WO2010009639A1 - 堆叠交换机分离和复合的方法及系统 - Google Patents

堆叠交换机分离和复合的方法及系统 Download PDF

Info

Publication number
WO2010009639A1
WO2010009639A1 PCT/CN2009/071741 CN2009071741W WO2010009639A1 WO 2010009639 A1 WO2010009639 A1 WO 2010009639A1 CN 2009071741 W CN2009071741 W CN 2009071741W WO 2010009639 A1 WO2010009639 A1 WO 2010009639A1
Authority
WO
WIPO (PCT)
Prior art keywords
soft
stack
switch
topology
stacking
Prior art date
Application number
PCT/CN2009/071741
Other languages
English (en)
French (fr)
Inventor
孙向东
Original Assignee
华为技术有限公司
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 华为技术有限公司 filed Critical 华为技术有限公司
Publication of WO2010009639A1 publication Critical patent/WO2010009639A1/zh

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00Packet switching elements
    • H04L49/45Arrangements for providing or supporting expansion
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00Packet switching elements
    • H04L49/65Re-configuration of fast packet switches

Definitions

  • the present invention relates to stacked switches, and more particularly to a method and system for separating and combining stacked switches.
  • a stack switch is a type of box-type Ethernet switch. It is an administrative domain consisting of Ethernet switches connected through stack ports. It consists of one master switch and several slave switches. The master switch and the slave switch are stack switches. Stack members. The stacked Ethernet switches can be viewed as one device, and the user can manage all the switches in the stack switch through the master switch.
  • Ethernet switches When multiple Ethernet switches are connected through a stack port, users can configure them on one of them, set them as a stack switch, and set the currently configured Ethernet switch as the master switch in the stack.
  • the master switch and the slave switch are connected by stacking modules and special stacking lines.
  • the stack switch is maintained through the connection between the master switch and the slave switch. As soon as the connection is found disconnected, the slave switch automatically exits the stack.
  • stack switches are easy to protect user investment, gradually expand capacity, facilitate management, save IP addresses, and support cross-device service backup, which is widely used in the market (especially the enterprise network market).
  • Star and daisy chain are common stacking methods, and the center of the star stack is a single point of failure. That is to say, when the central switch fails, the entire transmission network will be paralyzed, which has huge security risks, so the chrysanthemum is generally used.
  • Chain connection forming a ring top stacking ring.
  • the switch includes two stacking ports 1, 2, and the two stacking ports of the four switches are connected by stacking lines. A stacking ring is formed.
  • the switch from the switch to the stack switch is the first to physically disconnect the corresponding stack port or close the corresponding stack port, and then remove/join the stack switch.
  • the embodiments of the present invention provide a method and system for separating and recombining a stack switch, which can seamlessly leave/join the stack switch from the switch.
  • An embodiment of the present invention provides a method for separating a stack switch, including:
  • the soft split command triggering the master switch to send the soft split message to the stack member, where the soft split message includes soft split point information;
  • the stack member calculates the stack topology according to the soft split point information in the soft split message according to the topology calculation rule, and modifies the forwarding table according to the calculation result of the stack topology.
  • An embodiment of the present invention further provides a method for stacking switch composites, including:
  • the soft compound command triggering the master switch to send the soft composite message to each stack member, where the soft composite message includes soft composite point information matching the soft separation point;
  • the stack members calculate the stack topology according to the topology calculation rule according to the soft composite point information, and modify the forwarding table according to the calculation result of the stack topology.
  • An embodiment of the present invention further provides a system in which a stack switch is separated, including a stack topology management module and a sending module, where the stack topology management module communicates with a sending module;
  • the sending module is configured on the main switch, and is configured to send a soft separated message, where the soft separated message includes soft split point information;
  • the stacking topology management module is configured on the slave switch, and is configured to calculate a stack topology according to the topology calculation rule according to the soft separation point information in the soft separation packet, and modify the forwarding table according to the calculation result of the stack topology. .
  • An embodiment of the present invention further provides a system for stacking switch composites, including a stack topology management module and a sending module, where the stacking topology management module communicates with the sending module;
  • the sending module is disposed on the main switch, and is configured to send a soft composite message, where the soft composite message includes soft composite point information that matches a soft separation point when the soft separation command is executed;
  • the stacking topology management module is configured on the slave switch, and is configured to calculate a stack topology according to the topology calculation rule according to the soft compound point information in the soft composite packet, and modify the forwarding table according to the calculation result of the stack topology. .
  • the method and system for separating and combining the stack switches trigger the master switch to send soft-separated or soft-composite packets to the stack member by executing the soft-separation command or the soft-composite command.
  • the information includes a soft separation point or a soft composite point.
  • the stack member can calculate the stack topology according to the topology calculation rule according to the information of the soft separation point or the soft composite point, and modify the forwarding table according to the calculation result, so When a switch is removed or reassembled to a stack switch, packets are not lost, and the switch can be seamlessly removed from the switch.
