WO2009023996A1 - Procédé de mise en œuvre d'une interconnexion de réseau par l'intermédiaire d'une agrégation de liaisons - Google Patents

Procédé de mise en œuvre d'une interconnexion de réseau par l'intermédiaire d'une agrégation de liaisons Download PDF

Info

Publication number
WO2009023996A1
WO2009023996A1 PCT/CN2007/003422 CN2007003422W WO2009023996A1 WO 2009023996 A1 WO2009023996 A1 WO 2009023996A1 CN 2007003422 W CN2007003422 W CN 2007003422W WO 2009023996 A1 WO2009023996 A1 WO 2009023996A1
Authority
WO
WIPO (PCT)
Prior art keywords
port
information
aggregation
state
local
Prior art date
Application number
PCT/CN2007/003422
Other languages
English (en)
French (fr)
Inventor
Xiangwu Cui
Xiaosu Wu
Lijun Ma
Kai Zhao
Original Assignee
Zte Corporation
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Zte Corporation filed Critical Zte Corporation
Publication of WO2009023996A1 publication Critical patent/WO2009023996A1/zh

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/06Management of faults, events, alarms or notifications
    • H04L41/0654Management of faults, events, alarms or notifications using network fault recovery

Definitions

  • the present invention relates to the field of communication devices, and in particular to a method of implementing network interconnection using link aggregation.
  • BACKGROUND OF THE INVENTION Communication systems are required to provide users with reliable, uninterrupted service.
  • a common solution to this is redundant backup and active/standby switching. Redundant backup and active/standby switching technologies have multiple backup modes, such as 1+1, 1 : 1, 1 : N, and so on. Specifically, the 1+1 point is a protection mode of the single card.
  • the main/standby unit in the system forms a logical business function unit, that is, the protected device unit has a spare unit, and the main module is responsible for the service.
  • the standby service remains consistent with the data of the active unit, and when the original primary unit is detected to be faulty, the switching is initiated, and the standby unit takes over the service of the original primary unit.
  • the spare unit in the 1 : 1 backup only starts to work after the failure
  • the 1: N backup means that multiple unit units share one spare unit.
  • the standby unit takes over the work.
  • the switching In order to ensure the self-healing capability of the equipment during network failure, the switching must be completed within a specified time to restore the business as soon as possible, minimizing the impact on the communication system.
  • the method to improve network reliability is port bundling (PORT TRUNKING), which is also a low-cost bandwidth expansion solution.
  • Port Bundling also known as Link Aggregation, bundles multiple Ethernet ports to form a logical data link, which improves bandwidth and backups, improving system reliability.
  • the method is widely applied to interconnection of network devices such as network switches, routers, and broadband access servers.
  • Link aggregation technology also known as trunking or bonding, essentially combines several physical links between two devices into one logical link.
  • the current standard for link aggregation technology is the IEEE standard 802.3ad developed by the IEEE 802 committee.
  • the Link Aggregation Control Protocol (LACP) is one of the main contents of the IEEE 802. 3 ad standard and defines a standard aggregation control method.
  • the aggregated devices exchange information through protocols, and automatically aggregate and match the matched links according to the parameters and status of the two parties.
  • the switching device After the aggregation is formed, the switching device maintains the status of the aggregated link. Dynamically adjust or dissolve the aggregated link.
  • the protocol is based on the information and status of the transmission, not the command transmitted.
  • LACPDUs (or LACPDUs) are transmitted by the first party, the Actor, to the information known to them (including their own status and the status of the partner) to the second party, the initiator's partner. The information transmitted by the LACPDU is sufficient for the partner to decide the next step.
  • the LACP-Activity attribute which corresponds to each port. If the value is Passive LACP, the port does not send LACPDU unless the partner takes the value of Active LACP. When the value is Active LACP, it indicates that the port will participate in the protocol regardless of the partner's value. As long as the LACP-Activity of either party is Active LACP, it will cause the LACPDU to be sent periodically. These periodic transmissions are performed at a slow or fast rate, depending on the LACP_Timeout value of the partner. In addition to periodically transmitting LACPDUs, when a need is transmitted (Need To Transmit, abbr. NTT), the protocol sends LACPDUs.
  • LACPDU when the status of the initiator changes, or it is obvious from the partner's LACPDU that it does not know the current status of the initiator.
  • the protocol assumes that the LACPDU loss rate is very low and LACP does not use any explicit frame loss detection/retransmission mechanism.
  • the initiator if the information sent from the partner indicates that it does not have instant information from the initiator, or if the time is sent periodically, the initiator will send an LACPDU to help the partner update the information.
  • the structure of the LACPDU is shown in Figure 1, where Actor_State (local state) and
  • Partner—State Passive state
  • the bit structure is shown in Figure 2.
  • the operation of the protocol is controlled by several state machines, each of which performs different functions. These state machines are port-based, and events such as timer timeouts or receipt of LACPDUs can cause stateful migrations and actions to occur, including sending LACPDUs containing duplicate or updated information.
  • Periodic transmission and event-driven transmission are controlled by the state of the NTT variable.
  • the state machine includes the following types:
  • Receive machine This state machine receives the LACPDU from the partner, records the information contained in it, and then according to The short timer or long timer set by LACP_Timeout is timed. Based on the information received from the partner, the state machine determines whether the consistency of the protocol information exchanged between the initiator and the partner has reached a level at which the port and other ports can be aggregated or as a single port. If not, the state machine will
  • NTT is set to true to send new protocol information to the partner. If the protocol information of the waiting partner times out, the receiving state machine will fill in with the default parameters.
  • Periodic Transmission Machine This state machine decides whether to periodically exchange LACPDUs for maintaining aggregation between the initiator and the receiver.
  • the premise of sending periodic LACPDUs is that at least one party is set to Active LACP.
  • Selection Logic This state machine is responsible for selecting the appropriate aggregate for the port.
  • Mux machine This state machine is responsible for binding a port to a selected aggregate or detaching a port from an aggregate. It is also responsible for enabling or disabling the receiving or sending function of the port according to the current protocol information needs.
  • the two ports are aggregated.
  • the LACP protocol is used to periodically send LACP packets to each other.
  • the sending interval is usually 1 second.
  • Port 1 and port 2 are load sharing ones.
  • 003422 Group working port, the specific reserved physical egress port (port 1 or port 2) is determined by the hardware according to the hash algorithm selected by the MAC layer.
  • the fault occurs: The ports 2 of the devices A and B are faulty, such as the connected physical line is broken in both directions. 3. The ports 2 and B of the device A and B can no longer receive the packets from the peer. When the packets are not received, the device A and B consider that the port is faulty.
  • Port 2 is no longer a working port and does not carry services. Port 1 is the working port. At this point, all the packets sent by the two devices are sent to port 1. carry out.
  • port 2 of device A can also receive LACP protocol packets from device B. After receiving no packet, the device B considers that the port is faulty and stops sending LACP packets to device A. After three more transmission cycles, Device A finds that the port is faulty.
