates. According to the method and the apparatus for failure notification of the present invention, finally, the failure information is transmitted to a leaf node. After the failure information is obtained, the leaf node may take appropriate measure to ensure the normal service and improve the reliability of the network.

Abstract translation: 提供了一种用于故障通知的方法和装置。 该方法包括：当检测到组播标签交换机路径发生故障时，发送故障信息到失败点定位的组播树的下游节点，其中组播标签交换路径基于组播标签建立 分配协议; 并通过下游节点转发故障信息，故障信息到达失败点所在的组播树的叶节点。 根据本发明的故障通知方法和装置，最终将故障信息发送到叶节点。 获得故障信息后，叶节点可采取适当措施，保证正常业务，提高网络的可靠性。

公开(公告)号：US08325726B2

公开(公告)日：2012-12-04

申请号：US12702718

申请日：2010-02-09

Applicant: Alex Baban ,  Jean-Marc Frailong ,  Kireeti Kompella ,  Nitin Kumar ,  Jamsheed Wania ,  Murtuza Attarwala

Inventor： Alex Baban ,  Jean-Marc Frailong ,  Kireeti Kompella ,  Nitin Kumar ,  Jamsheed Wania ,  Murtuza Attarwala

IPC: H04L12/56 ,  H04J3/26

CPC classification number: H04L12/18 ,  H04L45/16 ,  H04L49/10 ,  H04L49/201

Abstract: A network device receives a packet with a multicast nexthop identifier, and creates a mask that includes addresses of egress packet forwarding engines, of the network device, to which to provide the packet. The network device divides the mask into two portions, generates two copies of the packet, provides a first portion of the mask in a first copy of the packet, and provides a second portion of the mask in a second copy of the packet. The network device also forwards the first copy of the packet to an address of a first egress packet forwarding engine provided in the first portion of the mask, and forwards the second copy of the packet to an address of a second egress packet forwarding engine provided in the second portion of the mask.

Abstract translation: 网络设备接收具有多播下一标识符的分组，并创建包括提供分组的网络设备的出站分组转发引擎的地址的掩码。 网络设备将掩码分成两部分，生成分组的两个副本，在分组的第一副本中提供该掩码的第一部分，并在分组的第二副本中提供该掩码的第二部分。 网络设备还将分组的第一副本转发到在掩码的第一部分中提供的第一出口分组转发引擎的地址，并将分组的第二副本转发到提供的第二出口分组转发引擎的地址 掩模的第二部分。

公开(公告)号：US08274971B2

公开(公告)日：2012-09-25

申请号：US12419945

申请日：2009-04-07

Applicant: James Battle ,  Daniel Tai

Inventor： James Battle ,  Daniel Tai

IPC: H04L12/50

CPC classification number: H04L12/5601 ,  H04L12/18 ,  H04L12/1881 ,  H04L45/00 ,  H04L45/245 ,  H04L45/54 ,  H04L45/745 ,  H04L47/10 ,  H04L47/2408 ,  H04L47/2425 ,  H04L47/2441 ,  H04L47/245 ,  H04L47/286 ,  H04L47/30 ,  H04L47/32 ,  H04L47/6205 ,  H04L47/6215 ,  H04L47/6225 ,  H04L49/101 ,  H04L49/201 ,  H04L49/203 ,  H04L49/205 ,  H04L49/25 ,  H04L49/254 ,  H04L49/30 ,  H04L49/309 ,  H04L49/352 ,  H04L49/608 ,  H04L49/90 ,  H04L49/901 ,  H04L49/9094 ,  H04L2012/5651 ,  H04L2012/5665 ,  H04L2012/5679 ,  H04L2012/5681 ,  H04L2012/5683 ,  Y02D50/30

Abstract: A method for controlling a flow of packet data in a memory management unit of a network switch fabric is disclosed. A first portion of a data packet is received at a port on an ingress bus ring of the network switch fabric. A class of service for the data packet is determined based on the first portion and the portion is stored in a packer RAM of the port based on the class of service. Subsequent portions of the data packet are stored in the packer RAM. Once the predetermined number of portions have been received, the predetermined number of portions is sent to a packet pool RAM. A reference pointer to a first predetermined number of portions is sent to a transaction queue once an end of packet is detected and an egress scheduler detects a presence of a ready packet in the transaction queue and notifies an unpacker of the ready packet. The unpacker puts the ready packet into a FIFO and the ready packet is sent to an ingress/egress module.