  • Figure 1 is a schematic diagram of an application of a daisy chain stack switch
  • FIG. 2 is a flowchart of a method for separating a stack switch according to an embodiment of the present invention
  • FIG. 3 is a flow chart of a method for stacking switches in an embodiment of the present invention.
  • FIG. 4 is a flow chart of an application embodiment of a method for separating and recombining stack switches provided by the present invention
  • FIG. 5 is an embodiment of a method for separating a stack switch provided by the present invention before being removed from the switch II Stacking diagram
  • FIG. 6 is a schematic diagram of stacking of an embodiment of a method for separating a stack switch according to the present invention, which is softly separated from a switch II;
  • FIG. 7 is a schematic diagram of stacking of an embodiment of a method for stacking switch stacks removed from switch II according to the present invention.
  • FIG. 8 is a schematic diagram of a stacking method of a method for stacking switches according to the present invention, which is removed from the switch II and soft-composited;
  • FIG. 9 is a schematic diagram of a system for separating and recombining stack switches according to an embodiment of the present invention. detailed description
  • the embodiment of the invention provides a method and a system for separating and recombining a stacking switch, so that when the daisy-chained ring topology of the switch is removed/joined from the switch, no packet loss occurs, and the switch can be seamlessly left/ Join the stack switch.
  • an embodiment of the present invention provides a method for separating a stack switch, which may include the following steps:
  • the soft split point information includes a number of the switch to be separated by the soft split command and a stack port number.
  • a stack port of a switch to be detached can be referred to as a soft split point.
  • the stacking member calculates the stack topology according to the topology calculation rule according to the soft separation point information, and modifies the forwarding table according to the calculation result of the stack topology.
  • An embodiment of the present invention further provides a method for stacking a switch. As shown in FIG. 3, the method may include:
  • the soft composite message is sent to each stack member, and the soft composite message includes soft composite point information that matches the soft separation point.
  • the soft separation point is determined by the number of the switch to be separated and the stack port number in the soft separation message in the embodiment shown in FIG. 2, where the soft composite point information includes the number and stack of the switch to be composited by the soft composite command.
  • the port number For example, the embodiment shown in FIG. 2 uses a certain port of a switch as a soft split point, and then, in soft recombination, a port of the switch is used as a soft recombination point, that is, the other switch is notified.
  • the aforementioned soft separation point is already available and does not need to be disconnected.
  • the stack member is configured to calculate a stack topology according to the topology calculation rule according to the soft composite point information, and modify the forwarding table according to the calculated result.
  • the method for separating and recombining a stack switch by executing a soft split command or a soft compound command, triggers the master switch to send a soft split or soft composite message to the stack member, where the soft split point or the soft composite point is included.
  • the stack member can calculate the stack topology according to the topology calculation rule according to the soft separation point or the soft composite point information, and modify the forwarding table according to the calculation result of the stack topology, so when the slave switch is removed or composited to the stack switch, Packets are lost, so that the switch can be seamlessly removed/joined from the switch.
  • FIG. 4 is a flow chart of an application embodiment of a method for separating and recombining stack switches provided by the present invention.
  • the method can include the following steps:
  • S401 Performing a soft-detach command on the master switch, the command triggers the master switch to send the soft-detach packet to all the stack members except itself; where the stack member refers to each of the stack switches.
  • the switch including the master switch and all slave switches; soft-detach can be designed as switch stack-port ⁇ slot/port> soft-detach, where slot refers to the number of the switch, and port refers to the stack port number.
  • the slot and the port jointly determine which stack port of the switch should be disconnected during soft separation. For example, if the slot/port is 1/1, it means that the stack port 1 of the I switch is a soft split point and will be disconnected.
  • the soft-detach command also requires the stack member to change the stack topology.
  • the soft-detached information carried by the soft-detach packet is the basis for the stack member to perform topology calculation.
  • the stack member After receiving the soft-detach packet in the S401, the stack member changes the stack topology and the forwarding table, so as to change the direction of the service data flow, and change the daisy-chain ring topology into a chain shape. Open soft separation point;
  • the no soft-detach command is executed on the master switch, and the command is sent to the master switch to send the soft-detach packet to all the stack members except the self.
  • the information of the soft-complex point; here, the no soft-detach packet can be designed as no switch stack-port ⁇ slot/port> soft-detach, which corresponds to the soft-detach, where slot refers to the number of the switch. Port refers to the stack port number.
  • the slot and port jointly determine the soft recombination point.
  • the soft recombination point should match the soft separation point in the soft-detach packet, indicating that the soft separation point in the soft-detach is already available.
  • the no soft-detach command is used to change the stack topology.
