公开(公告)号：US08503303B2

公开(公告)日：2013-08-06

申请号：US12989705

申请日：2008-04-25

申请人： András Császár ,  Attila Báder ,  Csaba Antal

发明人： András Császár ,  Attila Báder ,  Csaba Antal

IPC分类号： H04L12/26 ,  H04L12/56

CPC分类号： H04L47/10 ,  H04L12/18 ,  H04L47/12 ,  H04L47/15 ,  H04L47/31

摘要： The invention relates to techniques for traffic handling in congestion situations in a point-to-multipoint (‘PTM’) enabled network. A method embodiment of the inventive technique is performed in an egress node of the network and comprises the steps of detecting marked packets, the marking being indicative of a congestion situation in a core node of the PTM-enabled network; selecting, based on the detected markings, a particular PTM flow from one or more PTM flows passing the egress node for termination; and indicating the selected PTM flow in an upstream direction of the selected flow.

摘要翻译： 本发明涉及在点对多点（“PTM”）使能网络中的拥塞情况下的业务处理技术。 在网络的出口节点中执行本发明技术的方法实施例，并且包括以下步骤：检测标记分组，该标记指示PTM使能网络的核心节点中的拥塞情况; 基于检测到的标记，从经过出口节点的一个或多个PTM流中选择特定的PTM流来终止; 并指示所选择的流的上游方向上的所选PTM流。

公开(公告)号：US08411688B2

公开(公告)日：2013-04-02

申请号：US12933200

申请日：2009-03-16

申请人： János Farkas ,  Csaba Antal ,  Attila Takács

发明人： János Farkas ,  Csaba Antal ,  Attila Takács

IPC分类号： H04L12/28

CPC分类号： H04L12/4645 ,  H04L29/12349 ,  H04L45/02 ,  H04L45/22 ,  H04L45/28 ,  H04L45/66 ,  H04L61/2507 ,  H04L2212/00

摘要： A method is provided for use at a first node of a network. The first node (B) has an input path for receiving Ethernet frames destined for a second node of the network. The first node (B) also has a first output (working) path for forwarding the received Ethernet frames towards the second node. A failure along the first output (working) path is detected at the first node. Following such failure detection, the received Ethernet frames are caused, at the first node, to be re-routed along a second output (backup) path, different to the first output (working) path, towards the second node. The re-routing is performed by using (e.g. referring to and/or changing) respective indications (VLAN1) applied to the Ethernet frames to distinguish those Ethernet frames to be forwarded along the first output (working) path from those Ethernet frames to be forwarded along the second output (backup) path.

摘要翻译： 提供了一种在网络的第一节点使用的方法。 第一节点（B）具有用于接收去往网络的第二节点的以太网帧的输入路径。 第一节点（B）还具有用于将接收到的以太网帧转发到第二节点的第一输出（工作）路径。 在第一个节点处检测到沿着第一个输出（工作）路径的故障。 在这种故障检测之后，在第一节点处引起接收的以太网帧沿着与第一输出（工作）路径不同的第二输出（备用）路径被重新路由到第二节点。 通过使用（例如参考和/或改变）应用于以太网帧的各个指示（VLAN1）来执行重新路由以区分将沿着要转发的那些以太网帧沿着第一输出（工作）路径转发的以太网帧 沿着第二个输出（备份）路径。

公开(公告)号：US20110116418A1

公开(公告)日：2011-05-19

申请号：US13054430

申请日：2008-11-15

申请人： János Farkas ,  Csaba Antal ,  Attila Takács

发明人： János Farkas ,  Csaba Antal ,  Attila Takács

IPC分类号： H04L12/28

CPC分类号： H04L45/48 ,  H04L12/462 ,  H04L12/4641 ,  H04L45/02 ,  H04L45/66

摘要： A registration protocol for setting up and updating a tree that defines the active topology in an Ethernet network includes a plurality of application components and a plurality of attribute declaration components corresponding to the application components, wherein the application components declare and register attribute values and the corresponding attribute declaration components make and withdraw registration of the attributes.