  • the change of port 4 state must be the same at both ends. If the settings of the working ports on the two devices are inconsistent, for example, if the port 1 is the working port and the port 2 is the working port, the service 4 cannot be sent normally, and the service packets are discarded. The protection fails. . Therefore, the protection mode must be switched by the devices at both ends to switch; if the protection action is not completed at any end, the protection cannot be successful. Therefore, in the above case, only when port 2 is found to be faulty, the protection can be completed.
  • the time required from the failure of the second step to the completion of the protection of the fourth step is the network failure recovery time; , the time is at least 3 seconds, and the protection time for one-way fiber break is 6 seconds.
  • the network's failure recovery time is required to be shorter and shorter.
  • the transmission of voice services, recovery time greater than 50nis may lead to application problems, so only the standard definition of LACP protocol can not meet the needs of the development of business applications.
  • the shortcomings of the LACP protocol are:
  • the use of the timeout mechanism to detect faults is very time consuming; there is no fault information notification mechanism, and the peer cannot be notified when an error occurs on the local end.
  • SUMMARY OF THE INVENTION The present invention has been made in view of the above problems in the prior art, and it is therefore intended to provide a method for implementing network interconnection using link aggregation.
  • the monitored state change information is divided into different categories, and related processing is performed according to the category to which the state change information belongs.
  • the categories of the state change information include: an immediate update class, an immediate synchronization class, and a delayed synchronization class.
  • the related process includes: updating the local port state information, modifying the local aggregation information, and starting the protocol sending state machine.
  • the related processing includes: updating the local port state information, starting the protocol sending state, and waiting for the peer to respond, and then modifying the aggregation information.
  • the state change information is the delay synchronization class
  • the related process includes: updating the local port state information, and periodically transmitting the state machine trigger and the peer end message synchronization through the link aggregation control protocol, that is, LACP.
  • the reserved (Reserved) field in the link aggregation control protocol data unit that is, the LACPDU, carries at least one of the following private flags: The aggregation status information flag of the other ports in the aggregation group, the active send/response flag, and whether the status of the aggregation port of the local node changes.
  • the method may include the following operations: assigning an aggregation group ID according to the switching node and a port ID of each aggregation group corresponding to the aggregation; running the L ACP protocol state machine to maintain the aggregated link information; starting a timer to trigger the monitoring of the aggregation state, and receiving Interrupting the reported port status information; The monitored status change information is divided into different categories, and the related processing is performed according to the category to which the status change information belongs.
  • the local end After receiving the LACPDU, the local end first determines whether the local end is in the static aggregation mode. If the result is yes, it is converted to the dynamic aggregation mode, and the private label is determined in the LACPDU.
  • the LACPDU is directly determined whether there is a private flag.
  • the subsequent operation is performed according to the judgment result of whether the private label exists in the LACPDU. .
  • it indicates that the peer end sends the standard LACP protocol packet, and performs subsequent operations according to the standard LACP protocol rule.
  • it is determined that there is a private flag in the LACPDU it indicates that the peer sends an internal protocol packet, and extracts the private flag.
  • the private flag it is determined whether the local port state information needs to be updated, and whether the response processing result needs to be sent to the peer end.
  • the packet is extracted.
  • the receiving port information and other port information in the group are compared with the locally maintained port state information, the local port state information is updated, and the aggregation information is modified according to the updated port state information.
  • the response processing result is sent to the peer.
  • the restart timer triggers the monitoring aggregation state, and receives the port state information reported by the interrupt.
  • the method may further include the following: clearing the port timeout timer; when the port timeout timer arrives, determining whether other ports in the same aggregation group are also in the timeout state, and if yes, switching to the "state aggregation mode"
  • the status of the physical port is determined to be added to the aggregation group (Trunk). If not, the local port is elected from the aggregation group.
  • the switching node is a physical switching network board that supports link aggregation or load balancing.
  • the two physical switching stencils that work, the two physical switching stencils that use the load sharing work adopt the switching chip that supports the stacking technology, and the cascading between the chips is required.
  • the solution provided by the present invention basically follows the existing standards and universal lines. It can be connected to the universal switch running LACP in a trunk mode, which is easy to implement. It achieves the following beneficial effects: (1) The network communication link is effectively backed up, and the upper port can allocate more bandwidth; (2) Network Fast device fault recovery capability, LINK interrupt response, instant message interaction, etc. (3) Distinguish the fault level, restore the link communication capability in time, avoid frequent processing of protocol packets, and maximize product performance. (4) There is no restriction on the connection method. It is a switching network board, which can be physically two switched boards or universal switches; (5) The aggregate link management method is compatible with standard protocol specifications, and the performance is improved.
  • the accompanying drawings are used to provide the present invention.
  • FIG. 3 is a schematic diagram of the state of the port state information in the LACPDU of the prior art
  • FIG. 3 is a diagram of each state machine relationship and message flow in the LACP protocol according to the prior art
  • FIG. 4 is configured on two devices according to the prior art.
  • Schematic diagram of link aggregation of two ports 5 is a schematic diagram of networking in accordance with an embodiment of the present invention
  • FIG. 6 is a flowchart of a method for implementing network interconnection by link aggregation according to an embodiment of the present invention
  • FIG. 7(d) are based on A detailed processing flowchart of a method for implementing network interconnection by link aggregation according to an embodiment of the present invention; wherein, FIG. 7(a) shows a main flow; FIG. 7(b) shows a fault detection flow; FIG. 7(c) ) shows the protocol message processing flow; Figure 7 (d) shows the port event processing flow.
  • FIG. 7(a) shows a main flow
  • FIG. 7(b) shows a fault detection flow
  • FIG. 7(c) shows the protocol message processing flow
  • Figure 7 (d) shows the port event processing flow.
  • DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The preferred embodiments of the present invention are described with reference to the accompanying drawings. According to an embodiment of the present invention, there is provided a method of implementing network interconnection using link aggregation, which is capable of implementing fast protection when a failure occurs in an application that implements protection using link aggregation.
  • the method is improved on the basis of the standard L ACP protocol, and can not only meet the support of the standard L ACP protocol switching device, but also overcome the defects of the standard LACP protocol, and further improve the aggregation performance.
