CN104378297B - OSPF packets forwarding method and apparatus - Google Patents

OSPF packets forwarding method and apparatus Download PDF

Info

Publication number
CN104378297B
CN104378297B CN201310357517.1A CN201310357517A CN104378297B CN 104378297 B CN104378297 B CN 104378297B CN 201310357517 A CN201310357517 A CN 201310357517A CN 104378297 B CN104378297 B CN 104378297B
Authority
CN
China
Prior art keywords
mac address
gateway node
network
vlan
node
Prior art date
Application number
CN201310357517.1A
Other languages
Chinese (zh)
Other versions
CN104378297A (en
Inventor
宋玉兵
Original Assignee
新华三技术有限公司
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 新华三技术有限公司 filed Critical 新华三技术有限公司
Priority to CN201310357517.1A priority Critical patent/CN104378297B/en
Publication of CN104378297A publication Critical patent/CN104378297A/en
Application granted granted Critical
Publication of CN104378297B publication Critical patent/CN104378297B/en

Links

Abstract

本发明公开了一种报文转发方法及设备,网关节点通过一可达大二层网络中的所有节点设备的广播树,通告本节点设备配置的虚MAC地址及所属VLAN;非网关节点设备通过广播树接收到各个网关节点通告的该网关节点配置的虚MAC地址以及所属VLAN后,生成对应的第一MAC地址转发表项;当非网关节点接收到目的MAC地址为网关节点的虚MAC地址的数据报文时,根据该目的MAC地址在第一MAC地址转发表项中的匹配表项,直接转发该数据报文而不进行大二层网络协议封装;网关节点接收到该数据报文,且确定该数据报文的目的MAC地址为自身虚MAC地址时,根据该数据报文的目的IP地址,查找路由表转发该数据报文。 The present invention discloses a method and apparatus for forwarding packets, a gateway node tree up to all the node devices broadcast a large Layer 2 network, virtual MAC address and the VLAN configuration of the advertisement node equipment; non-gateway node devices receiving a broadcast configuration tree to each gateway node of the advertisement node gateway and the virtual MAC address belongs to VLAN, to generate a first MAC address corresponding forwarding entry; if the received virtual MAC address destination MAC address of the gateway node's gateway node data packet, based on the destination MAC address in a first MAC address forwarding matching entry in the entry and forwards the packet data network without large Layer protocol encapsulation; gateway node receives the data packet, and determining the destination MAC address of the data packet to its own virtual MAC address, the destination IP address of the data packet, the routing table to forward the data packet. 在本发明中,实现了基于大二层网络的L3单播流量转发。 In the present invention achieved based on L3 forwarding unicast layer 2 network traffic flow.

Description

一种报文转发方法及设备 OSPF packets forwarding method and apparatus

技术领域 FIELD

[0001] 本发明涉及通信领域的组播技术,尤其涉及一种报文转发方法及设备。 [0001] The present invention relates to a multicast communication field, particularly to a method and apparatus for packet forwarding.

背景技术 Background technique

[0002] 传统的三层数据中心架构中采用STP (Spanning Tree Protocol,生成树协议)来优化客户端到服务器的路径和支持连接冗余。 To optimize client to server connection support redundant paths and [0002] the traditional three-tier data center architecture using STP (Spanning Tree Protocol, the Spanning Tree Protocol). 由于STP的收敛性能等原因,一般情况下STP 的网络规模不会超过100台交换机。 Due to the convergence of performance such as STP, STP network size under normal circumstances will not exceed 100 switches. 同时由于STP需要阻塞冗余设备和链路,也降低了网络资源带宽利用率。 At the same time due to the need for STP blocking redundant links and equipment, but also reduces bandwidth utilization of network resources. 传统的基于STP备份设备和链路方案已经不能满足数据中心规模、带宽的需求,并且STP协议几秒至几分钟的故障收敛时间,也不能满足数据中心的可靠性要求。 Traditional backup device reliability based on STP link programs and data centers can not meet the size, bandwidth requirements, and the STP seconds to minutes convergence time of the fault, the data center can not be satisfied. 因此,需要能够有新的技术,在满足二层网络规模的同时,也能够充分利用冗余设备和链路, 提升链路利用率,而且数据中心的故障收敛时间能够降低到亚秒甚至毫秒级。 Thus, the need to have new technologies, while at Layer network scale, it is possible to take advantage of redundant links and devices, improve link utilization, and fault data center convergence time can be reduced to seconds or sub-millisecond .

[0003] 虚拟化能够有效地提高服务器的利用率,降低能源消耗,降低客户的运维成本,所以虚拟化技术得到了极大的发展。 [0003] virtualization can improve server utilization, reduce energy consumption, lower operation and maintenance costs of customers, so virtualization technology has been greatly developed. 虚拟化给数据中心带来的不仅是服务器利用率的提高, 还有网络架构的变化。 Virtualized data center brings not only improve server utilization, as well as changes in the network architecture. 虚拟机(Virtual Machine,VM)迀移技术可以使数据中心的计算资源得到灵活的调配,进一步提高虚拟机资源的利用率。 VM (Virtual Machine, VM) Gan-shift technology allows data center computing resources to obtain flexible deployment of virtual machines to further improve the utilization of resources. 但是虚拟机迀移要求虚拟机迀移前后的IP和MAC (Media AcCeSS Contr01,介质访问控制)地址不变,这就需要虚拟机迀移前后的网络处于同一个二层域内部。 But virtual machine virtual machine requirements shift Gan Gan IP before and after the shift and MAC (Media AcCeSS Contr01, MAC) address of the same, which is in the same network within a two-story front and rear virtual machine needs Gan shift. 由于客户要求虚拟机迀移的范围越来越大,甚至是跨越不同地域、不同机房之间的迀移,所以使得数据中心二层网络的范围越来越大,甚至出现了大规模二层网络(下文中简称为大二层网络)这一新领域专题。 Because virtual machines to customer requirements Gan shift range is growing, even across different regions, Gan move between the different rooms, so make the scope of the data center layer 2 network is growing, even a large-scale Layer 2 networks. (hereinafter referred to as large Layer 2 network) topic in this new field.

