公开(公告)号：US20100182907A1

公开(公告)日：2010-07-22

申请号：US12514016

申请日：2006-11-09

申请人： Ferenc Pinter ,  Attila Bader ,  Andras Csaszar ,  Attila Takacs

发明人： Ferenc Pinter ,  Attila Bader ,  Andras Csaszar ,  Attila Takacs

IPC分类号： H04L12/56

CPC分类号： H04L47/26 ,  H04L47/18 ,  H04L47/70

摘要： Reducing congestion in an IP domain wherein congested data flows arriving at an egress edge node of the network are identified. Total congestion represented by the congested flows is determined, and a congestion extent notification is sent from the egress edge node to its ingress edge node peer. The congestion extent notification includes information regarding the total congestion and is sent on a per-class basis. Congested core routers in the network insert DSCPs into data packets passing through them to enable the egress edge nodes to identify the affected flows. The core routers may also send congestion metric messages, designed to follow the same path as the marked packets, to inform the egress edge nodes of the extent of congestion. In an alternative method, the egress edge nodes inform their ingress peers that congestion is present, without initially identifying the extent. The ingress edge node sends a query downstream with a congestion metric.

摘要翻译： 识别其中拥塞数据流到达网络的出口边缘节点的IP域中的拥塞。 确定拥塞流表示的总拥塞，并且从出口边缘节点向其入口边缘节点对等体发送拥塞范围通知。 拥堵范围通知包括关于总拥塞的信息，并且以每个等级的形式发送。 网络中的拥塞核心路由器将DSCP插入到通过它们的数据包中，以使出口边缘节点能够识别受影响的流。 核心路由器还可以发送设计为遵循与标记分组相同的路径的拥塞度量消息，以向出口边缘节点通知拥塞程度。 在另一种方法中，出口边缘节点通知他们的入口对等体，拥塞是存在的，而没有初始地识别该范围。 入口边缘节点向下游发送具有拥塞度量的查询。

公开(公告)号：US08009568B2

公开(公告)日：2011-08-30

申请号：US12514016

申请日：2006-11-09

申请人： Ferenc Pinter ,  Attila Bader ,  Andras Csaszar ,  Attila Takacs

发明人： Ferenc Pinter ,  Attila Bader ,  Andras Csaszar ,  Attila Takacs

IPC分类号： G01R31/08

CPC分类号： H04L47/26 ,  H04L47/18 ,  H04L47/70

摘要： Reducing congestion in an IP domain wherein congested data flows arriving at an egress edge node of the network are identified. Total congestion represented by the congested flows is determined, and a congestion extent notification is sent from the egress edge node to its ingress edge node peer. The congestion extent notification includes information regarding the total congestion and is sent on a per-class basis. Congested core routers in the network insert DSCPs into data packets passing through them to enable the egress edge nodes to identify the affected flows. The core routers may also send congestion metric messages, designed to follow the same path as the marked packets, to inform the egress edge nodes of the extent of congestion. In an alternative method, the egress edge nodes inform their ingress peers that congestion is present, without initially identifying the extent. The ingress edge node sends a query downstream with a congestion metric.

摘要翻译： 识别其中拥塞数据流到达网络的出口边缘节点的IP域中的拥塞。 确定拥塞流表示的总拥塞，并且从出口边缘节点向其入口边缘节点对等体发送拥塞范围通知。 拥堵范围通知包括关于总拥塞的信息，并且以每个等级的形式发送。 网络中的拥塞核心路由器将DSCP插入到通过它们的数据包中，以使出口边缘节点能够识别受影响的流。 核心路由器还可以发送设计为遵循与标记分组相同的路径的拥塞度量消息，以向出口边缘节点通知拥塞程度。 在另一种方法中，出口边缘节点通知他们的入口对等体，拥塞是存在的，而没有初始地识别出该范围。 入口边缘节点向下游发送具有拥塞度量的查询。

公开(公告)号：US08509085B2

公开(公告)日：2013-08-13

申请号：US12514046

申请日：2006-11-10

申请人： Attila Bader ,  Attila Takacs ,  Andras Csaszar

发明人： Attila Bader ,  Attila Takacs ,  Andras Csaszar

IPC分类号： G01R31/08

CPC分类号： H04L47/12 ,  H04L47/10 ,  H04L47/11 ,  H04L47/26 ,  H04L47/31 ,  H04W36/00

摘要： An egress node is proposed for a network domain that comprises at least one ingress node for receiving and routing data units belonging to a plurality of flows into said network domain, a plurality of interior routers for routing said data units through said network domain, and said egress node(s). One or more of the interior routers is arranged to detect whether it is subject to a congestion condition and to congestion mark routed data units if it is subject to said congestion condition. The egress node comprises a control element arranged for performing a congestion handling function that comprises a part for detecting the presence of congestion marked data units arriving at said edge node, and a part for reacting to the detection of the presence of congestion marked data units by invoking a congestion control process. The congestion control comprises sending to said ingress node an instruction for terminating one or more flows to thereby reduce a traffic load, waiting a predetermined period of time and then determining whether congestion marked data units are still arriving at said edge node, and if congestion marked data units are still arriving, repeating the sending and waiting steps.