  • the devices at both ends of the communication network may be one physical switching network board that supports link aggregation, or two physical switching network boards that use load sharing. . If two physical switching stencils are used in the load sharing mode, the two stencils need to use switching chips that support stacking technology (currently supported by high-end Ethernet switching chips). Cascading between chips is required. A switching network board. In either case, both ends can be logically treated as a switching node.
  • the monitored state change information is divided into different categories, and the related processing is performed according to the category to which the state change information belongs.
  • the status of the state change information may include: an immediate update class, an immediate synchronization class, and a delayed synchronization class.
  • the related processing includes: updating the local port state information, modifying the local aggregation information, and starting the protocol sending.
  • the state change information is the immediate synchronization class
  • the related processing includes: updating the local port state information, starting the protocol sending state machine, and waiting for the peer to reply and then modifying the aggregation information.
  • the processing includes: updating the status information of the local port, and periodically transmitting the link aggregation control protocol, that is, LACP.
  • the status machine is triggered to synchronize with the peer information.
  • the transmission of the packet is generally triggered by the following conditions:
  • the LACP periodically transmits the state machine, and the real-time transmission triggered in the above processing.
  • the actual fault is usually that some ports in the aggregation group cannot communicate with each other.
  • the other port can exchange packets. For example, if port 2 of switch A and B is abnormal and port 1 works normally, LACP is used.
  • the protocol specifies that the port AC-based aggregation information L ACPDU cannot be used for real-time fault notification.
  • the carried flag information may include at least one of the following: aggregation state information of other ports in the same aggregation group of the switching node, and an active sending/responding flag (if the sending flag is true, the segment needs to be acknowledged after receiving the LACPDU), the node Whether the aggregation port status changes flag, so that the peer end can be notified to update the aggregation group information.
  • the packets of the aggregation link of the communication network are classified into service packets and LACP packets.
  • the protocol packet format is in accordance with the LACP protocol frame format, and the link state information of the two ends of the network is directly sent to the processor of the board management switch chip.
  • the service packets transmitted through the network link are exchanged by the switch chip according to the aggregation port distribution rule.
  • the standard LACP process port aggregation information update specifies that the wait timeout timer prevents possible false updates. Therefore, the present invention improves the fastness and correctness of the port update time without waiting for the timeout timer by improving the LACPDU packet transmission/response mechanism.
  • LACP specifies a dynamic aggregation mode, that is, when both parties to the exchange run the protocol, the connection communication can be successfully established.
  • the specified aggregation group may no longer be used for device interconnection, but only one physical medium.
  • the port is connected to a device that does not support aggregation, such as a PC.
  • the configuration of the aggregation group is changed without changing the configuration. That is, if all the ports in the same aggregation group cannot exchange protocol packets during the port timeout period, the protocol state machine is not operated. Configure the aggregation port based on the information of the local physical port. If one or more physical ports can exchange protocol packets, the protocol state machine is restarted to maintain the link information of the aggregation link.
  • a method for implementing network interconnection by using link aggregation may include the following processes: allocating a group ID and a port ID; running an LACP state machine; starting a timer TI to monitor local port information, enabling port interruption ; Start the LACP packet protocol processing process, as shown in Figure 7. 07 003422
  • the method for implementing network interconnection by using link aggregation may specifically perform the following operations (step S602 - step S610): Step S602, assigning an aggregation group ID according to the switching node.
  • Step S604 the aggregation party runs the LACP protocol state machine to maintain the aggregation link information; Step S606, start the timer T1 (for example, set to 200ms), trigger the monitoring aggregation state, and Receive port status information on the interrupt (such as port up/down interrupt, interrupt information is also classified as described above, such as port up interrupt is classified as immediate interaction class, update local port status information table, start immediate interaction protocol message
  • the process of the port-down interrupt is classified into an immediate synchronization class, which not only needs to update the local port state information table, initiates the immediate interaction protocol packet process, but also needs to update the local aggregation group state information;);
  • Step S608 the monitored state change information According to the priority, it is divided into different categories, and the relevant processing is performed according to the category to which the state change information belongs; Body, update class immediately: Update the local port status information table in time, modify the aggregation information table, sub-package according to LACP
  • Private new flag information is sent to the peer synchronization, synchronize the class immediately, and update the local port status in time.
  • the information table is distributed according to the LACPDU format and filled in the Reserve field (reserved field) to be sent to the peer synchronization, and the local wait for the peer to reply and then correct the aggregation information; delay synchronization class: only update the local port status information table, and
  • the information synchronization of the peer end is triggered by the LACP periodic transmission state machine; the specific process in this step can be understood by referring to FIG. 7 (b); (W1 timer to continuation port state change interrupt 4 ,, return to step S606 Step S610: After receiving the LACPDU, the local end first determines whether the local end is in the static aggregation mode.
  • the local end converts to the dynamic aggregation mode, and determines whether the private identifier exists in the LACPDU. If the judgment result is no, Directly determine whether there is a private flag in the LACPDU; according to whether it exists in the LACPDU With the mark determination result of the subsequent operation. When it is determined that there is no private label in the LACPDU, it indicates that the peer end sends the standard LACP protocol packet, and performs subsequent operations according to the standard LACP protocol rule.
  • the LACP port timeout timer is cleared. When the LACP port timeout timer expires, it is determined whether the other ports in the same aggregation group are also in the timeout state. If yes, the state machine is disabled and the static aggregation mode is disabled. The port status is determined to be added to the trunk (aggregation group). If not, the dynamic aggregation mode is maintained. The local port is elected from the aggregation group. For details, see Figure 7 (d). For example, if all ports in the same aggregation group cannot communicate with the peer in the LACP timeout timer for 30 seconds, the protocol state machine is stopped, and the aggregation information is maintained in static aggregation mode.
  • the solution provided by the present invention basically follows the existing standards and has high versatility. It can be connected to the universal switch running LACP in a trunk manner, which is easy to implement, and has effective backup of network communication links and network device failure recovery. The ability to be fast and so on.
  • the above is only the preferred embodiment of the present invention, and is not intended to limit the present invention, and various modifications and changes can be made to the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the scope of the present invention are intended to be included within the scope of the present invention.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Description