[0004] 隧道技术能够实现覆盖多个接入交换机和核心交换机的数据中心内的大二层网络。 [0004] Tunneling enables a large coverage in a plurality of Layer 2 network access and core switches in the data center. 隧道技术的典型代表是TRILL(TRansparent Interconnection of Lots of Links,多链路透明互联)和SPB (Shortest Path Bridging,最短路径桥)技术,两者均是通过借用IS_IS(Intermediate System to Intermediate System,中间系统到中间系统)路由协议的计算和转发模式,实现二层网络的大规模扩展。 Tunneling is typical TRILL (TRansparent Interconnection of Lots of Links, multilink Transparent Interconnection) and SPB (Shortest Path Bridging, the shortest path bridge) technique, both of which are by borrowing IS_IS (Intermediate System to Intermediate System, Intermediate System to intermediate system) protocol route calculation and forward mode, large-scale expansion layer 2 network. 这些技术目前正在标准化过程中,同时传统交换机不仅需要软件升级,还需要硬件支持。 These technologies are currently under standardization process, while traditional switch requires not only a software upgrade, you also need hardware support.

[0005] 现有大二层组网技术中,对于大二层环境下的跨VLAN的三层转发,还没有比较好的解决办法,限制了大二层组网的实际部署。 [0005] existing networking technology in large two-story, three-layer for inter-VLAN environment under the large two-story forward, there is no better solution, limiting the actual deployment of a large two-story networking.

[0006] 以Trill组网为例,当前Trill标准相关的RFC刚发布不久,配套的芯片尚不成熟, 相关芯片在一个pipeline中处理完Trill后,就不能再做L3转发;执行完L3转发后,就不能再做Tr ill处理。 After performing L3 forwarding; [0006] to Trill networking, for example, related to the current standard RFC Trill just released soon, supporting the chip is not yet mature, and related chips in a pipeline in after processing the Trill, you can not do L3 forwarding , you can not do Tr ill treatment.

[0007] 针对该问题,现有技术中有如下两种解决方式: [0007] To solve this problem, the prior art has the following two solutions:

[0008] 方式一、采用代理转发 [0008] In one embodiment, using forwarding agent

[0009] Tri 11板卡接收Tri 11报文,完成Tri 11终结后,若需要做L3转发,则将报文发送给L3代理板进行处理转发; [0009] Tri 11 Tri 11 card receiving packets, after the completion of end Tri 11, do For L3 forwarding, the packets will be sent for processing to a plate L3 forwarding agent;

[0010] L3代理板完成正常L3代理转发后,若有外部接收者,则将报文发送给外网;若Trill域内有接收者,报文需要再次转发到Trill板卡的专用回环口(Loopback端口),报文回环口回环后完成Trill封装,根据封装后的Trill报文查找Trill单播表,从相应的端口发送给Trill网络,其处理流程示意图可以参见图1。 [0010] L3 L3 agent board to complete the normal forwarding agent, if an external recipient, then transmits the packet to the external network; if the recipient has Trill art, packets need to be forwarded again to the loop port dedicated board Trill (the Loopback port), the packet is completed after loopback port loopback Trill package, according to the encapsulated packet Trill Trill unicast lookup table, the corresponding port Trill sent to the network, which may refer to a schematic process flow FIG.

[0011] 方式二、Trill/L3分离处理 [0011] The second approach, Trill / L3 separation process

[0012] Trill域外增加一层专门的设备来执行L3转发的功能。 [0012] Trill increased outside specialized apparatus to perform the L3 layer forwarding functions. Trill的终结点在core节点,core节点以上采用不带封装的native以太网报文转发,上行到L3设备处进行L3转发处理。 Trill endpoints employed in the above core node, a node without a core encapsulated native Ethernet packet forwarding, at the device for the uplink L3 to L3 forwarding.

[0013] 此外,现有大二层网络技术中,对于L3转发,目前存在的网关备份协议,存在固有的分担不均的问题。 [0013] In addition, the prior art a large layer 2 network, for L3 forwarding, the existing backup protocol gateway, there is an inherent problem of uneven sharing. 对于大规模的云计算数据中心(core节点大于4个),当数据中心内部跨VLAN的L3流量较大时,会形成转发瓶颈。 For large cloud computing data center (core node is greater than 4), when the L3 traffic data center across a large internal VLAN, the forwarding form a bottleneck.

[0014] 针对该问题,现有技术中,对于控制平面,通过建立备份组,并选举一个master, 客户端的所有ARP请求都只能由master回应,master根据一定的负载分担算法,为不同的ARP请求回应不同的网关ARP信息,这样不同的的主机获得的ARP回应不同。 [0014] To solve this problem, the prior art, the control plane, through the establishment of a backup group, and elected a master, all ARP requests the client can only respond by the master, master-sharing algorithm based on certain load, for different ARP gateway ARP request to respond to different information, different ARP response to such different hosts available.

[0015] 对于数据平面,不同主机封装的报文由于网关ARP不同,被发送给了不同的网关, 从而实现不同主机流量在不同L3网关出的分担。 [0015] For the data plane, different hosts encapsulated packets due to different gateways ARP, is sent to a different gateway, in order to achieve different hosts on different traffic sharing the L3 gateway.