摘要翻译： 提出了一种用于网络域的出口节点，其包括至少一个入口节点，用于接收和路由属于多个流的数据单元到所述网络域中，多个内部路由器用于通过所述网络域路由所述数据单元，并且所述多个内部路由器 出口节点。 内部路由器中的一个或多个被布置成检测它是否遭受拥塞状况，并且如果它遭受所述拥塞状况则阻塞路由数据单元。 出口节点包括一个控制元件，用于执行拥塞处理功能，该功能包括用于检测到达所述边缘节点的拥塞标记数据单元的存在的部分，以及响应于通过以下方式检测拥塞标记数据单元的部分： 调用拥塞控制过程。 拥塞控制包括向所述入口节点发送用于终止一个或多个流的指令，从而减少业务负载，等待预定时间段，然后确定拥塞标记数据单元是否仍然到达所述边缘节点，并且如果拥塞标记 数据单元仍然到达，重复发送和等待步骤。

公开(公告)号：US08446826B2

公开(公告)日：2013-05-21

申请号：US11718854

申请日：2004-11-12

申请人： Andras Csaszar ,  Attila Takacs ,  Attila Bader ,  Robert Szabo ,  Lars Westberg

发明人： Andras Csaszar ,  Attila Takacs ,  Attila Bader ,  Robert Szabo ,  Lars Westberg

IPC分类号： H04L12/26

CPC分类号： H04L47/115 ,  H04L43/026 ,  H04L43/0882 ,  H04L47/10 ,  H04L47/11 ,  H04L47/31 ,  Y02D50/30

摘要： A system and a method are described herein which provide for congestion handling in a packet switched network domain. In case of congestion, overload is measured by a core node, the data packets in proportion to the overload are marked and the signaled overload is stored. At least one egress node receives marked and not marked packets, decodes and counts the overload from the marked packets in a counting interval. Congestion report messages are sent to ingress nodes where flows are terminated.

摘要翻译： 这里描述了一种在分组交换网络域中提供拥塞处理的系统和方法。 在拥塞的情况下，由核心节点测量过载，与过载成比例的数据包被标记，并且信号过载被存储。 至少一个出口节点接收标记和未标记的分组，在计数间隔中对标记的分组进行解码和计数。 拥塞报告消息被发送到流终止的入口节点。

公开(公告)号：US20070268827A1

公开(公告)日：2007-11-22

申请号：US11718854

申请日：2004-11-12

申请人： Andras Csaszar ,  Attila Takacs ,  Attila Bader ,  Robert Szabo ,  Lars Westberg

发明人： Andras Csaszar ,  Attila Takacs ,  Attila Bader ,  Robert Szabo ,  Lars Westberg

IPC分类号： H04L12/56

CPC分类号： H04L47/115 ,  H04L43/026 ,  H04L43/0882 ,  H04L47/10 ,  H04L47/11 ,  H04L47/31 ,  Y02D50/30

摘要： A system a core (104), an edge node (102, 103) and a method are provided for congestion handling in a packet switched network domain. In case of congestion overload is measured, the data packets in proportion to the overload are marked and the signaled overload is stored. At least one egress node (103) receives marked and not marked packets, decodes and counts the overload from the marked packets in a counting interval. Congestion report messages are sent to ingress nodes (102) where flows are terminated. A core node (104) comprises a storage means (314) taking into account the previously signaled overload, which are already being handled in a congestion handling control loop. An edge node (103) comprises means for adding new flows to the set of affected flows, means for identifying affected flows to be terminated so that the number of identified affected flows should generate together traffic equal to the overload, and means for removing flows to be terminated from set of affected flows. In an advantageous embodiment the market module (315) applies a hysteresis algorithm with a higher bound used to detect congestion and to trigger the marking of user data packets in proportion to the measured overload, and a lower bound used as the reference to measure the overload ratio.