利用链路聚合实现网络互连的方法 技术领域 本发明涉及通讯设备领域,并且具体地, 涉及利用链路聚合实现网络互 连的方法。 背景技术 通讯系统要求能够为用户提供可靠的、 不间断的服务。 对此,一种常用 的解决方法是冗余备份和主备倒换技术。 冗余备份和主备倒换技术有多种备 份模式, 例如 1+1、 1 :1、 1 : N等。 具体的说, 1+1 分是一种筒单的保护模式, 系统中的主 /备用单元组成 一个逻辑上的业务功能单元, 即, 被保护设备单元拥有一个备用单元, 主用 模块负责业务的实时处理, 备用业务保持与主用单元的数据一致, 在检测到 原主用单元故障时发起倒换, 由备用单元接管原主用单元的业务。 与 1+1备 份不同, 1 : 1备份中备用单元只在出现故障后开始工作, 而 1 : N备份就是 多个设备单元公用一个备用单元, 当一个主用单元故障时, 备用单元接替其 工作。 网络故障时为了保证设备的自愈能力,倒换必须在指定的时间内完成, 使之尽快恢复业务, 把对通讯系统的影响降到最低。 另外, 提高网络可靠性的方法还有端口捆绑 ( PORT TRUNKING ), 同 时这也是一种低成本的带宽扩展方案。 端口捆绑, 也称链路聚合 ( Link Aggregation ) , 是把多个以太网端口捆绑起来, 形成一条逻辑上的数据链路 , 从而提高带宽并互为备份, 提高了系统的可靠性。 该方法被广泛地应用于网 络交换机、 路由器、 宽带接入服务器等网络设备的互连。 链路聚合技术亦称为主干技术(Trunking )或捆绑技术(Bonding ), 其 实质是将两台设备间的数条物理链路 "组合" 成一条逻辑链路。 目前链路聚 合技术的标准为由 IEEE802委员会制定的 IEEE standard 802 . 3ad。链路聚合 控制协议 ( LACP, Link Aggregation Control Protocol )是 IEEE 802 . 3 ad标准 的主要内容之一, 定义了一种标准的聚合控制方式。 聚合的双方设备通过协 议交互聚合信息, 根据双方的参数和状态, 自动将匹配的链路聚合在一起收 发数据。 聚合形成后, 交换设备维护聚合链路状态, 当双方配置变化时, 自 动调整或解散聚合链路。 协议依据的是传送的信息和状态, 而不是传送的命令。 LACPDU (或 LACPDUs ) 由第一方, 即发起方 (Actor ) 把其知道的信息 (包括自身的状 态和伙伴的状态) 传送往第二方, 即发起方的伙伴 (Partner )。 LACPDU 传 送的信息足以让伙伴方决定下一步的^ "为。
LACP中是否为主动( Active )或被动( Passive )参与由 LACP— Activity 属性决定, 该管理属性和每个端口相对应。 取值为 Passive LACP时表明端口 不发送 LACPDU, 除非伙伴方取值 Active LACP。 取值为 Active LACP时表 明端口会参与该协议而不管伙伴方的取值。 只要任何一方的 LACP— Activity取值为 Active LACP , 都会导致周期性 的发送 LACPDU。 这些周期性的发送是以慢或快的速率进行, 主要取决于伙 伴方的 LACP— Timeout取值。 除了周期性的发送 LACPDU, 当需要传送时 (Need To Transmit, abbr. NTT ), 协议发送 LACPDU。 比如, 发起方的状态变化时, 或者从伙伴方的 LACPDU可以明显看出其不知道发起方目前的状态时。 协议假设 LACPDU的丢失率非常低, LACP没有采用任何明确的帧丢 失检测 /重发机制。但是如果从伙伴方传来的信息表明其没有发起方的即时信 息, 或者周期性发送时间到, 则发起方将会发送一个 LACPDU 来帮助伙伴 更新信息。 LACPDU 的结构如图 1 所示, 其中 Actor— State (本端状态) 和
Partner— State (被动端状态) '位结构如图 2所示。 协议的运行由几个状态机控制, 每个状态机执行不同的功能。 这些状态 机在是基于端口运行的, 像定时器超时或者收到 LACPDU 的事件会导致状 态的迁移以及动作的发生, 这些动作包括发送包含重复或者更新信息的 LACPDU。 周期性发送以及事件驱动的发送是由 NTT变量的状态控制的。 状态机包括以下几种类型:
Receive machine (接收状态机) 该状态机从伙伴方接收 LACPDU , 记录下所包含的信息, 然后根据 LACP_Timeout 设定的短定时器或长定时器进行计时。 状态机根据从伙伴方 接收到的信息, 判断发起方和伙伴方交换的协议信息的一致性是否已经达到 可以将端口和其他端口聚合或是作为单个端口的程度。 如果不能, 状态机将
NTT设置为真, 从而向伙伴方发送新协议信息。 如果等待伙伴方的协议信息 超时, 则接收状态机将使用默认参数填入。
Periodic Transmission machine (周期发送状态机 ) 该状态机决定是否周期性地在发起方和接收方之间交换用于维持聚合 的 LACPDU。 发送周期性 LACPDU 的前提是至少有一方被设置成 Active LACP。 Selection Logic (选择逻辑状态机) 该状态机负责为端口选择相应的聚合体。 Mux machine ( Mu 状态机 ) 该状态机负责将某个端口绑定于选定的聚合体,或者把端口从聚合体拆 离。 同时还负责根据当前的协议信息需要, 把该端口的接收或发送功能使能 或关闭。
Transmit machine (发送状态机) 该状态机负责根据其他状态或周期性状态机的需要来发送 LACPDU。 图 3说明了各个状态机之间的关系以及它们之间的消息流向。 其中, 箭 头标注 "对端状态信息( Partner State Information )" 代表更新的伙伴方信息, 该信息可能来自接收到的 LACPDU或自身的默人设定值, 该信息由 Receive machine发往其他状态机。箭头标注 "本端状态信息( Actor State Information )" 代表更新的发起方消息, 导致该 LACPDU 的发送的原因有可能是 Periodic Transmission machine周期到,或者是发起方状态发生变化,需要通知伙伴方。 现在参照图 4对链路聚合的保护功能进行说明。 如图 4所示, 当端口 1 和端口 2聚合时, 现有技术利用 LACP协议实现保护的过程如下:
1、 正常运行: 两个端口已经聚合, 运行 LACP协议, 周期性地互相发 送 LACP协议报文, 发送周期一般为 1秒。 端口 1和端口 2为负荷分担的一 003422 组工作端口, 具体保留的物理出端口 (端口 1 或端口 2 ) 由硬件按照 MAC 层选定的 hash (哈希) 算法决定。
2、 故障发生: 设备 A、 B的端口 2都发生故障, 如连接的物理线路双 向断纤。 3、 设备 A、 B的端口 2无法再接收对端的报文, 当三个发送周期无法 接收到对端报文时, 设备 A、 B认为此端口出现故障。