摘要翻译： 用于设置和更新在以太网中定义活动拓扑的树的注册协议包括多个应用组件和对应于应用组件的多个属性声明组件，其中应用组件声明并注册属性值和对应的 属性声明组件制作和撤销属性的注册。

公开(公告)号：US08155030B2

公开(公告)日：2012-04-10

申请号：US12523781

申请日：2008-01-18

申请人： János Farkas ,  Csaba Antal ,  Attila Takács ,  Panagiotis Saltsidis

发明人： János Farkas ,  Csaba Antal ,  Attila Takács ,  Panagiotis Saltsidis

IPC分类号： H04L12/28 ,  H04L12/56

CPC分类号： H04L12/462 ,  H04L45/02 ,  H04L45/025 ,  H04L45/04 ,  H04L45/48

摘要： There is disclosed a manner of managing the active topology, that is the scheme for forwarding data, in a computer network such as an Ethernet local area network. A root bridge is selected from a plurality of bridges in the network; the root bridge calculating at least one spanning tree from a topology database that has been populated by bridge neighbor information gleaned from link state advertising messages. The root bridge may also calculate an alternate topology for implementation in the event a failure condition that affects the primary spanning tree is detected. The root bridge then advertises the spanning tree and remotely configures the port states of the bridges in the spanning tree through tree advertising messages that the bridges are arranged to process and set their ports accordingly.

摘要翻译： 公开了在诸如以太网局域网的计算机网络中管理活动拓扑的方式，即用于转发数据的方案。 从网络中的多个桥中选择根桥; 根桥从已经从链接状态广告消息收集的桥接邻居信息填充的拓扑数据库中计算至少一个生成树。 如果检测到影响主生成树的故障条件，则根桥也可以计算备用拓扑。 根桥然后通告生成树，并通过树形广告消息远程配置生成树中的网桥的端口状态，桥梁被布置为相应地处理和设置其端口。

公开(公告)号：US20080291822A1

公开(公告)日：2008-11-27

申请号：US11916565

申请日：2005-06-14

申请人： Janos Farkas ,  Csaba Antal ,  Lars Westberg

发明人： Janos Farkas ,  Csaba Antal ,  Lars Westberg

IPC分类号： H04L12/24

CPC分类号： H04L12/4641 ,  H04L12/462 ,  H04L41/0213 ,  H04L41/06 ,  H04L43/10 ,  H04L45/00

摘要： The invention relates to failure handling in a tree sructure network (NW1) that has edge nodes (EN1 . . . EN4) and switching nodes (SW1 . . . SW4) interconnected by lines (L1). VLANs (VLAN1-VLAN3) are established such that at least one thereof provides connectivity in case of any single failure in the network. The VLANs can be established by using spanning trees (ST1,ST2,ST3). Among the edge nodes, emitters (EN3) broadcast alive messages (A1,A2,A3) regularly on the VLANs and notifiers (EN2) note the alive messages. A missing alive message indicates a failure (CD1) on one VLAN (VLAN2) and the notifier (EN2) broadcasts corresponding failure messages (F1,F2,F3) on the VLANs. When all the alive messages (A1,A2,A3) appear again the notifier (EN2) broadcasts corresponding repair messages (R1,R2,R3). If the notifiers don't note a failure the nodes (EN1,EN4) with no special role performs a similar function as the notifier (EN2) somewhat slower. The failure handling is fast, robust, uses few messages, increases only slightly the traffic load in the network (NW1) and is compliant with present standards.

摘要翻译： 本发明涉及具有通过线路（L1）互连的边缘节点（EN1 ... EN4）和交换节点（SW1 ... SW4）的树结构网络（NW1）中的故障处理。 建立VLAN（VLAN1-VLAN3），使得其中的至少一个在网络中的任何单个故障的情况下提供连接性。 VLAN可以通过生成树（ST1，ST2，ST3）建立。 在边缘节点中，发射器（EN3）定期在VLAN和通知器（EN2）上广播有效消息（A1，A2，A3），注意活动消息。 丢失的活着消息表示一个VLAN（VLAN2）发生故障（CD1），通知器（EN2）广播VLAN上相应的故障消息（F1，F2，F3）。 当所有活动消息（A1，A2，A3）再次出现时，通知器（EN2）广播相应的修复消息（R1，R2，R3）。 如果通知程序不记录故障，则没有特殊角色的节点（EN1，EN4）执行与通知程序（EN2）类似的功能稍慢。 故障处理快速，稳健，使用少量消息，仅增加网络流量负载（NW1），符合现行标准。

公开(公告)号：US20060198304A1

公开(公告)日：2006-09-07

申请号：US10595092

申请日：2004-07-27

申请人： Gergely Matefi ,  Csaba Antal ,  Janos Farkas

发明人： Gergely Matefi ,  Csaba Antal ,  Janos Farkas

IPC分类号： H04L12/26 ,  H04L12/54

CPC分类号： H04L47/10 ,  H04L47/14 ,  H04L47/15 ,  H04L47/2408 ,  H04L47/2416 ,  H04L47/283 ,  H04L47/70 ,  H04L47/801 ,  H04L47/805 ,  H04L47/822 ,  H04L47/824