  • the soft-complex information carried by the no-soft-detach packet is the topology of the stack member. in accordance with;
  • the stack member After receiving the no soft-detach packet in the S404, the stack member changes the stack topology and the forwarding table to restore the daisy chain ring topology.
  • the main stack I and the two stack ports 1 and 2 of the slave switches II, III, and IV are sequentially connected by a stacking line to form a daisy chain ring topology.
  • stack members I and III transmit data streams over the link of member II.
  • the forwarding table on stack member III is as follows:
  • the member switch After the member switch receives the soft-separated packet, the member switch performs topology calculation based on the soft-separation point information (that is, 1/1 disconnected information) in the packet, and recalculates each stack member relative to the local The topology location and path distance of the switch are recalculated according to the shortest path priority principle.
  • the original daisy chain ring topology is modified to a chain topology.
  • the stack member After the topology calculation, the stack member changes its forwarding table. At this time, the forwarding table on the stack member I is changed to:
  • the service data forwarding path on the stack switch is as shown in Figure 6.
  • the data flow between the stack members I and III is changed from the stack member II to the stack member IV. There is no flow on the stacking line between I and II. Then physically The stacking line of the 1/1 port is disconnected, that is, the stacking ring is disconnected. Since the data flow between I and III is not transmitted on the link between I and II, the service is not affected, and the stacking ring is changed without loss. For the topology of the chain, it is now possible to smoothly move out of Switch II without affecting the services between other stack members.
  • the stack port 1 of the switch I and the stack port 2 of the switch III are connected by a stacking line, and a daisy-chain connection is formed between the stack members I, III, and IV.
  • the stack topology is shown in Figure 7. There are three switches I, III, and IV in the ring topology. The data flow between the stack members I and III is still transmitted through the stack member IV.
  • each stack member can know the stack of the I switch. Port 1 is already available, so recalculate the topology, change the stack topology and forwarding table, and restore the ring topology.
  • the member switch After the member switch receives the soft composite message, the member switch performs topology calculation based on the soft composite point information (that is, 1/1 composite information) in the packet, and recalculates each stack member relative to the switch.
  • the topology location and path distance are recalculated according to the shortest path priority principle to the outbound port of each member switch traffic.
  • the chain topology is re-modified into a daisy chain ring topology.
  • the stack member After the topology calculation, the stack member changes its forwarding table. At this time, the forwarding table on the stack member I is changed to:
  • the method for separating or recombining the stack switch according to the foregoing stack switch is to separate or combine the stack switches according to the embodiment of the present invention.
  • the data stream Before physically disconnecting the stack ring or closing the stack port, the data stream is changed into a chain shape.
  • Topology when compounding, change the chain topology to a ring topology, so when the switch is removed or composited to the stack switch, it will not cause short-term packet loss, thus enabling seamless exit from the switch.
  • the embodiment of the present invention further provides a system in which a stack switch is separated, including a stack topology management module 901 and a sending module 902, where the stack topology management module 901 communicates with a sending module 902.
  • the sending module 902 is configured on the main switch, and is configured to send a soft split message, where the soft split message includes soft split point information;
  • the stack topology management module 901 is disposed on the slave switch, and is configured to calculate a stack topology according to the topology calculation rule according to the soft separation point information, and modify the forwarding according to the calculation result of the stack topology.
  • the soft separation point information includes a number of the switch to be separated and a stack port number of the system separated by the stacking switch.
  • a stack port of a switch to be detached can be referred to as a soft split point.
  • the embodiment of the present invention further provides a system for stacking switches, comprising: a stacking topology management module and a sending module, wherein the stacking topology management module communicates with the sending module;
  • the sending module is disposed on the main switch, and configured to send a soft composite message, where the soft composite message includes soft composite point information that matches a soft separation point when the soft separation command is executed;
  • the stacking topology management module is configured on the slave switch, and is configured to calculate a stack topology according to the topology calculation rule according to the soft compound point information in the soft composite packet, and modify the forwarding table according to the calculation result of the stack topology. .
  • the soft separation point is determined by the number of the switch to be separated and the stack port number in the soft separation message when the soft separation command is executed.
  • the soft composite point information includes a number of the switch to be combined and a stack port number of the system in which the stack switch is composited.
  • the relationship between the sending module and the stack topology management module can also refer to FIG. 9.
  • the system in which the stack switch is separated and combined changes the data stream to a chain before physically disconnecting the stack ring or closing the stack port.
  • Topology when compounding, change the chain topology to a ring topology, so when the switch is removed or composited to the stack switch, it will not cause short-term packet loss, thus enabling seamless exit from the switch. Join the stack switch.