4、 设备 A、 B重新设置端口的聚合状态, 端口 2不再为工作端口, 不 承载业务, 端口 1为工作端口, 至此, 所有两端设备发送的报文都会皮发送 到端口 1上, 保护完成。 还有一种情况是单个方向上发生故障, 例如单向断 纤, 例如设备 A的端口 2到设备 B的端口 2的光纤断纤。 在这种情况下, 设 备 A的端口 2还可以接收到设备 B的 LACP协议报文。 设备 B在三个发送 周期没有收到报文后, 则认为此端口故障, 停止向设备 A发送 LACP协议报 文。再过三个发送周期,设备 A发现此端口故障。 利用链路聚合进行保护时, 两端设备对端口 4 态的更改必须一致。 如果两端设备对工作端口的设置不一 致, 比如 A 殳置端口 1为工作端口, Β设置端口 2为工作端口, 则业务 4艮文 的发送不能正常进行, 业务报文就会被丢弃, 保护失败。 因此, 保护的方式 必须由两端设备共同完成动作才能切换;如果任何一端设备未完成保护动作, 则保护不能成功。 所以, 在上述情况下, 只有 Α发现端口 2发生故障后才能 完成保护。 对上面的过程进行分析: 采用 IEEE 802 . 3 ad标准定义的 LACP协议进 行链路聚合保护时, 从第 2步发生故障到第 4步保护完成所需要的时间为网 络的故障恢复时间; 理论上, 该时间最少为 3秒, 单向断纤的保护时间为 6 秒。 然而, 随着各种网络应用对网络可用性要求的提高, 要求网络的故障恢 复时间越来越短。 特别是语音业务的传送, 恢复时间大于 50nis 就可能导致 应用方面的问题, 因此仅采用标准定义的 LACP协议进 4亍保护不能满足发展 的业务应用需求。 综上, LACP协议的缺点在于: 采用超时机制检测故障, 非常耗时; 没 有故障信息通告机制, 当本端发生错误时无法通知对端。 2 发明内容 考虑到现有技术中存在的上述问题而提出本发明, 为此, 本发明旨在提 供一种利用链路聚合实现网络互连的方法。 在该方法中, 在通讯网络的两端监测本地聚合组端口的状态时, 将监测 到的状态变化信息分为不同类别, 并根据状态变化信息所属的类别进行相关 处理。 具体地, 状态变化信息的类别包括: 立即更新类、 立即同步类、 延迟同 步类。 当状态变化信息为立即更新类时, 相关处理包括: 更新本地端口状态 信息, 修改本地聚合信息, 启动协议发送状态机。 当状态变化信息为立即同 步类时, 相关处理包括: 更新本地端口状态信息, 启动协议发送状态.机, 等 待对端应答后再修改聚合信息。 当状态变化信息为延迟同步类时, 相关处理 包括: 更新本地端口状态信息, 通过链路聚合控制协议即 LACP周期性发送 状态机触发和对端的 ^言息同步。 另外, 在该方法中, 在启动链路聚合控制协议发送状态机时, 在链路聚 合控制协议数据单元即 LACPDU中的保留 ( Reserve ) 字段携带以下私有标 志中的至少之一: 本交换节点同一聚合组中其他端口的聚合状态信息标志、 主动发送 /应答标志、 本节点聚合端口状态是否变化标志。 该方法具体可以包括如下操作: 按交换节点分配聚合组 ID以及每个聚 合组对应的参与聚合的端口 ID; 运行 L ACP协议状态机维护聚合链路信息; 启动定时器触发监测聚合状态, 并接收中断上报的端口状态信息; 对监测到 的状态变化信息分为不同类别, 并根据状态变化信息所属的类别进行相关处 理; 本端接收到 LACPDU后, 首先判断本端是否处于静态聚合方式, 如果 判断结果为是, 则转换为动态聚合方式, 并判断 LACPDU 中是否存在私有 标志, 如果判断结果为否, 则直接判断 LACPDU 中是否存在私有标志; 根 据对 LACPDU中是否存在私有标志的判断结果进行后续操作。 其中, 当判断 LACPDU中不存在私有标志时, 表示对端发送的是标准 LACP协议报文, 按标准 LACP协议规则进行后续操作。 当判断 LACPDU中存在私有标志时, 表示对端发送的是内部协议包, 并提取私有标志, 根据私有标志判断是否需要更新本地端口状态信息、 是否 需要应答处理结果给对端。 当私有标志标识对端状态有变化时, 提取报文中 的接收端口信息和组内其他端口信息,与本地维护的端口状态信息进行比较, 更新本地端口状态信息, 并根据更新后的端口状态信息修改聚合信息。 当私 有标志标识需要应答处理结果时, 应答处理结果给对端。 在上述方法中, 当定时器到时或有端口状态变化中断上^ :艮时, 重新启动 定时器触发监测聚合状态, 并接收中断上报的端口状态信息。 该方法可以进一步包括以下处理: 将端口超时定时器清零; 当端口超时 定时器到达时, 判断同一聚合组内其他端口是否也处于超时状态, 如果是, 则转为 "态聚合方式, 居本端物理端口状态决定是否加入聚合组( Trunk ), 如果否, 则将本端口从聚合组中选出。 另外, 在上述方法中, 交换节点为支持链路聚合的物理交换网板或采用 负荷分担工作的两个物理交换网板, 采用负荷分担工作的两个物理交换网板 采用支持堆叠技术的交换芯片, 且芯片之间需要级联。 本发明提供的方案基本上遵循现有标准、 通用行高, 可以和运行 LACP 的通用交换机之间以 Trunk方式对接, 易于实现。 其实现了以下有益效果: ( 1 ) 网络通讯链路有效备份, 上连口可以分配更多的带宽; (2 ) 网络设备 故障恢复能力快, 采用 LINK中断响应、 即时消息交互等手段, 可以快速 测链路的故障情况; (3 ) 区分故障等级, 及时恢复链路通讯能力, 同时避免 频繁处理协议报文, 最大限度提升了产品性能; (4 ) 对连接方法没有限制, 网络两端只要逻辑上是一个交换网板, 物理上可以是两个堆叠在一起的交换 单板或者通用交换机; ( 5 ) 聚合链路管理方法兼容标准协议规范, 性能得到 提高。 附图说明 附图用来提供对本发明的进一步理解, 并且构成说明书的一部分, 与本 发明的实施例一起用于解释本发明, 并不构成对本发明的限制。 在附图中: 图 1是根据现有技术的 LACPDU结构示意图; 图 2是 居现有技术的 LACPDU中端口状态信息位结构示意图; 图 3是根据现有技术的 LACP协议中各状态机关系及消息流向图; 图 4是根据现有技术的在两个设备上配置的两个端口链路聚合示意图; 图 5是 居本发明实施例的组网示意图; 图 6 是根据本发明实施例的利用链路聚合实现网络互连的方法的流程 图; 以及 图 7 ( a ) 至图 7 ( d ) 是根据本发明实施例的利用链路聚合实现网络互 连的方法的详细处理流程图; 其中, 图 7 ( a )示出了主流程; 图 7 ( b )示出 了故障检测流程; 图 7 ( c )示出了协议报文处理流程; 图 7 ( d )示出了端口 事件处理流程。 具体实施方式 以下结合附图对本发明的优选实施例进行说明, 应当理解, 此处所描述 的优选实施例仅用于说明和解释本发明, 并不用于限定本发明。 根据本发明实施例, 提供了一种利用链路聚合实现网络互连的方法, 其 能够在利用链路聚合实施保护的应用中发生故障时, 实施快速保护。 该方法 在标准 L ACP协议基.础上改进, 既能满足和支持标准 L ACP协议的交换设备 对接, 又能克服标准 LACP协议的缺陷, 进一步提高聚合性能。 首先, 在根据本发明实施例的方法中, 如图 5所示, 通讯网络两端的设 备可以是支持链路聚合的一个物理交换网板, 也可以是采用负荷分担工作的 两个物理交换网板。 