[0016] 在实现本发明的过程中,发明人发现现有技术中至少存在以下问题: [0016] During the implementation of the present invention, the inventor finds at least the following problems in the prior art:

[0017] 对于现有大二层环境下的跨VLAN的三层转发问题的解决方式中,方式一需要集中式处理,性能较差,且代理配置复杂,需要专门的代理板卡或回环端口,造成资源浪费;方式二中需要额外的L3设备,设备层次增加,增加了成本,同时,Trill终结点和L3网关中间的链路存在很多的环路,需要考虑环路避免技术,且L3转发处理的性能还是较低。 Solution [0017] forwarded to the problem of inter-VLAN Layer under the existing two-story great environment, a way needs to centralized processing, poor performance, and the proxy configuration complexity, require specialized proxy cards or loopback port, waste of resources; the second approach requires additional L3 equipment, increased equipment levels, increased costs, and, Trill endpoint and gateway links L3 middle there are many loop, loop avoidance technology needs to be considered, and L3 forwarding the performance is low.

[0018] 对于现有网关备份协议分担不均问题的解决方式中,对于同一个客户端发送的不同数据流,无法在不同L3网关之间分担,造成分担不均。 Solution [0018] The problem of a disproportionate share of prior backup gateway protocol, different data streams transmitted to the same client, the gateway can not be shared between different L3, resulting in a disproportionate share.

发明内容 SUMMARY

[0019] 本发明实施例提供一种报文转发方法及设备,用以在大二层网络中实现L3单播流量转发。 Embodiment [0019] The present invention provides a packet forwarding method and apparatus for a large network, which implements Layer L3 forwarding unicast traffic.

[0020] 为了达到以上目的,本发明实施例提供了一种报文转发方法,应用于多个节点设备基于大二层网络协议组建的大二层网络中,所述多个节点设备包括作为网关设备的网关节点和作为接入设备的非网关节点,同一VLAN的各网关节点配置有相同的虚IP地址和虚MAC地址,该方法包括: [0020] To achieve the above object, embodiments of the present invention there is provided a packet forwarding method applied to a large plurality of node devices Layer Layer large networking based on network protocols, said apparatus comprising a plurality of nodes as a gateway gateway node device and a gateway node as the non-access device, the same VLAN gateway nodes each configured with the same virtual IP address and the virtual MAC address, the method comprising:

[0021] 非网关节点通过一可达所述大二层网络中的所有节点设备的广播树,接收各个网关节点通告的该网关节点配置的虚MAC地址以及所属VLAN信息; Configuration [0021] All non-broadcast by the gateway node tree nodes of the device up to a large Layer 2 network, each gateway node receiving the advertisement gateway node virtual MAC address and the VLAN information;

[0022] 非网关节点根据大二层网络协议和接收到的同一虚MAC地址对应的各个网关节点的大二层网络标识,计算去往同一VLAN的各网关节点的单播路由,进而生成所述虚MC地址对应的第一MAC地址转发表项,其中,所述第一MAC地址转发表项中的VLAN为所述虚MAC所属VLAN、目的MAC地址为所述虚MAC地址、出接口为所述单播路由中代价最小的路由对应的出接口; [0022] According to a non-gateway node identified large Layer 2 network gateway node of each network protocol and the same Layer large virtual MAC address corresponding to the received calculated for each unicast route to the gateway node of the same VLAN, thereby generating a a first virtual MAC address of the MC address forwarding table entries, wherein the first MAC address forwarding table to the virtual MAC VLAN entry in the VLAN, destination MAC address is the virtual MAC address, said interface unicast route with the lowest cost of the route corresponding to the interface;

[0023] 非网关节点接收到目的MAC地址为一网关节点的虚MAC地址的数据报文时,根据该数据报文的目的MAC地址,查找匹配的第一MAC地址转发表项,并根据查找到的匹配表项,直接转发所述数据报文而不进行大二层网络协议封装。 [0023] When a data packet is received virtual MAC address to the destination MAC address of a gateway node of the non-gateway node, according to the MAC address of the data packet, looks up the MAC address forwarding entries first, and according to the found the matching entry and forwards the packet data network without large Layer protocol encapsulation.

[0024] 其中,该方法还包括: [0024] wherein the method further comprises:

[0025] 在非网关节点与网关节点之间的链路中断时,非网关节点按照更新后的拓扑,重新计算去往同一VLAN的各网关节点的单播路由,并更新重新计算得到的单播路由,更新第一MAC地址转发表项的出接口; [0025] When the link between the non-gateway node and the gateway node an interrupt, a non-gateway node according to the updated topology recalculated each unicast route to the gateway node same VLAN and update unicast recalculated route, updating a first MAC address forwarding entries of the interfaces;

[0026] 在网关节点宕机时,非网关节点更新该网关节点配置的虚MAC地址对应的第一MAC 地址转发表项的出接口。 First MAC address [0026] In the gateway node goes down, the non-gateway node configured to update the gateway node virtual MAC address forwarding entries corresponding to the interface.

[0027] 本发明实施例还提供了一种报文转发方法,应用于多个节点设备基于大二层网络协议组建的大二层网络中,所述多个节点设备包括作为网关节点的网关节点和作为接入设备的非网关节点,同一VLAN的各网关节点配置有相同的虚IP地址和虚MAC地址,该方法包括: Example [0027] The present invention also provides a packet forwarding method applied to a large plurality of Layer 2 network node apparatus based on a large set of Layer 2 network protocol, the apparatus comprising a plurality of nodes gateway node as a gateway node as a non-access node and gateway device, each of the same VLAN gateway node configured with the same virtual IP address and the virtual MAC address, the method comprising:

[0028] 网关节点通过一可达所述大二层网络中的所有节点设备的广播树,通告本节点设备配置的虚MAC地址以及所属VLAN信息,以使非网关节点设备根据大二层网络协议和接收到的同一虚MAC地址对应的各个网关节点的大二层网络标识,计算去往同一VLAN的各网关节点的单播路由,进而生成所述虚MAC地址对应的第一MAC地址转发表项,其中,所述第一MAC地址转发表项中的VLAN为所述虚MAC所属VLAN、目的MAC地址为所述虚MAC地址、出接口为所述单播路由中代价最小的路由对应的出接口; [0028] The gateway node broadcasting through all the tree nodes apparatus up to a large Layer 2 network, the configuration of the advertisement node device and the VLAN virtual MAC address information, the gateway node so that the non apparatus according to the large network protocol Layer large Layer 2 network gateway node and identifiers of the same virtual MAC address corresponding to the received calculated for each unicast route to the gateway node of the same VLAN, thereby generating a first MAC address corresponding to the virtual MAC address forwarding entry wherein the first MAC address VLAN forwarding table entry belongs to the virtual MAC VLAN, destination MAC address of the virtual MAC address, the unicast interface minimum route cost of routing outbound interface ;

[0029] 网关节点接收到目的MAC地址为自身虚MAC地址的数据报文,所述数据报文是非网关节点根据报文的目的MAC地址,匹配所述第一MAC地址转发表项后直接转发的且没有进行大二层网络协议封装; [0029] The gateway node to the destination MAC address receives the data packets across the virtual MAC address, the data packet gateway node according to a non-packet destination MAC address matching the MAC address and forwards the first forwarding entry and without major protocol encapsulation layer 2 network;

[0030] 所述网关节点在判断出所述数据报文的目的MAC地址为自身虚MAC地址时,根据所述数据报文的目的IP地址,查找路由表转发所述数据报文。 [0030] In the gateway node is determined that the data packet when the destination MAC address for its own virtual MAC address, the destination IP address of the data packet, the routing table to forward the data packet.

[0031] 其中,该方法还包括: [0031] wherein the method further comprises:

[0032] 各个节点设备通过所述广播树,将新学习到的本地MAC地址以及所属VLAN通告给其他节点设备,所述本地MAC地址包括从本节点设备连接的用户网络侧学习到的MAC地址; [0032] The respective node devices through the broadcast tree, the newly learned MAC address and the local network node to learn from the user-side device is connected to the MAC address belongs to another VLAN advertisement node device comprises the local MAC address;

[0033] 网关节点从所述广播树上接收其他节点设备通告的所述其他节点设备新学习到的本地MAC地址以及所属VLAN,并根据预先基于该大二层网络协议生成的大二层网络转发表项,生成所述其他节点设备的本地MAC地址对应的第二MAC地址转发表项,所述第二MAC地址转发表项中的目的MAC地址为所述其他节点设备的本地MAC地址,出接口为所述大二层网络转发表项中去往所述其他节点的出接口; [0033] The gateway node receiving the advertisement of another apparatus other nodes from the broadcast apparatus tree newly learned MAC address and belongs to the local VLAN, based on the large and based on pre-Layer Network Layer Protocol Network turn generated a large entry, the generated second MAC address MAC address of the local node device corresponding to another forwarding entry, the second MAC address destination MAC address forwarding table entry for the local MAC address of the other node device, the interfaces the network is large Layer forwarding items destined for other nodes in the outgoing interface;

[0034] 所述网关节点接收到目的MAC地址不是自身端口MAC地址或所述虚MAC地址的单播数据报文后,根据所述单播数据报文的目的MAC地址及所属VLAN,在所述第二MAC地址转发表项中进行匹配,并通过匹配表项的出接口直接转发所述单播数据报文而不进行大二层网络协议封装。 After [0034] the gateway node receives the destination MAC address is not its own MAC address or port of the virtual MAC address unicast data packets, according to the destination MAC address and the VLAN said unicast data message, the second MAC address forwarding matching entry, and forward the unicast packet data network without large Layer protocol encapsulation matching entry directly through the interface of the publication.

[0035] 其中,所述网关节点还与所述大二层网络外部的路由设备连接; [0035] wherein said gateway node is further connected to the external network routing Layer large equipment;

[0036] 在各个节点设备上设置VLAN的广播域范围,其中,所述广播域范围包括该节点设备上允许该VLAN通过的用户侧端口和该节点设备上的位于该VLAN裁剪树上的大二层网络侧接口,所述裁剪树是基于VLAN对所述广播树进行分发树的剪枝得到的; [0036] range of VLAN broadcast domains is provided in each node device, wherein the broadcast domain located VLAN range includes second year on cut tree allows the VLAN specified by the user-side port of the node device and the node device a network side interface layer, the tree was cut based VLAN pruning tree for the broadcast distribution tree;

[0037] 网关节点在接收到所述路由设备发送的数据报文时,查找匹配所述数据报文的目的IP地址的ARP表项,并根据查找到的ARP表项的目的MAC地址以及出接口,直接转发所述数据报文而不进行大二层网络协议封装。 [0037] In the gateway node receives data packets sent by the routing device, the ARP lookup entry matching the destination IP address of the data packet, and according to the obtained destination MAC address and an ARP entry interfaces and forwards the data packet network without large Layer protocol encapsulation.