Landscapes

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

Description

本申请要求 2008年 7月 25 日提交的、 申请号为 200810142654.2、 发明 名称为一种堆叠交换机分离或复合的方法及设备的中国专利申请的优选权, 其全部内容通过引用结合在本申请中。 技术领域
本发明涉及堆叠交换机, 特别是涉及堆叠交换机分离和复合的方法及系 统。
背景技术
在大型网络中, 通常采用堆叠交换机以满足大型网络对端口数量的要求。 堆叠交换机是盒式以太网交换机的一种, 是由一些通过堆叠端口相连的以太 网交换机组成的一个管理域, 其中包括一个主交换机和若干个从交换机, 主 交换和从交换机都是堆叠交换机的堆叠成员。 堆叠在一起的以太网交换机可 看作为一个设备, 用户可通过主交换机实现对堆叠交换机内所有交换机的管 理。
当多个以太网交换机通过堆叠端口 (Stack Port )相连时, 用户可以在其 中一台上进行配置, 把它们设置成堆叠交换机, 并把当前进行配置的以太网 交换机设置为堆叠中的主交换机。 主交换机和从交换机之间通过堆叠模块及 特殊的堆叠线连接起来。 堆叠交换机是通过主交换机和从交换机之间的连接 来维持的, 只要发现连接断开, 从交换机就会自动退出堆叠。
由于堆叠交换机便于保护用户投资, 逐步扩容, 方便管理, 节省 IP地址, 同时可以支持跨设备的业务备份, 其在市场上(尤其是企业网市场)得到了 广泛应用。
星形和菊花链式是常用的堆叠方式, 而星形堆叠的中心是单点故障点, 也就是说中心交换机出现故障时, 整个传输网就将瘫痪, 有巨大的安全隐患, 所以一般采用菊花链式连接, 形成环形拓朴堆叠环。 如图 1 所示, 交换机均 包括有两堆叠端口 1、 2, 四台交换机的两堆叠端口之间用堆叠线依次连接之, 形成堆叠环。
对于菊花链式堆叠交换机, 目前从交换机离开 /加入堆叠交换机主要采用 先物理上断开相应堆叠端口或者先关闭相应堆叠端口, 然后从交换机再离开 / 加入堆叠交换机的方法。
在实现本发明过程中, 发明人发现现有技术中至少存在如下问题: 当从交换机离开 /加入堆叠交换机的堆叠环时,会造成短时间的报文丟失, 无法实现从交换机无缝离开 /加入堆叠交换机。
发明内容
本发明实施例提供了堆叠交换机分离和复合的方法及系统, 能够实现从 交换机无缝离开 /加入堆叠交换机。
本发明的实施例采用如下技术方案:
本发明实施例提供了一种堆叠交换机分离的方法, 包括:
执行软分离命令, 所述软分离命令触发主交换机将软分离报文发送给堆 叠成员, 所述软分离报文包括软分离点信息;
所述堆叠成员根据所述软分离报文中的软分离点信息, 按照拓朴计算规 则计算堆叠拓朴并根据堆叠拓朴的计算结果修改转发表。
本发明的实施例还提供了一种堆叠交换机复合的方法, 包括:
按照上述的方案实现了堆叠交换机的分离时, 在软分离点处移出或加入 从交换机;
执行软复合命令, 所述软复合命令触发主交换机将软复合报文发送给各 堆叠成员, 所述软复合报文包括与所述软分离点匹配的软复合点信息;
所述各堆叠成员根据所述软复合点信息, 按照拓朴计算规则计算堆叠拓 朴并根据堆叠拓朴的计算结果修改转发表。
本发明的实施例还提供了一种堆叠交换机分离的系统, 包括堆叠拓朴管 理模块和发送模块, 所述堆叠拓朴管理模块与发送模块通信;
所述发送模块设置于主交换机上, 用于发送软分离报文, 所述软分离报 文包括软分离点信息; 所述堆叠拓朴管理模块设置于从交换机上, 用于根据所述软分离报文中 的软分离点信息, 按照拓朴计算规则计算堆叠拓朴, 并根据堆叠拓朴的计算 结果修改转发表。
本发明的实施例还提供了一种堆叠交换机复合的系统,包括堆叠拓朴管理 模块和发送模块, 所述堆叠拓 4卜管理模块与发送模块通信;
所述发送模块设置于主交换机上, 用于发送软复合报文, 所述软复合报 文包括与执行软分离命令时的软分离点相匹配的软复合点信息;
所述堆叠拓朴管理模块设置于从交换机上, 用于根据所述软复合报文中 的软复合点信息, 按照拓朴计算规则计算堆叠拓朴, 并根据堆叠拓朴的计算 结果修改转发表。
由以上技术方案可以看出, 本发明的实施例提供的堆叠交换机分离和复 合的方法及系统, 通过执行软分离命令或软复合命令, 触发主交换机向堆叠 成员发送软分离或软复合报文, 其中包括了软分离点或软复合点的信息, 堆 叠成员能够根据该软分离点或软复合点的信息, 按照拓朴计算规则计算堆叠 拓朴, 并根据计算的结果修改转发表, 因而在将从交换机移出或复合到堆叠 交换机时, 不会造成报文丟失, 能够实现从交换机无缝离开 /加入堆叠交换机。 附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案, 下面将对实 施例或现有技术描述中所需要使用的附图作简单地介绍, 显而易见地, 下面 描述中的附图仅仅是本发明的一些实施例, 对于本领域普通技术人员来讲, 在不付出创造性劳动的前提下, 还可以根据这些附图获得其它的附图。
图 1是菊花链式堆叠交换机应用示意图;
图 2是本发明实施例中堆叠交换机分离的方法的流程图;
图 3是本发明的实施例中堆叠交换机复合的方法的流程图;
图 4是本发明提供的堆叠交换机分离和复合的方法的应用实施例的流程 图;
图 5是本发明提供的堆叠交换机分离的方法的实施例从交换机 II移出前 的堆叠示意图;
图 6是本发明提供的堆叠交换机分离的方法的实施例从交换机 II软分离 后的堆叠示意图;
图 7是本发明提供的堆叠交换机复合的方法的实施例从交换机 II移出后 的堆叠示意图;
图 8是本发明提供的堆叠交换机复合的方法的实施例从交换机 II移出并 经过软复合后的堆叠示意图;
图 9是本发明实施例提供的堆叠交换机分离和复合的系统的示意图。 具体实施方式