如果是采用负荷分担方式的两个物理交换网板, 则两个 网板需要采用支持堆叠技术 (目前高端的以太网交换芯片都支持) 的交换芯 片, 芯片之间需要级联, 逻辑上看成一个交换网板。 无论那种情况, 两端逻 辑上都可看作是一个交换节点。 基于此,在该方法中,在通讯网络的两端监测本地聚合组端口的状态时, 将监测到的状态变化信息分为不同类别, 并根据状态变化信息所属的类别进 ^"相关处理。 具体地, 状态变化信息的类别可以包括: 立即更新类、 立即同步类、 延 迟同步类。 当状态变化信息为立即更新类时, 相关处理包括: 更新本地端口 状态信息, 修改本地聚合信息, 启动协议发送状态机。 当状态变化信息为立 即同步类时, 相关处理包括: 更新本地端口状态信息, 启动协议发送状态机, 等待对端应答后再修改聚合信息。 当状态变化信息为延迟同步类时, 相关处 理包括: 更新本地端口状态信息, 通过链路聚合控制协议即 LACP周期性发 送状态机触发和对端的信息同步。 这样, 在本发明中, 发送 ^佥测报文一般可 以由以下几种条件触发: LACP周期性发送状态机, 上述处理中触发的实时 发送。 实际中的故障通常是聚合组中一部分端口异常不能交互报文,另一部分 端口能正常交互报文, 如附图 1 中交换设备 A、 B的端口 2异常而端口 1工 作正常的情况, 而 LACP协议规定的是基于端口交互聚合信息 L ACPDU , 不 能实时进行故障通告。 因此, 在启动 LACP协议发送状态机时, 本发明利用 LACPDU中 Reserve字段携带标志 4言息。 携带的标志信息可以包括以下至少之一:本交换节点同一聚合组中其他 端口的聚合状态信息、 主动发送 /应答标志(如果发送标志为真, 则对段接收 到 LACPDU后需要应答)、 本节点聚合端口状态是否变化标志, 以便及时通 知对端更新聚合组信息。 经过通讯网络聚合链路的报文分为业务报文和 LACP协议报文。 协议报文格式符合 LACP协议帧格式, 携带网络两端的链 路状态信息, 直接送到单板管理交换芯片的处理器; 经过网络链路传递的业 务报文, 由交换芯片根据聚合端口分发规则交换至对端设备。 标准 LACP 流程端口聚合信息更新中规定等待超时定时器预防可能出 现的误更新, 这样, 本发明通过改进 LACPDU报文发送 /应答机制, 保证未 经过等待超时定时器更新端口的快速性和正确性。 另外, LACP规定的是动态聚合方式, 即, 交换双方都运行协议时才可 以成功建立连接通讯, 而在通讯设备应用中, 指定的聚合組可能不再用于设 备互联, 而只用其中一个物理端口同不支持聚合的设备如 PC机连接。 本发 明实施例中增加静态聚合方式支持不改动配置的情况下改变聚合组用途,即, 如果同一聚合组内所有端口在端口超时周期内都不能交互协议报文, 则不运 行协议状态机, 不必通过链路双方交互协议报文, 仅根据本地物理端口状态 信息配置聚合端口, 如果其中一个或多个物理端口能够进行协议报文交互, 则重新启动协议状态机维护聚合双方链路信息。 以下将进一步结合附图描述本发明。 一般地,根据本发明实施例的利用链路聚合实现网絡互连的方法可以包 括以下处理: 分配组 ID和端口 ID; 运行 LACP状态机; 启动定时器 TI监测 本地端口信息, 使能端口发生中断; 启动 LACP报文协议处理流程, 如图 7 07 003422
( a )所示。 具体地, 如图 6所示,据本发明实施例的利用链路聚合实现网络互连的 方法具体可以包 4舌如下操作 (步骤 S602-步骤 S610 ): 步骤 S602,按交换节点分配聚合组 ID以及每个聚合组对应的参与聚合 的端口 ID; 步骤 S604 , 聚合双方运行 LACP协议状态机维护聚合链路信息; 步骤 S606, 启动定时器 T1 (例如, 设置为 200ms ), 触发监测聚合状 态, 并接收中断上 4艮的端口状态信息 (如端口 up/down中断, 中断信息同样 如上文所述进行分类, 如端口 up 中断归为立即交互类, 更新本地端口状态 信息表, 启动立即交互协议报文流程, 对于端口 down中断归为立即同步类, 不仅需要更新本地端口状态信息表、 启动立即交互协议报文流程, 还需要更 新本地聚合組状态信息;); 步驟 S608 , 对监测到的状态变化信息按照优先级分为不同类别, 并冲艮 据状态变化信息所属的类别进行相关处理; 具体地, 立即更新类: 及时更新 本地端口状态信息表, 修改聚合信息表, 按 LACPDU格式分装并在 Reserve 字段填充.私有新标志信息发送给对端同步, 立即同步类, 及时更新本地端口 状态信息表, 按 LACPDU格式分装并在 Reserve字段(保留字段)填充私有 标志信息发送给对端同步, 本地等待对端应答后再修正聚合信息; 延迟同步 类: 只更新本地端口状态信息表, 和对端的信息同步由 LACP周期性发送状 态机触发; 该步骤中的具体流程可以参照图 7 ( b )来理解; ( T 1定时器到或 有端口状态变化中断上 4艮时, 返回至步骤 S606 ) 步骤 S610 , 本端接收到 LACPDU后, 首先判断本端是否处于静态聚合 方式, 如果判断结果为是, 则转换为动态聚合方式, 并判断 LACPDU 中是 否存在私有标志, 如果判断结果为否, 则直接判断 LACPDU 中是否存在私 有标志; 根据对 LACPDU中是否存在私有标志的判断结果进行后续操作。 其中, 当判断 LACPDU中不存在私有标志时, 表示对端发送的是标准 LACP协议报文, 按标准 LACP协议规则进行后续操作。 当判断 LACPDU中存在私有标志时, 表示对端发送的是内部协议包, 并提取私有标志, 根据私有标志判断是否需要更新本地端口状态信息、 是否 需要应答处理结果给对端。 当私有标志标识对端状态有变化时, 提取报文中 的接收端口信息和组内其他端口信息,与本地维护的端口状态信息进行比较, 更新本地端口状态信息, 并根据更新后的端口状态信息修改聚合信息。 当私 有标志(例如, 主动发送 /应答标志)标识需要应答处理结果时, 应答处理结 果给对端。 具体操作在图 7 ( c ) 中示出。 将 LACP端口超时定时器清零; 当 LACP端口超时定时器到达时, 判 断同一聚合组内其他端口是否也处于超时状态, 如果是, 则禁止运行状态机, 转为静态聚合方式, 根据本端物理端口状态决定是否加入 Trunk (聚合组), 如果否, 则保持动态聚合方式不变, 将本端口从聚合组中选出, 具体可以参 见图 7 ( d )。 例如, 如果本地同一聚合组内全部端口在 LACP协议超时定时 器 30s内都不能和对端交互协议报文, 则停止运行协议状态机, 采用静态聚 合方式维护聚合信息, 即如果本地端口 up就加入聚合组中, 端口 down时就 从聚合组中剔除。 如上所述, 本发明提供的方案基本上遵循现有标准、 通用性高, 可以和 运行 LACP的通用交换机之间以 Trunk方式对接 , 易于实现, 并具有网络通 讯链路有效备份、 网絡设备故障恢复能力快等优势。 以上所述仅为本发明的优选实施例而已, 并不用于限制本发明, 对于本 领 i或的技术人员来说, 本发明可以有各种更改和变化。 凡在本发明的 4青神和 原则之内, 所作的任何修改、 等同替换、 改进等, 均应包含在本发明的保护 范围之内。