Abstract translation: 公开了一种用于控制网络交换结构的存储器管理单元中的分组数据流的方法。 在网络交换结构的入口总线环上的端口处接收数据分组的第一部分。 基于第一部分确定数据包的一类服务，并且该部分基于服务等级被存储在端口的封隔器RAM中。 数据包的后续部分被存储在封隔器RAM中。 一旦已经接收到预定数量的部分，则将预定数量的部分发送到分组池RAM。 一旦检测到分组的结束，一个第一预定数量的部分的参考指针被发送到事务队列，并且出口调度器检测到事务队列中是否存在准备好的分组，并且向解包器通知就绪分组。 解包器将准备好的数据包放入FIFO中，并将准备好的数据包发送到入口/出口模块。

公开(公告)号：US08218440B2

公开(公告)日：2012-07-10

申请号：US11396617

申请日：2006-04-04

Applicant: William Dai

Inventor： William Dai

IPC: G08C15/00

CPC classification number: H04L49/505 ,  H04L49/201 ,  H04L49/352 ,  H04L49/354

Abstract: A high speed transmission protocol that is implemented in a network device that includes a plurality of modules which are connected by a plurality of high speed links. The protocol includes a transmission link aspect for providing at least one of variable-sized packet based transmission with fixed sized messaging capability and pre-emptive transmission capability and a fabric forwarding aspect supporting at least one of class differentiations for packet flows, a plurality of addressable physical and logical modules, generic multicast forwarding port level indication for physical or logical ports, and explicit parameter for packet-content agnostic fabric operation. The protocol also include a packet processing descriptor aspect for providing at least one of a flexibility for various packet-processing descriptor adaptations and packet processing flow continuity across the network device for system design scalability. The protocol further includes an in-band messaging aspect for providing at least one of congestion management protocols, system resiliency protocols, database synchronization protocols and component access protocols and an encoding aspect for providing a structured header design.

Abstract translation: 一种在包括由多个高速链路连接的多个模块的网络设备中实现的高速传输协议。 该协议包括传输链路方面，用于提供具有固定大小的消息传递能力和优先传输能力的可变大小的分组传输中的至少一个以及支持分组流的类别差异中的至少一个的结构转发方面，多个可寻址 物理和逻辑模块，用于物理或逻辑端口的通用组播转发端口级别指示，以及用于分组内容无关结构操作的显式参数。 该协议还包括分组处理描述符方面，用于提供用于各种分组处理描述符适配的灵活性和跨网络设备的分组处理流连续性的系统设计可扩展性中的至少一个。 该协议还包括用于提供拥塞管理协议，系统弹性协议，数据库同步协议和组件访问协议中的至少一个的带内消息传送方面以及用于提供结构化头部设计的编码方面。

公开(公告)号：US20120155473A1

公开(公告)日：2012-06-21

申请号：US13406145

申请日：2012-02-27

Applicant: Norival R. FIGUEIRA ,  Paul A. BOTTORFF ,  Guyves ACHTARI ,  Huiwen LI

Inventor： Norival R. FIGUEIRA ,  Paul A. BOTTORFF ,  Guyves ACHTARI ,  Huiwen LI

IPC: H04L12/56

CPC classification number: H04L49/602 ,  H04L45/74 ,  H04L49/201 ,  H04L61/2015

Abstract: Frame contained destination information may be used by a switch to identify an appropriate output port for a given frame without performing a table access operation. This reduces the processing requirements of the switch to enable the switch to handle frames more efficiently. The frame contained destination information may be contained in the frame's local destination MAC addresses (DA) such that a portion of the DA directly indicates, for each switch that handles the frame, an output port for that switch. Different portions of the DA may be used by different switches, depending on where they are in the network hierarchy. Large switches may also use sub-fields within their allocated portion in the DA to identify internal switching components. A location resolution server may be provided to store and distribute IP and MAC addresses and respond to local ARP requests on the local domain.