摘要： A call admission control system and method for Internet Protocol (IP) Differentiated Services (DiffServ) network having at least one node for interpreting signaling messages and controlling traffic load in the network. The method consists of an initialization (601) and a real-time phase (602). In initialization phase (601), coefficients of the approximating hyperplanes are computed (61) and stored (62). This phase is repeated when the descriptor of a traffic class changes (63), which usually happens when nodes are configured or reconfigured. A traffic mix is admissible (67), if for each real-time traffic class both the stability (65) and the delay (66) constraints are fulfilled. Stability is tested by evaluating the number of lost packets and comparing it to the tolerated packet loss ratio for each class in that queue. Delay constraint is tested by checking if the traffic mix is below at least one of the approximating hyperplanes in the space of number of sessions for each class.

摘要翻译： 一种用于互联网协议（IP）差分服务（DiffServ）网络的呼叫接纳控制系统和方法，其具有用于解释信令消息和控制网络中的业务负载的至少一个节点。 该方法由初始化（601）和实时阶段（602）组成。 在初始化阶段（601）中，计算（61）并存储（62）近似超平面的系数。 当业务类的描述符改变（63）时，会重复此阶段，通常在配置或重新配置节点时发生。 如果对于每个实时交通类来说，流量混合是可接受的（67），则稳定性（65）和延迟（66）约束都得到满足。 通过评估丢失数据包的数量并将其与该队列中的每个类别的容忍丢包率进行比较来测试稳定性。 通过检查流量混合是否低于每个类的会话空间中的近似超平面中的至少一个来测试延迟约束。

公开(公告)号：US20050169179A1

公开(公告)日：2005-08-04

申请号：US10939970

申请日：2004-09-14

申请人： Csaba Antal ,  Janos Harmatos ,  Alpar Juttner ,  Lars Westberg

发明人： Csaba Antal ,  Janos Harmatos ,  Alpar Juttner ,  Lars Westberg

IPC分类号： H04L12/24 ,  H04L12/26 ,  H04L12/28

CPC分类号： H04L41/0806 ,  H04L41/5003 ,  H04L41/5009 ,  H04L41/5054

摘要： In the area of network provisioning, there is a problem of selecting a suitable traffic-provisioning model for large networks due to the high management complexity of the resource-efficient trunk model and the poor bandwidth efficiency of the easy-to-configure hose model. The invention is based on the idea of partitioning at least part of the network into multi-node clusters, and defining traffic limitations on at least two levels, including the intra-cluster level and the inter-cluster level, where the traffic limitations include one or more node-to-cluster traffic limitations for inter-cluster traffic. Subsequently, cluster-based provisioning of the network is performed based on the traffic limitations. The novel node-to-cluster limitations proposed by the invention are preferably applied in a cluster-based trunk or hose model on the inter-cluster level. In other words, for the description of the inter-cluster traffic (traffic between the clusters) cluster-based trunk or hose models can be used, preferably depending on the available information about the traffic. The cluster-based provisioning makes it possible to find a trade-off between management complexity and overprovisioning.

摘要翻译： 在网络配置领域，由于资源高效的主干模型的高管理复杂性和易于配置的软管模型的较差带宽效率，存在为大型网络选择合适的流量供应模型的问题。 本发明基于将网络的至少一部分划分成多节点集群并且在包括集群内级别和集群间级别的至少两个级别上定义业务限制的想法，其中流量限制包括一个 或更多的群集间流量的节点到群集流量限制。 随后，基于流量限制来执行基于群集的网络供应。 本发明提出的新颖的节点到群集限制优选地应用于群集间级别上的基于群集的中继线或软管模型。 换句话说，对于集群间流量（集群之间的流量）的描述，可以使用基于集群的中继或软件模型，优选地取决于关于流量的可用信息。 基于集群的配置使得可以在管理复杂性和过度配置之间找到权衡。

公开(公告)号：US08837329B2

公开(公告)日：2014-09-16

申请号：US12525062

申请日：2008-01-18

申请人： János Farkas ,  Csaba Antal ,  Attila Takács ,  Panagiotis Saltsidis

发明人： János Farkas ,  Csaba Antal ,  Attila Takács ,  Panagiotis Saltsidis

IPC分类号： H04L12/28 ,  H04L12/705 ,  H04L12/721 ,  H04L12/751 ,  H04L12/703 ,  H04L12/753 ,  H04L12/46 ,  H04L12/24

CPC分类号： H04L12/462 ,  H04L41/069 ,  H04L41/0853 ,  H04L41/0856 ,  H04L41/12 ,  H04L45/02 ,  H04L45/12 ,  H04L45/18 ,  H04L45/28 ,  H04L45/48