本发明实施例提供了一种堆叠交换机分离和复合的方法及系统, 使从交 换机离开 /加入堆叠交换机的菊花链式环形拓朴时, 不会造成报文丟失, 能够 实现从交换机无缝离开 /加入堆叠交换机。
为使本发明的目的、 技术方案和优点更加清楚明白, 下面结合实施例和 附图, 对本发明进一步详细说明。
如图 2所示, 本发明实施例提供了一种堆叠交换机分离的方法, 可以包 括如下步骤:
5201、 执行软分离命令, 所述软分离命令触发主交换机将软分离报文发 送给堆叠成员, 所述软分离报文包括软分离点信息。 所述软分离点信息包括 所述软分离命令要分离的交换机的编号和堆叠端口号。 可以将要分离的某个 交换机的某个堆叠端口称为软分离点。
5202、 所述各堆叠成员根据所述软分离点信息, 按照拓朴计算规则计算 堆叠拓朴并根据堆叠拓朴的计算结果修改转发表。
本发明的实施例还提供了一种堆叠交换机复合的方法, 如图 3 所示, 该 方法可以包括:
5301、 按照图 2所示的实施例实现了堆叠交换机的分离时, 在软分离点 处移出或加入从交换机。
5302、 执行软复合命令, 所述软复合命令触发主交换机将软复合报文发 送给各堆叠成员, 所述软复合报文包括与所述软分离点匹配的软复合点信息。 所述软分离点由图 2所示的实施例中软分离报文中要分离的交换机的编 号和堆叠端口号决定, 所述软复合点信息包括所述软复合命令要复合的交换 机的编号和堆叠端口号。 例如, 图 2所述的实施例将某个交换机的某个端口 作为软分离点, 那么在软复合时, 则将所述的某个交换机的某个端口作为软 复合点, 即通知其它的交换机前述的软分离点处已经能够使用, 无需再断开。
S303、 所述各堆叠成员能够根据所述软复合点信息, 按照拓朴计算规则 计算堆叠拓朴并根据计算的结果修改转发表。
本发明实施例提供的堆叠交换机分离和复合的方法, 通过执行软分离命 令或软复合命令, 触发主交换机向堆叠成员发送软分离或软复合报文, 其中 包括了软分离点或软复合点的信息, 堆叠成员能够根据软分离点或软复合点 信息, 按照拓朴计算规则计算堆叠拓朴并根据堆叠拓朴的计算结果修改转发 表, 因而在将从交换机移出或复合到堆叠交换机时, 不会造成报文丟失, 从 而能够实现从交换机无缝离开 /加入堆叠交换机。
图 4是本发明提供的堆叠交换机分离和复合的方法的一个应用实施例的 流程图。 该方法可以包括如下步骤:
S401 : 在主交换机上执行 soft-detach (软分离 )命令, 该命令触发主交换 机将 soft-detach报文发送给除自身之外的所有堆叠成员; 这里, 堆叠成员指 的是组成堆叠交换机的各交换机, 包括主交换机和所有的从交换机; soft-detach才艮文可以设计为 switch stack-port <slot/port> soft-detach, 其中 slot 指的是交换机的编号, port指的是堆叠端口号, slot和 port共同决定软分离时 应该断开哪一个交换机的哪一个堆叠端口, 比如 slot/port如果为 1/1 , 指的就 是 I号交换机的堆叠端口 1是软分离点,将会被断开;需要说明的是, soft-detach 命令事实上也是要求堆叠成员改变堆叠拓朴的命令, soft-detach报文携带的软 分离点信息则是堆叠成员进行拓朴计算的依据;
S402: 堆叠成员接收到 S401中的 soft-detach报文后, 改变堆叠拓朴和转 发表, 目的是为了改变业务数据流走向, 使菊花链式环形拓朴变为链形, 避 开软分离点;
S403: 在完成 S402后, 将堆叠交换机在软分离点处物理断开或关闭, 然 后从交换机离开或新的从交换机加入堆叠, 并用堆叠线连接各堆叠成员, 使 之再次成为菊花链式;
S404: 在主交换机上执行 no soft-detach (软复合)命令, 该命令触发主 交换机将 no soft-detach报文发送给除自身之外的所有堆叠成员, 其中 no soft-detach报文中包括了软复合点的信息; 这里,可以将 no soft-detach报文设 计为 no switch stack-port <slot/port> soft-detach, 与 soft-detach才艮文相对应 , 其 中 slot指的是交换机的编号, port指的是堆叠端口号, slot和 port共同决定软 复合点,该软复合点应与 soft-detach报文中的软分离点相匹配,表明 soft-detach 中的软分离点处已经能够使用, 无需再断开; 需要说明的是, no soft-detach 命令事实上也是要求堆叠成员改变堆叠拓朴的命令, no soft-detach报文携带 的软复合点信息则是堆叠成员进行拓朴计算的依据;
S405: 堆叠成员接收到 S404中的 no soft-detach报文后, 改变堆叠拓朴和 转发表, 恢复菊花链式环形拓朴。
下面结合图 5、 图 6、 图 7和图 8, 以移出从交换机 II为例对上述方法进 行更伴细的说明。
在图 5中, 主交换机 I和从交换机 II、 III、 IV的两个堆叠端口 1、 2通过 堆叠线依次连接, 形成菊花链式环形拓朴。 在如图 5 所示的例子中, 堆叠成 员 I和 III通过成员 II的链路传输数据流。
此时, 在堆叠成员 I上的转发表如下:
目的交换机 出堆叠端口
II Stack Port 1
III Stack Port 1
IV Stack Port 2
在堆叠成员 III上的转发表如下:
目的交换机 出堆叠端口 I Stack Port 2
II Stack Port 2
IV Stack Port 1