[0038] 其中,若网关节点未查找到匹配所述数据报文的目的IP地址的ARP表项,所述方法还包括: [0038] wherein, if the ARP entry gateway node is not a match to said data packet destination IP address, the method further comprising:

[0039] 所述网关节点在该数据报文对应的VLAN的广播域内,直接广播发送用于请求所述数据报文的目的IP地址对应的MAC地址的第一ARP请求报文而不进行大二层网络协议封装, 所述第一ARP请求报文的发送端sender IP地址是所述网关节点的发送端口IP地址,源MAC 地址及发送端MC地址是所述网关节点的发送端口MC地址; The first object of the IP address of the ARP [0039] The gateway node in the broadcast domain to the VLAN corresponding to the packet data directly transmitting broadcast data packet for requesting the MAC address of the corresponding request packet without sophomore network protocol package, the first ARP request packet sender SENDER IP address is the IP address of the transmission port of the gateway node, and sending the source MAC address of MC port address is transmitted to the MC address of the gateway node;

[0040] 所述网关节点根据接收到的针对第一ARP请求报文的第一ARP应答报文,建立所述数据报文的目的IP地址对应的ARP表项,并根据该建立的ARP表项,转发所述数据报文而不进行大二层网络协议封装。 [0040] The gateway node according to the received first ARP response packet for the first packet of the ARP request, the ARP entry destination IP address of the corresponding data packet, and according to the ARP table entry created forwarding the data packet network without large Layer protocol encapsulation.

[0041] 其中,同一VLAN中的一个网关节点被选举为该VLAN的主网关节点,该方法还包括: [0041] wherein a gateway node in the same VLAN gateway is elected as the master node VLAN, the method further comprising:

[0042] 网关节点接收到主机发送的用于请求网关节点的MAC地址的第二ARP请求报文,且确定自身为对应VLAN的主网关节点时,生成第二ARP应答报文,该第二ARP应答报文的目的MAC地址为所述主机的MAC地址,源MAC地址和发送端sender MAC地址均为所述网关节点的虚MAC地址,sender IP地址为所述网关节点的虚IP地址,以使所述主机学习并建立所述网关节点的虚IP地址与虚MC地址之间的关联关系; When [0042] the second gateway node receives ARP request for the MAC address of the gateway node sent by the host request message, and determines its own corresponding primary VLAN gateway node generates a second ARP response message, the second ARP reply message to the destination MAC address of the host MAC address, source MAC address and the MAC address of both the sending end SENDER virtual MAC address of the gateway node, the IP address of the virtual SENDER IP address of the gateway node, so that association between the virtual address and the virtual IP address of the host MC and study of the establishment of said gateway node;

[0043] 所述网关节点根据所述ARP应答报文的目的MAC地址,在第二MAC地址转发表项中进行匹配,并通过匹配表项的出接口直接转发所述ARP应答报文而不进行大二层网络协议封装。 [0043] The gateway node according to the destination MAC address of the ARP response packet, the second MAC address forwarding matching entry, and forward the ARP entry by matching the response packet interfaces directly without large network Layer protocol encapsulation.

[0044] 其中,所述主网关节点的选举包括: [0044] wherein the primary gateway node election comprises:

[0045] 同一VLAN中的各个网关节点在该VLAN内广播携带有各自端口IP地址、端口MAC地址、虚IP及虚MAC地址的邻居发现及主网关节点选举协议报文; [0045] The respective gateway nodes in the same VLAN broadcast carries a respective port IP address and port MAC address, IP address and the virtual MAC address of the virtual master gateway node and neighbor discovery election protocol packets within the VLAN;

[0046] 同一VLAN中的各个网关节点接收其他网关节点发送的所述协议报文,选举端口IP 地址和/或端口MC地址满足预设规则的网关节点,作为所述主网关节点。 [0046] The other gateway node receives the transmitted packet protocol, the election port IP address and / or port address of each gateway node MC same VLAN gateway node satisfies a predetermined rule, as the primary gateway node.

[0047] 其中,还包括: [0047] wherein, further comprising:

[0048] 在非网关节点与网关节点之间的链路中断时,该网关节点更新达到该非网关节点下所有主机的ARP表项的出接口; When [0048] the link between the gateway node and the non-interrupt gateway node, the gateway node to update the outgoing interface to reach all hosts ARP entry in the non-gateway node;

[0049] 在非网关节点发生宕机down时,网关节点删除该非网关节点下的所有主机的ARP 表项; [0049] When the down down occurs in a non-gateway node, the gateway node deletes the ARP entry for all hosts in the non-gateway node;

[0050] VLAN内的主网关节点down时,该VLAN内其他网关节点重新选举该VLAN内的主网关节点; [0050] The gateway node within the primary VLAN down, this VLAN other gateway node re-election of the master gateway node within the VLAN;

[0051] 网关节点在发现外部路由器不可达时,退出本网关节点所在VLAN的主网关节点的选举,以触发其他网关节点重新选举主网关节点,禁用自身所有的三层接口,并通过所述广播树通告虚MAC地址的撤销消息,以使得非网关节点设备更新所述虚MAC地址对应的所述第一MAC地址转发表项的出接口。 Returning gateway node [0051] the gateway node when it finds an external router does not reach, the point where the exit Ben joint VLAN to trigger the re-election of other primary gateway node gateway node, disable all three of its own interface and through the broadcast tree advertisement virtual MAC address withdraw message, so that the non-gateway node device updates the virtual MAC address corresponding to the first MAC address forwarding entry outgoing interface.

[0052] 本发明实施例还提供一种网络设备,应用于多个节点设备基于大二层网络协议组建的大二层网络中,所述多个节点设备包括作为网关设备的网关节点和作为接入设备的非网关节点,同一VLAN的各网关节点配置有相同的虚IP地址和虚MAC地址,所述网络设备作为非网关节点应用于所述大二层网络中,该网络设备包括: [0052] Embodiments of the present invention further provides a network device, a plurality of network nodes applied to large Layer Layer device based on a large set of network protocols, said apparatus comprising a plurality of nodes as a gateway node as an access gateway device and the device non-gateway nodes, the gateway nodes each of the same VLAN configured with the same virtual IP address and the virtual MAC address, the network device as a non-gateway nodes in the floor of the large network, the network device comprising:

[0053] 地址管理模块,用于通过一可达所述大二层网络中的所有节点设备的广播树,接收各个网关节点通告的该网关节点配置的虚MAC地址以及所属VLAN信息; [0053] The address management module configured to broadcast all tree nodes by the device up to a large Layer 2 network, each gateway node arranged to receive the advertisement gateway node virtual MAC address and the VLAN information;