Claims

权 利 要 求 书
1. 一种利用链路聚合实现网络互连的方法, 其特征在于, 在通讯网络的两 端监测本地聚合组端口的状态时, 将监测到的状态变化信息分为不同类 别, 并根据状态变化信息所属的类别进行相关处理。
2. 根据权利要求 1 所述的方法, 其特征在于, 所述状态变化信息的类别包 立即更新类、 立即同步类、 延迟同步类。
3. 根据权利要求 2所述的方法, 其特征在于, 当所述状态变化信息为立即 更新类时, 所述相关处理包 4舌: 更新本地端口状态ί言息, 改本地聚合 信息, 启动协议发送状态机。
4. 根据权利要求 2所述的方法, 其特征在于, 当所述状态变化信息为立即 同步类时, 所述相关处理包括: 更新本地端口状态信息, 启动协议发送 状态机, 等待对端应答后再 ^^改聚合信息。
5. 根据权利要求 2所述的方法, 其特征在于, 当所述状态变化信息为延迟 同步类时, 所述相关处理包^■: 更新本地端口状态信息, 通过链路聚合 控制协议即 LACP周期性发送状态机触发和对端的信息同步。
6. 根据权利要求 1至 5 中任一项所述的方法, 其特征在于, 在启动链路聚 合控制协议发送状态机时,在链路聚合控制协议数据单元即 LACPDU中 的保留字段携带以下私有标志中的至少之一: 本交换节点同一聚合组中 其他端口的聚合状态信息标志、 主动发送 /应答标志、 本节点聚合端口状 态是否变化标志。
7. 根据权利要求 6所述的方法, 其特征在于, 具体包括如下操作:
按交换节点分配聚合组 ID 以及每个聚合组对应的参与聚合的端口
ID;
运行 LACP协议状态机维护聚合链路信息;
启动定时器触发监测聚合状态, 并接收中断上报的端口状态信息; 对监测到的状态变化信息分为不同类别,并 居状态变化信息所属 的类别进行所述相关处理; 本端接收到 LACPDU后, 首先判断本端是否处于静态聚合方式, 如果判断结果为是, 则转换为动态聚合方式, 并判断所述 LACPDU中是 否存在所述私有标志, 如果判断结果为否, 则直接判断所述 LACPDU中 是否存在所述私有标志;
才艮据对所述 LACPDU中是否 ^"在所述私有标志的判断结果进 4亍后 续操作。
8. 根据权利要求 7所述的方法,其特征在于, 当判断所述 LACPDU中不存 在所述私有标志时, 表示对端发送的是标准 LACP 协议报文, 按标准 LACP协议规则进^ "后续操作。
9. 根据权利要求 7所述的方法, 其特征在于, 当判断所述 LACPDU中存在 所述私有标志时, 表示对端发送的是内部协议包, 并提取所述私有标志, 根据所述私有标志判断是否需要更新本地端口状态信息、 是否需要应答 处理结果给对端。
10. 居权利要求 9所述的方法, 其特征在于, 当所述私有标志标识对端状 态有变化时, 提取报文中的接收端口信息和组内其他端口信息, 与本地 维护的端口状态信息进行比较, 更新本地端口状态信息, 并根据更新后 的端口状态信息修改聚合信息。
11. 根据权利要求 10所述的方法, 其特征在于, 当所述私有标志标识需要应 答处理结果时, 应答处理结果给对端。
12. 根据权利要求 7所述的方法, 其特征在于, 当定时器到时或有端口状态 变化中断上报时, 重新启动定时器触发监测聚合状态, 并接收中断上报 的端口状态信息。
13. 根据权利要求 7至 12中任一项所述的方法, 其特征在于, 进一步包括: 将端口超时定时器清零;
当端口超时定时器到达时,判断同一聚合组内其他端口是否也处于 超时状态, 如果是, 则转为静态聚合方式, 根据本端物理端口状态决定 是否加入聚合组, 如果否, 则将本端口从聚合組中选出。 根据权利要求 7至 12中任一项所述的方法, 其特征在于, 所述交换节点 板, 所述采用负荷分担工作的两个物理交换网板采用支持堆叠技术的交 换芯片, 且芯片之间需要级联。
PCT/CN2007/003422 2007-08-20 2007-12-03 Procédé de mise en œuvre d'une interconnexion de réseau par l'intermédiaire d'une agrégation de liaisons WO2009023996A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN200710142048.6 2007-08-20
CN2007101420486A CN101094157B (zh) 2007-08-20 2007-08-20 利用链路聚合实现网络互连的方法

Publications (1)

Publication Number Publication Date
WO2009023996A1 true WO2009023996A1 (fr) 2009-02-26

Family

ID=38992196

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2007/003422 WO2009023996A1 (fr) 2007-08-20 2007-12-03 Procédé de mise en œuvre d'une interconnexion de réseau par l'intermédiaire d'une agrégation de liaisons

Country Status (2)

Country Link
CN (1) CN101094157B (zh)
WO (1) WO2009023996A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105187323A (zh) * 2015-09-29 2015-12-23 盛科网络(苏州)有限公司 实现多级流量负载分担的装置及方法
CN111049765A (zh) * 2019-12-12 2020-04-21 北京东土军悦科技有限公司 聚合端口切换方法、装置、芯片、交换机及存储介质
CN115174708A (zh) * 2022-07-14 2022-10-11 北京物芯科技有限责任公司 报文接收、信息确定方法、装置、交换芯片、设备及系统