Abstract translation: 可以由交换机使用帧包含的目的地信息来识别给定帧的适当输出端口，而不执行表访问操作。 这减少了开关的处理要求，使开关更有效地处理框架。 帧的目的地信息可以包含在帧的本地目的地MAC地址（DA）中，使得DA的一部分对于处理帧的每个交换机直接指示用于该交换机的输出端口。 取决于它们在网络层次中的位置，DA的不同部分可以被不同的交换机使用。 大型交换机也可以使用它们在DA中分配的部分内的子场来识别内部交换组件。 可以提供位置解析服务器来存储和分发IP和MAC地址并响应本地域上的本地ARP请求。

公开(公告)号：US08194561B2

公开(公告)日：2012-06-05

申请号：US12559277

申请日：2009-09-14

Applicant: Hiroyuki Kogata ,  Akio Shinohara

Inventor： Hiroyuki Kogata ,  Akio Shinohara

IPC: H04L12/54

CPC classification number: H04L47/32 ,  H04L45/00 ,  H04L47/10 ,  H04L49/201 ,  H04L49/3009

Abstract: A packet relay device receives an input packet via a first communication port out of a plurality of communication ports of the packet relay device and transmits the input packet via a second communication port out of the plurality of communication ports. The packet relay device includes a learning table storage, a header adder, a discard determiner, and a packet reducer. The learning table storage stores a source address contained in the input packet, in association with an input port identifier capable of identifying the first communication port. The header adder adds an internal transfer header containing the input port identifier to the input packet to generate an internal transfer packet. The discard determiner determines whether to discard the internal transfer packet. The packet reducer reduces the internal transfer packet to preserve at least the source address and the input port identifier upon determining to discard the internal transfer packet.

Abstract translation: 分组中继装置经由分组中继装置的多个通信端口经由第一通信端口接收输入分组，并经由多个通信端口中的第二通信端口发送输入分组。 分组中继装置包括学习表存储，标题加法器，丢弃确定器和分组减少器。 学习表存储器与能够识别第一通信端口的输入端口标识符相关联地存储包含在输入分组中的源地址。 标题加法器将包含输入端口标识符的内部传送标头添加到输入分组以生成内部传送分组。 丢弃决定器确定是否丢弃内部传输数据包。 分组减少器减少内部传输分组，以在确定丢弃内部传送分组时至少保留源地址和输入端口标识符。

公开(公告)号：US08160047B2

公开(公告)日：2012-04-17

申请号：US12903792

申请日：2010-10-13

Applicant: Junwei Wang ,  Xiaoan Fan ,  Jianghua Liu

Inventor： Junwei Wang ,  Xiaoan Fan ,  Jianghua Liu

IPC: H04J3/00

CPC classification number: H04L5/14 ,  H04J11/00 ,  H04L5/0007 ,  H04L5/003 ,  H04L5/0091 ,  H04L5/22 ,  H04L49/201 ,  H04W4/06 ,  H04W72/005 ,  H04W72/02 ,  H04W72/042 ,  H04W72/0446 ,  H04W72/0493

Abstract: A method and device for service time division multiplexing as well as a method and a device for transmitting service are disclosed. The method for service time division multiplexing includes: selecting a part or all of radio frames in one time unit as specific radio frames; and selecting a part or all of subframes in the specific radio frames as specific subframes for sending a specific service. The specific service is a multimedia broadcast multicast service, or a unicast service, or one or more than one kind of service transmitted in broadcast or multicast mode.