摘要： There is disclosed a method and apparatus for facilitating a network, such as an Ethernet LAN, for efficient forwarding of data traffic by collecting neighbor information, generating and distributing link state advertisements, populating a topology database for the network, and calculating trees for each bridge serving as a root bridge. In a preferred embodiment the method and system also prepare one or more backup topologies, and store them for use if they are needed due to a failure condition detected in the network. In a particularly preferred embodiment, probabilities are assigned to various potential failure conditions, and the probability values are used to decide which backup topologies to calculate, store, or use.

摘要翻译： 公开了一种方法和装置，用于通过收集邻居信息，生成和分发链路状态通告，填充网络的拓扑数据库以及计算每个网桥的树，从而方便诸如以太网LAN的网络来有效地转发数据流量 作为根桥。 在优选实施例中，方法和系统还准备一个或多个备用拓扑，并且如果由于在网络中检测到故障状况而需要，则存储它们以供使用。 在特别优选的实施例中，将概率分配给各种潜在的故障条件，并且概率值用于决定要计算，存储或使用哪些备份拓扑。

公开(公告)号：US08467317B2

公开(公告)日：2013-06-18

申请号：US13414927

申请日：2012-03-08

申请人： János Farkas ,  Csaba Antal ,  Attila Takács ,  Panagiotis Saltsidis

发明人： János Farkas ,  Csaba Antal ,  Attila Takács ,  Panagiotis Saltsidis

IPC分类号： H04L12/28 ,  H04L12/56

CPC分类号： H04L12/462 ,  H04L45/02 ,  H04L45/025 ,  H04L45/04 ,  H04L45/48

摘要： There is disclosed a manner of managing the active topology, that is the scheme for forwarding data, in a computer network such as an Ethernet local area network. A root bridge is selected from a plurality of bridges in the network; the root bridge calculating at least one spanning tree from a topology database that has been populated by bridge neighbor information gleaned from link state advertising messages. The root bridge may also calculate an alternate topology for implementation in the event a failure condition that affects the primary spanning tree is detected. The root bridge then advertises the spanning tree and remotely configures the port states of the bridges in the spanning tree through tree advertising messages that the bridges are arranged to process and set their ports accordingly.

摘要翻译： 公开了在诸如以太网局域网的计算机网络中管理活动拓扑的方式，即用于转发数据的方案。 从网络中的多个桥中选择根桥; 根桥从已经从链接状态广告消息收集的桥接邻居信息填充的拓扑数据库中计算至少一个生成树。 如果检测到影响主生成树的故障条件，则根桥也可以计算备用拓扑。 根桥然后通告生成树，并通过树形广告消息远程配置生成树中的网桥的端口状态，桥梁被布置为相应地处理和设置其端口。

公开(公告)号：US20120170489A1

公开(公告)日：2012-07-05

申请号：US13414927

申请日：2012-03-08

申请人： János Farkas ,  Csaba Antal ,  Attila Takács ,  Panagiotis Saltsidis

发明人： János Farkas ,  Csaba Antal ,  Attila Takács ,  Panagiotis Saltsidis

IPC分类号： H04L12/44

CPC分类号： H04L12/462 ,  H04L45/02 ,  H04L45/025 ,  H04L45/04 ,  H04L45/48

摘要： There is disclosed a manner of managing the active topology, that is the scheme for forwarding data, in a computer network such as an Ethernet local area network. A root bridge is selected from a plurality of bridges in the network; the root bridge calculating at least one spanning tree from a topology database that has been populated by bridge neighbor information gleaned from link state advertising messages. The root bridge may also calculate an alternate topology for implementation in the event a failure condition that affects the primary spanning tree is detected. The root bridge then advertises the spanning tree and remotely configures the port states of the bridges in the spanning tree through tree advertising messages that the bridges are arranged to process and set their ports accordingly.

摘要翻译： 公开了在诸如以太网局域网的计算机网络中管理活动拓扑的方式，即用于转发数据的方案。 从网络中的多个桥中选择根桥; 根桥从已经从链接状态广告消息收集的桥接邻居信息填充的拓扑数据库中计算至少一个生成树。 如果检测到影响主生成树的故障条件，则根桥也可以计算备用拓扑。 根桥然后通告生成树，并通过树形广告消息远程配置生成树中的网桥的端口状态，桥梁被布置为相应地处理和设置其端口。