[0054] 表项维护模块,用于根据大二层网络协议和所述地址管理模块接收到的同一虚MAC地址对应的各个网关节点的大二层网络标识,计算去往同一VLAN的各网关节点的单播路由,进而生成所述虚MAC地址对应的第一MAC地址转发表项,其中,所述第一MAC地址转发表项中的VLAN为所述虚MAC所属VLAN、目的MAC地址为所述虚MAC地址、出接口为所述单播路由中代价最小的路由对应的出接口; [0054] The maintenance module entries, each gateway node for the same virtual MAC address based on the received Layer large network management protocol and the address corresponding to the module identifier of the network Layer large, calculated for each gateway node destined for a same VLAN unicast route, thereby generating a first MAC address corresponding to the virtual MAC address forwarding entry, wherein the first MAC address forwarding entry to the virtual MAC VLAN published VLAN belongs, to the destination MAC address the virtual MAC address, the unicast interface corresponding to the route with the lowest cost route the interface;

[0055] 报文转发模块,用于当所述网络设备接收到目的MAC地址为一网关节点的虚MAC地址的数据报文时,根据该数据报文的目的MAC地址,查找匹配的第一MAC地址转发表项,并根据查找到的匹配表项,直接转发所述数据报文而不进行大二层网络协议封装。 When the [0055] message forwarding module, configured to, when the network device receives the destination MAC address is the MAC address of the virtual data packets of a gateway node, according to the MAC address of the data packet, looks for the first MAC matches forwarding entry, according to find a matching entry and forwards the packet data network without large Layer protocol encapsulation.

[0056] 其中,所述表项维护模块还用于, [0056] wherein the entry module is further configured to maintain,

[0057] 在所述网络设备与网关节点之间的链路中断时,按照更新后的拓扑,重新计算去往同一VLAN的各网关节点的单播路由,并更新重新计算得到的单播路由,更新第一MAC地址转发表项的出接口; [0057] When the link between the network device and the gateway node an interrupt, in accordance with the updated topology recalculated each unicast route to the gateway node same VLAN and update unicast route recalculated, updating the first MAC address forwarding entries of the interfaces;

[0058] 在网关节点宕机时,更新该网关节点配置的虚MAC地址对应的第一MAC地址转发表项的出接口。 First MAC address [0058] In the gateway node goes down, updating the configuration of the gateway node virtual MAC address forwarding entry corresponding to the outgoing interface.

[0059] 本发明实施例还提供了一种网络设备,应用于多个节点设备基于大二层网络协议组建的大二层网络中,所述多个节点设备包括作为网关设备的网关节点和作为接入设备的非网关节点,同一VLAN的各网关节点配置有相同的虚IP地址和虚MAC地址,所述网络设备作为网关节点应用于所述大二层网络中,该网络设备包括: [0059] Embodiments of the present invention further provides a network device, a large layer 2 network device based on a plurality of nodes applied to a large set of Layer 2 network protocol, the apparatus comprising a plurality of nodes as a gateway device and a gateway node as a non-gateway node access device, each of the same VLAN gateway node configured with the same virtual IP address and the virtual MAC address, the network device as a gateway node to said large Layer 2 network, the network device comprising:

[0060] 地址管理模块,用于通过一可达所述大二层网络中的所有节点设备的广播树,通告本节点设备配置的虚MAC地址以及所属VLAN信息,以使非网关节点设备根据大二层网络协议和接收到的同一虚MAC地址对应的各个网关节点的大二层网络标识,计算去往同一VLAN的各网关节点的单播路由,进而生成所述虚MAC地址对应的第一MAC地址转发表项,其中,所述第一MAC地址转发表项中的VLAN为所述虚MAC所属VLAN、目的MAC地址为所述虚MAC 地址、出接口为所述单播路由中代价最小的路由对应的出接口; [0060] The address management module configured to broadcast all tree nodes by the device up to a large Layer 2 network, the virtual MAC addresses advertised to the present node device, and the VLAN information so that a non-gateway node apparatus according to the large large Layer 2 network gateway node identifiers of network protocol and the same Layer virtual MAC address corresponding to the received calculated for each unicast route to the gateway node of the same VLAN, thereby generating a first MAC address corresponding to the virtual MAC forwarding entry, wherein the first MAC address VLAN forwarding table entry belongs to the virtual MAC VLAN, destination MAC address of the virtual MAC address, the unicast interface of the route with the smallest cost of the route outbound interface;

[0061] 报文转发模块,用于当所述网络设备接收到目的MAC地址为自身虚MAC地址的数据报文,且判断出所述数据报文的目的MAC地址为自身虚MAC地址时,根据所述数据报文的目的IP地址,查找路由表转发所述数据报文;其中,所述数据报文是非网关节点根据报文的目的MAC地址,匹配所述第一MAC地址转发表项后直接转发的且没有进行大二层网络协议封装。 When the [0061] message forwarding module, configured to, when the network device receives the destination MAC address of the data packets across the virtual MAC address, destination MAC address and it is determined that the data packet is its own virtual MAC address, in accordance with destination IP address of the data packet, routing table look forward the data packet; wherein said non-packet data gateway node according to the destination MAC addresses of the packets, directly after the first matching MAC address forwarding entries and forwarding the network without major Layer protocol encapsulation.