这里, 欲将从交换机 II移出, 以断开主交换机 I的堆叠端口 1为例进行 说明(事实上, 断开从交换机 II的堆叠端口 2也是可以的)。 在主交换机上执 行 soft-detach命令, 该命令触发主交换机通过堆叠协议向其他各堆叠成员同 步 switch stack-port <I/1> soft-detach报文, 根据软分离报文中携带的成员编号 和堆叠端口编号信息(即 <1/1> ), 各堆叠成员可以得知 I号交换机的堆叠端口 1将被断开, 于是重新计算拓朴, 改变堆叠拓朴和转发表。 具体而言, 各成员 交换机收到软分离报文后, 成员交换机根据报文中的软分离点信息 (即 1/1断 开的信息)进行拓朴计算, 重新计算出各堆叠成员相对于本交换机的拓朴位 置和路径距离, 根据最短路径优先原则重新计算出到各成员交换机流量的出 堆叠口, 如此, 原来的菊花链式环行拓朴即被修改为链形拓朴。
经过拓朴计算后, 堆叠成员改变其转发表, 这时, 在堆叠成员 I上的转发 表变更为:
Figure imgf000009_0001
堆叠成员的转发表变更后, 堆叠交换机上的业务数据流转发路径如图 6 所示, 堆叠成员 I和 III之间的数据流由原来的经过堆叠成员 II变为经过堆叠 成员 IV, 在堆叠成员 I和 II之间的堆叠线上没有了流量。 然后再从物理上将 1/1口的堆叠线断开, 即断开堆叠环, 由于 I和 III之间的数据流已经不在 I和 II之间的链路上传输, 所以业务不会受影响, 堆叠环无损地变为链形拓朴, 此时即可将从交换机 II顺利移出而不影响其它堆叠成员间的业务。
当将从交换机 II移出后, 将交换机 I的堆叠端口 1和从交换机 III的堆叠 端口 2用堆叠线连接, 堆叠成员 I、 III、 IV之间再次形成了菊花链式连接。 这 时, 堆叠拓朴如图 7所示, 环形拓朴中有 I、 III和 IV三个交换机, 堆叠成员 I和 III之间的数据流仍然经过堆叠成员 IV传输。
然后, 在主交换机上执行 no soft-detach命令, 该命令触发主交换机通过 堆叠协议向各堆叠成员同步 no switch stack-port <I/1> soft-detach才艮文到各成 员交换机(这里, 发送 no switch stack-port <III/2> soft-detach也可以), 根据 软复合 文中携带的成员编号和堆叠端口编号信息 (即<1/1> ), 各堆叠成员可 以得知 I号交换机的堆叠端口 1已经能够使用, 于是重新计算拓朴, 改变堆叠 拓朴和转发表, 恢复环形拓朴。 具体而言, 各成员交换机收到软复合报文后, 成员交换机根据报文中的软复合点信息 (即 1/1复合的信息 )进行拓朴计算, 重新计算出各堆叠成员相对于本交换机的拓朴位置和路径距离, 根据最短路 径优先原则重新计算出到各成员交换机流量的出堆叠口, 如此, 链形拓朴即 重新被修改为菊花链式环行拓朴。
经过拓朴计算后, 堆叠成员改变其转发表, 这时, 在堆叠成员 I上的转发 表变更为:
Figure imgf000010_0001
此时,如图 8所示,堆叠成员 I和 III之间的数据流不再经过堆叠成员 IV, 而是选择 I和 III之间直接相连的链路。
至于加入从交换机的方法, 和上面的移出从交换机 II类似, 这里不再贅 述。
由上述堆叠交换机分离或复合的方法实施例可以看出, 本发明实施例提 供的堆叠交换机分离或复合的方法, 在物理上断开堆叠环或关闭堆叠端口之 前, 先将数据流改为链形拓朴, 在复合时, 再将链形拓朴改为环形拓朴, 因 而将从交换机移出或复合到堆叠交换机时, 不会造成短时间的报文丟失, 从 而能够实现从交换机无缝离开 /加入堆叠交换机。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分步骤 是可以通过程序来指令相关的硬件完成, 所述的程序可以存储于计算机可读 存储介质中, 所述存储介质可以是 ROM/RAM, 磁盘或光盘等。 如图 9所示, 本发明实施例还提供了一种堆叠交换机分离的系统, 包括 堆叠拓朴管理模块 901和发送模块 902,所述堆叠拓朴管理模块 901与发送模 块 902通信;
所述发送模块 902设置于主交换机上, 用于发送软分离报文, 所述软分 离报文包括软分离点信息;
所述堆叠拓朴管理模块 901 设置于从交换机上, 用于根据所述软分离点 信息, 按照拓朴计算规则计算堆叠拓朴, 并根据堆叠拓朴的计算结果修改转 发表。
其中, 所述软分离点信息包括所述堆叠交换机分离的系统要分离的交换 机的编号和堆叠端口号。 可以将某个要分离的交换机的某个堆叠端口称为软 分离点。
本发明的实施例还提供了一种堆叠交换机复合的系统, 包括: 堆叠拓 朴管理模块和发送模块, 所述堆叠拓朴管理模块与发送模块通信;
所述发送模块设置于主交换机上, 用于发送软复合报文, 所述软复合报 文包括与执行软分离命令时的软分离点相匹配的软复合点信息; 所述堆叠拓朴管理模块设置于从交换机上, 用于根据所述软复合报文中 的软复合点信息, 按照拓朴计算规则计算堆叠拓朴, 并根据堆叠拓朴的计算 结果修改转发表。
其中, 所述软分离点由执行软分离命令时的软分离报文中要分离的交换 机的编号和堆叠端口号决定。 所述软复合点信息包括所述堆叠交换机复合的 系统要复合的交换机的编号和堆叠端口号。
本实施例堆叠交换机复合的系统中, 发送模块和堆叠拓朴管理模块之间 的关系也可以参考图 9。
上述堆叠交换机分离的系统与复合的系统实施例的具体实施方式, 可以 参考本发明提供的堆叠交换机分离和复合的方法各实施例, 这里不再贅述。
由上述堆叠交换机分离和复合的系统实施例可以看出, 本发明实施例提 供的堆叠交换机分离和复合的系统, 在物理上断开堆叠环或关闭堆叠端口之 前, 先将数据流改为链形拓朴, 在复合时, 再将链形拓朴改为环形拓朴, 因 而将从交换机移出或复合到堆叠交换机时, 不会造成短时间的报文丟失, 从 而能够实现从交换机无缝离开 /加入堆叠交换机。
以上对本发明所提供的一种堆叠交换机分离或复合的方法及系统进行了 以上实施例的说明只是用于帮助理解本发明的方法及其核心思想; 同时, 对 于本领域的一般技术人员, 依据本发明的思想, 在具体实施方式及应用范围 上均会有改变之处, 综上所述, 本说明书内容不应理解为对本发明的限制。