摘要翻译： 提供了一种核心（104），边缘节点（102,103）和方法，用于分组交换网络域中的拥塞处理。 在测量拥塞过载的情况下，标记与过载成比例的数据包，并存储信号过载。 至少一个出口节点（103）接收标记和未标记的分组，在计数间隔中对标记的分组进行解码和计数过载。 拥塞报告消息被发送到流停止的入口节点（102）。 核心节点（104）包括存储装置（314），其考虑了已经在拥塞控制控制环路中处理的先前发信号的过载。 边缘节点（103）包括用于将新的流添加到受影响的流集合的装置，用于识别待终止的受影响的流的装置，使得所识别的受影响的流的数量应该一起生成等于过载的流量，以及用于去除流到 由一组受影响的流量终止。 在有利的实施例中，市场模块（315）应用具有用于检测拥塞的较高边界的滞后算法，并且与所测量的过载成比例地触发用户数据分组的标记，并且将下限用作测量过载的参考 比。

公开(公告)号：US20090274165A1

公开(公告)日：2009-11-05

申请号：US12064127

申请日：2005-08-23

申请人： Andras Csaszar ,  Attila Takacs ,  Attila Bader

发明人： Andras Csaszar ,  Attila Takacs ,  Attila Bader

IPC分类号： H04B7/212

CPC分类号： H04L47/72 ,  H04L47/70 ,  H04L47/724 ,  H04L47/823

摘要： A network node for reserving resources for data flows in a communication network. The node detects a request for resource reservation for a new data flow and computes an admission test count based on descriptors of the new data flow and reservation counts, which represent reservation functions of descriptors of previously admitted data flows and which specify resources of the admitted data flows. The new data flow is admitted if the admission test count is smaller than a maximum allowable count. The reservation counts are then updated based on the descriptors of the new data flow. A plurality of admission formulas may thus be used to maintain aggregated reservation counts for data flows, the number of counts being independent from the number of flows.

摘要翻译： 一种用于在通信网络中预留数据流的资源的网络节点。 节点检测对新数据流的资源预留请求，并且基于新数据流和预留计数的描述符来计算准入测试计数，其表示先前允许的数据流的描述符的预留功能，并且指定允许数据的资源 流动。 如果入学测试计数小于最大允许计数，则允许新的数据流。 然后基于新数据流的描述符更新预约计数。 因此，可以使用多个允许公式来维护数据流的聚合预留计数，计数的数量与流的数量无关。

公开(公告)号：US07849215B2

公开(公告)日：2010-12-07

申请号：US12295219

申请日：2006-03-31

申请人： Andras Csaszar ,  Attila Takacs ,  Attila Bader

发明人： Andras Csaszar ,  Attila Takacs ,  Attila Bader

IPC分类号： G06F15/173

CPC分类号： H04L47/724 ,  H04L45/02 ,  H04L45/04 ,  H04L45/22 ,  H04L47/70 ,  H04L47/746 ,  H04L47/783

摘要： Methods, edge routers and an edge-router-refresh network signalling message used to update state information in edge routers. A data session is established on a path from a source towards a destination connected from the source via a plurality of Autonomous Systems (AS). The edge-router-refresh network signalling message is created by an edge router acting an an ingress edge router. The edge-router-refresh network signalling message comprises an identifier of the data session, an identifier of the edge router, which issued the edge-router-refresh message and an indication that the edge-router-refresh message is meant to be used by the edge routers present on the modified path. Optionally, the edge-router-refresh network signalling message further comprises a list of the plurality of AS traversed by the path before the modification.

摘要翻译： 方法，边缘路由器和边缘路由器刷新网络信令消息，用于更新边缘路由器中的状态信息。 在从源经由多个自治系统（AS）从源连接的目的地的路径上建立数据会话。 边缘路由器刷新网络信令消息由作为入口边缘路由器的边缘路由器创建。 边缘路由器刷新网络信令消息包括数据会话的标识符，发出边缘路由器刷新消息的边缘路由器的标识符，以及边缘路由器刷新消息将被使用的指示 边缘路由器存在于修改的路径上。 可选地，边缘路由器刷新网络信令消息还包括由修改之前的路径遍历的多个AS的列表。