[0062] 其中,该网络设备还包括:表项维护模块; [0062] wherein, the network device further comprising: a maintenance module entries;

[0063] 所述地址管理模块还用于,从所述广播树上接收其他节点设备通告的所述其他节点设备新学习到的本地MAC地址以及所属VLAN;所述本地MAC地址包括从所述其他节点设备连接的用户网络侧学习到的MAC地址; [0063] The address management module is further configured to receive the advertised another apparatus other nodes from the broadcast device to learn new tree local MAC address and the VLAN belongs; the local MAC address from the other comprising a user-side network node apparatus connected MAC address learning;

[0064] 表项维护模块,用于根据预先基于该大二层网络协议生成的大二层网络转发表项,生成所述其他节点设备的本地MAC地址对应的第二MAC地址转发表项,所述第二MAC地址转发表项中的目的MAC地址为所述其他节点设备的本地MAC地址,出接口为所述大二层网络转发表项中去往所述其他节点设备的出接口; [0064] The maintenance module entry, based on the large advance according Layer Network Layer Protocol Network generate large forwarding entry, the second MAC address of other nodes generating a local device address corresponding to the MAC forwarding entry, the said second MAC address destination MAC address forwarding table entry for the local MAC address of the other node device, a network interface of the large Layer forwarding entry device destined for the other nodes in the outgoing interface;

[0065] 所述报文转发模块还用于,当所述网络设备接收到目的MAC地址不是自身端口MAC 地址或所述虚MAC地址的单播数据报文后,根据所述单播数据报文的目的MAC地址及所属VLAN,在所述第二MAC地址转发表项中进行匹配,并通过匹配表项的出接口直接转发所述单播数据报文而不进行大二层网络协议封装。 [0065] The message forwarding module is further configured to, when the network device receives the destination MAC address is not its own MAC address or port of the virtual MAC address unicast data packets, according to the unicast data packets destination MAC address and the VLAN belongs, in the second MAC address forwarding matching entry, and forward the unicast packet data network without large Layer protocol encapsulation matching entry directly through the interface of the publication.

[0066] 其中,所述网络设备还与所述大二层网络外部的路由设备连接; [0066] wherein the network device is further connected to the external network routing Layer large equipment;

[0067] 所述大二层网络中各节点设备上设置有VLAN的广播域范围,其中,所述广播域范围包括该节点设备上允许该VLAN通过的用户侧端口和该节点设备上的位于该VLAN裁剪树上的大二层网络侧接口,所述裁剪树是基于VLAN对所述广播树进行分发树的剪枝得到的; [0067] The Network Layer is provided with a large broadcast domain VLAN range in each node device, wherein the broadcast domain on the range comprises allowing the VLAN specified by the user-side port of the node device and the node device large Layer 2 network side interface VLAN cut tree, the tree was cut based VLAN pruning tree for the broadcast distribution tree;

[0068] 所述报文转发模块还用于,在接收到所述路由设备发送的数据报文时,查找匹配所述数据报文的目的IP地址的ARP表项,并根据查找到的ARP表项的目的MAC地址以及出接口,直接转发所述数据报文而不进行大二层网络协议封装。 [0068] The message forwarding module is further configured to, upon receiving a data packet sent by the routing device, the ARP lookup entry matching the destination IP address of the data packet, and according to the search the ARP object of the MAC address and the interface to directly forward the data packet network without large Layer protocol encapsulation.

[0069] 其中,所述报文转发模块还用于,若未查找到匹配所述数据报文的目的IP地址的ARP表项,在该数据报文对应的VLAN的广播域内,直接广播发送用于请求所述数据报文的目的IP地址对应的MAC地址的第一ARP请求报文而不进行大二层网络协议封装,所述第一ARP 请求报文的发送端sender IP地址是所述网络设备的发送端口IP地址,源MAC地址及发送端MAC地址是所述网络设备的发送端口MAC地址; [0069] wherein the message forwarding module is further configured to, if not found ARP entry matching the destination IP address of the data packet, the data in the broadcast domain to the VLAN corresponding to the packet, a direct broadcast transmission the first object of the ARP request to the IP address of the data packet corresponding to the MAC address request message without a large network Layer protocol encapsulation, the first ARP request packet sender IP address is the network SENDER IP address of the device transmitting port, and the transmission source MAC address MAC address is the MAC address of the transmission port of the network device;

[0070] 所述表项维护模块还用于,当所述网络设备接收到根据接收到针对第一ARP请求报文的第一ARP应答报文时,建立所述数据报文的目的IP地址对应的ARP表项; [0070] the entry maintenance module is further configured to, when the network device receives a response packet according to the received first ARP ARP request for the first packet, the destination IP address establishing the data packet corresponding to ARP entries;

[0071] 所述报文转发模块还用于,根据所述表项维护模块建立的ARP表项,转发所述数据报文而不进行大二层网络协议封装。 [0071] The message forwarding module is further configured to maintain ARP entry module established according to the entry, not to forward the data packet Network Layer protocol encapsulation large.

[0072] 其中,同一VLAN中的一个网关节点被选举为该VLAN的主网关节点, [0072] wherein a gateway node in the same VLAN is elected as the VLAN primary gateway node,

[0073] 所述报文转发模块还用于,当接收到主机发送的用于请求网关节点的MAC地址的第二ARP请求报文,且确定所述网络设备为对应VLAN的主网关节点时,生成第二ARP应答报文,该第二ARP应答报文的目的MAC地址为所述主机的MAC地址,源MAC地址和发送端sender MAC地址均为所述网络设备的虚MAC地址,sender IP地址为所述网络设备的虚IP地址,以使所述主机学习并建立所述网络设备的虚IP地址与虚MAC地址之间的关联关系;根据所述ARP 应答报文的目的MAC地址,在第二MAC地址转发表项中进行匹配,并通过匹配表项的出接口直接转发所述ARP应答报文而不进行大二层网络协议封装。 [0073] The message forwarding module is further configured to, when receiving the second ARP request for the MAC address of the gateway node sent by the host request message, and determining the network device is a master gateway node corresponding to the VLAN, generating a second ARP response packet, the second ARP response packet to the destination MAC address of the host MAC address, source MAC address and the MAC address of both the end SENDER virtual MAC address of the network device, sender IP address the virtual IP address of the network device, so that the host and establish a relationship between learning virtual IP address and the virtual MAC address of the network device; ARP response packet according to the destination MAC address, in the two MAC forwarding matching entry, and the ARP forwarding matching entry directly through the interface of the response packet network without large Layer protocol encapsulation published.