Claims

权利 要求 书
1、 一种堆叠交换机分离的方法, 其特征在于, 包括:
执行软分离命令, 所述软分离命令触发主交换机将软分离报文发送给堆叠 成员, 所述软分离报文包括软分离点信息;
所述堆叠成员根据所述软分离报文中的软分离点信息, 按照拓朴计算规则 计算堆叠拓朴并根据堆叠拓朴的计算结果修改转发表。
2、 根据权利要求 1所述的方法, 其特征在于, 所述软分离点信息包括所述 软分离命令要分离的交换机的编号和堆叠端口号。
3、 一种堆叠交换机复合的方法, 其特征在于, 包括:
当根据权利要求 1 所述的方法实现了堆叠交换机的分离时, 在软分离点处 移出或力 p入从交换机;
执行软复合命令, 所述软复合命令触发主交换机将软复合报文发送给堆叠 成员, 所述软复合报文包括与所述软分离点匹配的软复合点信息;
所述各堆叠成员根据所述软复合报文中的软复合点信息, 按照拓朴计算规 则计算堆叠拓朴并根据堆叠拓朴的计算结果修改转发表。
4、 根据权利要求 3所述的方法, 其特征在于, 所述软分离点由软分离报文 中要分离的交换机的编号和堆叠端口号决定。
5、 根据权利要求 3或 4所述的方法, 其特征在于, 所述软复合点信息包括 所述软复合命令要复合的交换机的编号和堆叠端口号。
6、 一种堆叠交换机分离的系统, 其特征在于, 包括: 堆叠拓朴管理模块和 发送模块, 所述堆叠拓朴管理模块与发送模块通信;
所述发送模块设置于主交换机上, 用于发送软分离报文, 所述软分离报文 包括软分离点信息;
所述堆叠拓朴管理模块设置于从交换机上, 用于根据所述软分离报文中的 软分离点信息, 按照拓朴计算规则计算堆叠拓朴, 并根据堆叠拓朴的计算结果 修改转发表。
7、 根据权利要求 6所述的系统, 其特征在于, 所述软分离点信息包括所述 堆叠交换机分离的系统要分离的交换机的编号和堆叠端口号。
8、 一种堆叠交换机复合的系统, 其特征在于, 包括: 堆叠拓朴管理模块和 发送模块, 所述堆叠拓朴管理模块与发送模块通信;
所述发送模块设置于主交换机上, 用于发送软复合报文, 所述软复合报文 包括与权利要求 6中所述软分离点匹配的软复合点信息;
所述堆叠拓朴管理模块设置于从交换机上, 用于根据所述软复合报文中的 软复合点信息, 按照拓朴计算规则计算堆叠拓朴, 并根据堆叠拓朴的计算结果 修改转发表。
9、 根据权利要求 8所述的系统, 其特征在于, 所述软分离点由软分离报文 中要分离的交换机的编号和堆叠端口号决定。
10、 根据权利要求 8或 9所述的的系统, 其特征在于, 所述软复合点信息 包括所述堆叠交换机复合的系统要复合的交换机的编号和堆叠端口号。
PCT/CN2009/071741 2008-07-25 2009-05-12 堆叠交换机分离和复合的方法及系统 WO2010009639A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN2008101426542A CN101335716B (zh) 2008-07-25 2008-07-25 一种堆叠交换机分离或复合的方法及设备
CN200810142654.2 2008-07-25

Publications (1)

Publication Number Publication Date
WO2010009639A1 true WO2010009639A1 (zh) 2010-01-28

Family

ID=40198031

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2009/071741 WO2010009639A1 (zh) 2008-07-25 2009-05-12 堆叠交换机分离和复合的方法及系统

Country Status (2)

Country Link
CN (1) CN101335716B (zh)
WO (1) WO2010009639A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111698155A (zh) * 2020-06-17 2020-09-22 深圳市信锐网科技术有限公司 一种stp协议报文分布式处理方法、系统及相关设备