公开(公告)号：US20100034090A1

公开(公告)日：2010-02-11

申请号：US12514046

申请日：2006-11-10

申请人： Attila Bader ,  Attila Takacs ,  Andras Csaszar

发明人： Attila Bader ,  Attila Takacs ,  Andras Csaszar

IPC分类号： H04L12/56

CPC分类号： H04L47/12 ,  H04L47/10 ,  H04L47/11 ,  H04L47/26 ,  H04L47/31 ,  H04W36/00

摘要： An egress node is proposed for a network domain that comprises at least one ingress node for receiving and routing data units belonging to a plurality of flows into said network domain, a plurality of interior routers for routing said data units through said network domain, and said egress node(s). One or more of the interior routers is arranged to detect whether it is subject to a congestion condition and to congestion mark routed data units if it is subject to said congestion condition. The egress node comprises a control element arranged for performing a congestion handling function that comprises a part for detecting the presence of congestion marked data units arriving at said edge node, and a part for reacting to the detection of the presence of congestion marked data units by invoking a congestion control process. The congestion control comprises sending to said ingress node an instruction for terminating one or more flows to thereby reduce a traffic load, waiting a predetermined period of time and then determining whether congestion marked data units are still arriving at said edge node, and if congestion marked data units are still arriving, repeating the sending and waiting steps.

摘要翻译： 提出了一种用于网络域的出口节点，其包括至少一个入口节点，用于接收和路由属于多个流的数据单元到所述网络域中，多个内部路由器用于通过所述网络域路由所述数据单元，并且所述多个内部路由器 出口节点。 内部路由器中的一个或多个被布置成检测它是否遭受拥塞状况，并且如果它遭受所述拥塞状况则阻塞路由数据单元。 出口节点包括一个控制元件，用于执行拥塞处理功能，该功能包括用于检测到达所述边缘节点的拥塞标记数据单元的存在的部分，以及响应于通过以下方式检测拥塞标记数据单元的部分： 调用拥塞控制过程。 拥塞控制包括向所述入口节点发送用于终止一个或多个流的指令，从而减少业务负载，等待预定时间段，然后确定拥塞标记数据单元是否仍然到达所述边缘节点，并且如果拥塞标记 数据单元仍然到达，重复发送和等待步骤。

公开(公告)号：US20090049194A1

公开(公告)日：2009-02-19

申请号：US12295219

申请日：2006-03-31

申请人： Andras Csaszar ,  Attila Takacs ,  Attila Bader

发明人： Andras Csaszar ,  Attila Takacs ,  Attila Bader

IPC分类号： G06F15/173

CPC分类号： H04L47/724 ,  H04L45/02 ,  H04L45/04 ,  H04L45/22 ,  H04L47/70 ,  H04L47/746 ,  H04L47/783

摘要： Methods, edge routers and an edge-router-refresh network signalling message used to update state information in edge routers. A data session is established on a path from a source towards a destination connected from the source via a plurality of Autonomous Systems (AS). The edge-router-refresh network signalling message is created by an edge router acting an ingress edge router. The edge-router-refresh network signalling message comprises an identifier of the data session, an identifier of the edge router, which issued the edge-router-refresh message and an indication that the edge-router-refresh message is meant to be used by the edge routers present on the modified path. Optionally, the edge-router-refresh network signalling message further comprises a list of the plurality of AS traversed by the path before the modification.

摘要翻译： 方法，边缘路由器和边缘路由器刷新网络信令消息用于更新边缘路由器中的状态信息。 在从源经由多个自治系统（AS）从源连接的目的地的路径上建立数据会话。 边缘路由器刷新网络信令消息由作为入口边缘路由器的边缘路由器创建。 边缘路由器刷新网络信令消息包括数据会话的标识符，发出边缘路由器刷新消息的边缘路由器的标识符，以及边缘路由器刷新消息将被使用的指示 边缘路由器存在于修改的路径上。 可选地，边缘路由器刷新网络信令消息还包括由修改之前的路径遍历的多个AS的列表。

公开(公告)号：US09706270B2

公开(公告)日：2017-07-11

申请号：US12676895

申请日：2008-06-30

申请人： Raimo Vuopionpera ,  Attila Takacs ,  Boris Dortschy ,  Henrik Almeida ,  Attila Bader

发明人： Raimo Vuopionpera ,  Attila Takacs ,  Boris Dortschy ,  Henrik Almeida ,  Attila Bader

IPC分类号： H04L12/28 ,  H04Q11/00 ,  H04L12/707 ,  H04L12/701

CPC分类号： H04Q11/0005 ,  H04L45/00 ,  H04L45/24 ,  H04Q11/0066 ,  H04Q2011/0039 ,  H04Q2011/0073

摘要： A method of routing packets between nodes of a packet switched network, the method comprising receiving a packet at a first node, the packet comprising routing information and payload data, sending said routing information to a second node in the routing path via a first transmission medium having a relatively high transmission speed, and sending said payload data to said second node via a second transmission medium having a relatively low transmission speed. Upon receipt of said routing information, said second node is able to prepare or begin preparation for onward routing of the packet in advance of receipt of said payload.