Families Citing this family (36)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101674230B (zh) * 2008-09-12 2012-08-08 华为技术有限公司 一种链路聚合组信息管理的方法和网络节点
WO2010028570A1 (zh) * 2008-09-12 2010-03-18 华为技术有限公司 一种链路聚合组信息管理的方法和网络节点
US8477791B2 (en) * 2009-05-13 2013-07-02 Avaya Inc. Method and apparatus for locally implementing port selection via synchronized port state databases maintained by the forwarding plane of a network element
US8619605B2 (en) * 2009-05-13 2013-12-31 Avaya Inc. Method and apparatus for maintaining port state tables in a forwarding plane of a network element
CN101945042B (zh) * 2010-09-03 2014-01-01 中兴通讯股份有限公司 一种链路聚合的方法及设备
CN101938780A (zh) * 2010-09-19 2011-01-05 中兴通讯股份有限公司 一种降低多链路点到点协议组丢包率的方法及系统
CN101977145B (zh) * 2010-10-20 2015-04-01 中兴通讯股份有限公司 一种聚合链路业务流转发的方法及系统
CN101984573B (zh) * 2010-11-15 2015-07-22 中兴通讯股份有限公司 分布式实现lacp标准状态机的方法及系统
CN102118280A (zh) * 2011-03-02 2011-07-06 浪潮(北京)电子信息产业有限公司 一种实现服务器内网络服务启动的方法及装置
CN102752126B (zh) * 2011-04-22 2017-05-24 中兴通讯股份有限公司 一种自动保护倒换方法、主设备及设备
CN102404172A (zh) * 2011-12-21 2012-04-04 迈普通信技术股份有限公司 一种多激活检测方法和设备
JP5826381B2 (ja) * 2012-04-27 2015-12-02 アライドテレシスホールディングス株式会社 スイッチ、送信方法、プログラム、記録媒体
CN102724030A (zh) * 2012-06-29 2012-10-10 杭州迪普科技有限公司 一种高可靠性的堆叠系统
CN103905326B (zh) * 2012-12-28 2017-11-24 迈普通信技术股份有限公司 以太网链路聚合的报文转发控制方法及网络设备
US9497132B2 (en) 2013-04-23 2016-11-15 Telefonaktiebolaget Lm Ericsson (Publ) Method and system of implementing conversation-sensitive collection for a link aggregation group
US9660861B2 (en) 2013-04-23 2017-05-23 Telefonaktiebolaget L M Ericsson (Publ) Method and system for synchronizing with neighbor in a distributed resilient network interconnect (DRNI) link aggregation group
US9553798B2 (en) * 2013-04-23 2017-01-24 Telefonaktiebolaget L M Ericsson (Publ) Method and system of updating conversation allocation in link aggregation
CN104219075B (zh) * 2013-05-31 2017-10-24 新华三技术有限公司 一种基于开放流协议的lacp环路检测方法和装置
CN103401707B (zh) * 2013-07-26 2016-09-28 杭州华三通信技术有限公司 链路聚合方法及接入设备
CN104683264B (zh) * 2013-11-26 2018-05-25 上海斐讯数据通信技术有限公司 交换机接口的速率模式切换方法
CN103780472B (zh) * 2014-01-06 2017-08-22 新华三技术有限公司 一种聚合端口组的选中口管理方法及设备
JP6189783B2 (ja) * 2014-04-08 2017-08-30 APRESIA Systems株式会社 中継システムおよびスイッチ装置
CN104468203A (zh) * 2014-11-26 2015-03-25 上海斐讯数据通信技术有限公司 交换机聚合组中端口故障切换的配置方法
CN105703933B (zh) * 2014-11-27 2019-03-15 杭州迪普科技股份有限公司 聚合链路恢复方法及装置
CN105721234B (zh) * 2014-12-05 2019-04-05 杭州迪普科技股份有限公司 端口聚合方法及装置
CN107154861B (zh) * 2016-03-02 2019-11-29 华为技术有限公司 一种堆叠系统的数据传输方法及装置
CN107517114B (zh) * 2016-06-16 2022-08-02 中兴通讯股份有限公司 一种通信端口的链接状态更新方法及装置
CN106210871B (zh) * 2016-07-14 2019-04-30 四川长虹电器股份有限公司 用于智能电视的网络切换系统及方法
CN107547452B (zh) * 2017-07-11 2020-09-25 新华三技术有限公司 一种分布式通信设备及其报文转发方法和装置
CN110391979B (zh) * 2018-04-23 2021-01-15 华为技术有限公司 一种网络设备堆叠方法、网络设备及存储介质
CN109889442B (zh) * 2018-12-24 2021-07-27 广州芯德通信科技股份有限公司 一种以太网接口链路聚合方法、电子设备、存储介质
CN110971519A (zh) * 2019-12-12 2020-04-07 迈普通信技术股份有限公司 一种端口互联的管理方法及装置
CN112087378B (zh) * 2020-08-18 2022-03-22 浪潮思科网络科技有限公司 一种LACP forceup功能的实现方法、设备及介质
CN112134797B (zh) 2020-09-04 2022-05-13 苏州浪潮智能科技有限公司 一种改善链路聚合协议超时的方法和设备
CN112350853B (zh) * 2020-10-21 2022-09-30 杭州迪普科技股份有限公司 一种网络设备主备切换的方法、装置、存储介质及设备
CN114124666B (zh) * 2021-11-19 2024-02-23 深信服科技股份有限公司 网络处置方法、装置、计算机设备和存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030128706A1 (en) * 2001-06-14 2003-07-10 Mark Barry Ding Ken Extension of link aggregation protocols over the network
JP2005347943A (ja) * 2004-06-01 2005-12-15 Hitachi Ltd ネットワーク中継装置及びその制御方法
CN1913414A (zh) * 2006-08-29 2007-02-14 华为技术有限公司 利用链路聚合实施保护的方法、设备和系统
CN101018228A (zh) * 2006-12-22 2007-08-15 华为技术有限公司 一种端口聚合方法及装置

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN100477601C (zh) * 2005-04-12 2009-04-08 富士通株式会社 用于链路聚合标志协议的特定标志消息
CN1941683A (zh) * 2005-09-30 2007-04-04 华为技术有限公司 一种触发链路聚合组重分发业务的方法及系统
CN100512194C (zh) * 2006-12-25 2009-07-08 华为技术有限公司 链路聚合方法、装置、mac帧收发方法和系统

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030128706A1 (en) * 2001-06-14 2003-07-10 Mark Barry Ding Ken Extension of link aggregation protocols over the network
JP2005347943A (ja) * 2004-06-01 2005-12-15 Hitachi Ltd ネットワーク中継装置及びその制御方法
CN1913414A (zh) * 2006-08-29 2007-02-14 华为技术有限公司 利用链路聚合实施保护的方法、设备和系统
CN101018228A (zh) * 2006-12-22 2007-08-15 华为技术有限公司 一种端口聚合方法及装置

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105187323A (zh) * 2015-09-29 2015-12-23 盛科网络(苏州)有限公司 实现多级流量负载分担的装置及方法
CN111049765A (zh) * 2019-12-12 2020-04-21 北京东土军悦科技有限公司 聚合端口切换方法、装置、芯片、交换机及存储介质
CN115174708A (zh) * 2022-07-14 2022-10-11 北京物芯科技有限责任公司 报文接收、信息确定方法、装置、交换芯片、设备及系统

Also Published As

Publication number Publication date
CN101094157A (zh) 2007-12-26
CN101094157B (zh) 2011-09-21

Similar Documents

Publication Publication Date Title
WO2009023996A1 (fr) Procédé de mise en œuvre d'une interconnexion de réseau par l'intermédiaire d'une agrégation de liaisons
US9270524B2 (en) Method and device for LACP link switching and data transmission
WO2011100882A1 (zh) 链路检测方法、装置和系统
US8509059B2 (en) Method for operating a virtual router redundancy protocol router and communication system therefor
WO2011157151A2 (zh) 实现容灾备份的方法、设备及系统
WO2012000234A1 (zh) 链路间快速切换的方法、装置和系统
WO2009009977A1 (fr) Procédé de commutation maître/secours pour un dispositif de routage et système de secours pour un dispositif de routage
US20130238738A1 (en) Distributed method and system for implementing link aggregation control protocol (lacp) standard state machines
WO2009046655A1 (fr) Système, procédé et dispositif de commutation de protection automatique
EP1982447A2 (en) System and method for detecting and recovering from virtual switch link failures
TW200421777A (en) High availability Ethernet backplane architecture
CN102904818A (zh) 一种arp信息表项更新方法及装置
WO2007062573A1 (fr) Procede de gestion de groupe auxiliaire de protocole de redondance du routeur virtuel
WO2008014696A1 (fr) Méthode et dispositif pour effectuer un transfert de communications
WO2014059844A1 (zh) 分布式弹性网络互连中网关动态切换方法和装置
WO2008046358A1 (fr) Procédé et dispositif destinés à réaliser une pénétration d'un statut de liaison de réseau point à multipoint
CN107911291A (zh) Vrrp路由器切换方法、路由器、vrrp主备切换系统及存储介质
CN101465859A (zh) 一种触发主备用接口板倒换的方法及装置
WO2011147312A1 (zh) 一种业务接入路由器的端口备份方法、装置和系统
CN102075343A (zh) 一种实现带外管理的方法、系统和带外管理交换机
WO2012058895A1 (zh) 聚合链路切换方法及装置
CN101197733A (zh) 网络连通性的自动检测方法及装置
WO2012048585A1 (zh) 切换方法和路由器
WO2012062069A1 (zh) 双向转发检测报文的发送方法及设备
CN111343019A (zh) 一种重连方法、装置和存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 07845784

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 07845784

Country of ref document: EP

Kind code of ref document: A1