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101335716B (zh) * 2008-07-25 2011-06-08 华为技术有限公司 一种堆叠交换机分离或复合的方法及设备
CN101599909B (zh) * 2009-07-08 2012-02-08 杭州华三通信技术有限公司 一种流量转发的方法和装置
CN103780496B (zh) * 2012-10-17 2017-07-14 新华三技术有限公司 应用于星型堆叠系统中的报文转发方法和设备
CN105939226A (zh) * 2016-07-01 2016-09-14 浪潮集团有限公司 一种基于三层交换机的冗余网络设备
CN108650118B (zh) * 2018-04-20 2020-12-29 烽火通信科技股份有限公司 堆叠交换机的拓扑管理方法及系统
CN115396385B (zh) * 2021-05-19 2023-07-21 中国移动通信集团浙江有限公司 堆叠交换机快速恢复业务的方法、装置及计算设备

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1859290A (zh) * 2005-10-28 2006-11-08 华为技术有限公司 一种网络拓扑发生变化时网络节点配置信息的处理方法
CN101110782A (zh) * 2007-08-28 2008-01-23 中兴通讯股份有限公司 一种用于以太网交换机堆叠系统管理的方法
CN101141404A (zh) * 2007-10-16 2008-03-12 中兴通讯股份有限公司 堆叠系统拓扑管理方法及拓扑改变通知方法
EP1655905B1 (en) * 2004-10-28 2008-05-07 Alcatel Lucent Stack switch manager protocol with automatic assignment of stack switch identifiers
CN101335716A (zh) * 2008-07-25 2008-12-31 华为技术有限公司 一种堆叠交换机分离或复合的方法及设备
CN101345708A (zh) * 2008-08-25 2009-01-14 杭州华三通信技术有限公司 一种生成转发表的方法和装置

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN100370740C (zh) * 2004-03-06 2008-02-20 鸿富锦精密工业(深圳)有限公司 堆叠式交换机管理方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1655905B1 (en) * 2004-10-28 2008-05-07 Alcatel Lucent Stack switch manager protocol with automatic assignment of stack switch identifiers
CN1859290A (zh) * 2005-10-28 2006-11-08 华为技术有限公司 一种网络拓扑发生变化时网络节点配置信息的处理方法
CN101110782A (zh) * 2007-08-28 2008-01-23 中兴通讯股份有限公司 一种用于以太网交换机堆叠系统管理的方法
CN101141404A (zh) * 2007-10-16 2008-03-12 中兴通讯股份有限公司 堆叠系统拓扑管理方法及拓扑改变通知方法
CN101335716A (zh) * 2008-07-25 2008-12-31 华为技术有限公司 一种堆叠交换机分离或复合的方法及设备
CN101345708A (zh) * 2008-08-25 2009-01-14 杭州华三通信技术有限公司 一种生成转发表的方法和装置

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111698155A (zh) * 2020-06-17 2020-09-22 深圳市信锐网科技术有限公司 一种stp协议报文分布式处理方法、系统及相关设备
CN111698155B (zh) * 2020-06-17 2022-10-04 深圳市信锐网科技术有限公司 一种stp协议报文分布式处理方法、系统及相关设备

Also Published As

Publication number Publication date
CN101335716B (zh) 2011-06-08
CN101335716A (zh) 2008-12-31

Similar Documents

Publication Publication Date Title
WO2010009639A1 (zh) 堆叠交换机分离和复合的方法及系统
US10003552B2 (en) Distributed bidirectional forwarding detection protocol (D-BFD) for cluster of interconnected switches
WO2015180040A1 (zh) 流表管理方法及相关设备和系统
EP2640013A1 (en) Method And Apparatus Providing Network Redundancy And High Availability To Remote Network Nodes
US20120110393A1 (en) Method and apparatus providing failover for a point to point tunnel for wireless local area network split-plane environments
JP6173833B2 (ja) ネットワーク中継システムおよびスイッチ装置
WO2013123846A1 (zh) 一种分布式网络控制方法及装置
CN105450525B (zh) 用于路由交换设备使用的方法和设备
US8861334B2 (en) Method and apparatus for lossless link recovery between two devices interconnected via multi link trunk/link aggregation group (MLT/LAG)
WO2010022576A1 (zh) 基于以太多环网的控制报文的传输方法和装置
WO2017000832A1 (zh) Mac地址的同步方法、装置及系统
WO2013053276A1 (zh) 一种网络路由收敛处理方法和装置
WO2014059867A1 (zh) 链路聚合组中流量路径的协商方法及装置
WO2014206103A1 (zh) 一种数据包交换方法、装置以及接入交换机和交换系统
WO2021082812A1 (zh) 报文的发送方法和第一网络设备
EP3213441B1 (en) Redundancy for port extender chains
WO2017113929A1 (zh) 一种软件定义网络中控制转发设备集群的方法及控制器
CN102821099B (zh) 报文转发方法、设备及系统
WO2012163032A1 (zh) 处理组播的方法和装置
WO2011009248A1 (zh) 基于主控板倒换的数据处理方法、线卡及主控板
WO2012000374A1 (zh) 接口业务集中处理方法和系统
WO2011127849A2 (zh) 一种数据流传送方法及网络设备
US20150301571A1 (en) Methods and apparatus for dynamic mapping of power outlets
WO2012130084A1 (zh) 一种报文多次转向方法、装置和系统
WO2011110037A1 (zh) 一种环网架构下的组播实现方法、接入层设备及系统

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: 09799955

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: 09799955

Country of ref document: EP

Kind code of ref document: A1