[0074] 其中,所述网络设备还包括: [0074] wherein, the network device further comprises:

[0075] 选举模块,用于在所述网络设备所属的VLAN内广播携带有所述网络设备自身端口IP地址、端口MAC地址、虚IP及虚MAC地址的邻居发现及主网关节点选举协议报文;接收所述VLAN内其他网关节点发送的所述协议报文,选举端口IP地址和/或端口MAC地址满足预设规则的网关节点,作为所述主网关节点。 [0075] election means for VLAN within the network device belongs to its own neighbor port carries broadcast IP address, port MAC address, IP address and the virtual MAC address of the virtual network device discovery and the main gateway node election protocol packets ; the other gateway node receiving the transmission protocol packets within the VLAN, election port IP address and / or port MAC address preset rule satisfies the gateway node as the primary gateway node.

[0076] 其中,所述表项维护模块还用于,在非网关节点与所述网络设备之间的链路中断时,更新达到该非网关节点下所有主机的ARP表项的出接口;在非网关节点发生宕机down 时,删除该非网关节点下的所有主机的ARP表项; [0076] wherein the entry module is further configured to maintain, when the link between the gateway node and the non-network device interrupt, update the outgoing interface to reach all hosts ARP entry in the non-gateway node; in when a non-gateway node occurs downtime down, delete the ARP entries for all hosts in the non-gateway node;

[0077] 所述地址管理模块还用于,在所述网络设备发现外部路由器不可达时,禁用自身所有的三层接口,并通过所述广播树通告虚MC地址的撤销消息,以使得非网关节点设备更新所述虚MC地址对应的所述第一MC地址转发表项的出接口; [0077] The address management module is further configured to, when the network device discovery external routers unreachable, disable all three interfaces themselves, and notify the MC address withdraw message through the virtual broadcast tree, so that the non-gateway updating the node device corresponding to the virtual address of the first MC MC address forwarding entries of the interfaces;

[0078] 所述选举模块还用于,在所述网络设备发现外部路由器不可达时,退出本网关节点所在VLAN的主网关节点的选举,以触发其他网关节点重新选举主网关节点。 [0078] The election module is further configured, when the external network device discovery unreachable router, gateway node exit RO Ben articulation point is located VLAN to trigger the re-election of the master gateway node to other gateway nodes.

[0079] 本发明上述实施例中,网关节点通过一可达大二层网络中的所有节点设备的广播树,通告本节点设备配置的虚MC地址以及所属VLAN;非网关节点设备通过该广播树接收到各个网关节点通告的该网关节点配置的虚MAC地址以及所属VLAN后,根据大二层网络协议和接收到的同一虚MAC地址对应的各个网关节点的大二层网络标识,计算去往同一VLAN的各网关节点的单播路由,进而生成该虚MAC地址对应的第一MAC地址转发表项;当非网关节点接收到数据报文,且确定该数据报文的目的MAC地址为网关节点的虚MAC地址时,该非网关节点根据该数据报文的目的MAC地址,在第一MAC地址转发表中查询匹配的表项,并根据查询到的匹配表项,转发该数据报文而不进行大二层网络协议封装;网关节点接收到该数据报文,且确定该数据报文的目的MAC地址为自身虚MAC地址时,网关节点 The above embodiments [0079] In the present invention, all broadcast by the gateway node of a tree node devices up to large Layer 2 network, the virtual address and the VLAN configuration MC advertisement node equipment; non-gateway apparatus through the broadcast node tree arranged to each gateway node receiving the advertisement gateway node virtual MAC address belongs to the VLAN and a rear, large floor according to each gateway node network protocol and the same Layer large virtual MAC address corresponding to the received network identification, destined for the same computing unicast route to the gateway node the VLAN, thereby generating a first MAC address corresponding to the virtual MAC address forwarding entry; if a non-gateway node receiving the data packet, and determining that the data packet destination MAC address of gateway node when the virtual MAC address, the non-gateway node according to the destination MAC address of the data packet, the lookup table entries that match the first MAC address forwarding table, and matching the query entry to forward the data packet without large network Layer protocol encapsulation; gateway node receives the data packet and determine the destination MAC address of the data packet to its own virtual MAC address, the gateway node 根据该数据报文的目的IP地址,查找路由表转发该数据报文,实现了基于大二层网络的L3单播流量转发。 According to the destination IP address of the data packet, the routing table to forward the data packets to carry out the L3 unicast traffic forwarding based on a large layer 2 network.

附图说明 BRIEF DESCRIPTION

[0080] 图1为现有技术中一种采用代理转发解决L3转发和Trill处理不能共存问题的方式的示意图; [0080] FIG. 1 is a prior art schematic diagram of one employed L3 forwarding mode and coexistence issues Trill processing can not solve the forwarding agent;

[0081] 图2为本发明实施例提供的一种Trill组网中VLAN广播域的示意图; [0081] FIG. 2 Trill schematic diagram of a networking embodiment of VLAN broadcast domains provided in the embodiment of the present invention;

[0082] 图3为本发明实施例提供的一种Tr i 11组网的架构示意图; [0082] FIG. 3 Tr i 11 architecture diagram of a network according to an embodiment of the present invention;

[0083] 图4为本发明实施例提供的一种三层单播报文转发方法的流程示意图; [0083] FIG 4 provides a schematic flow diagram of a three unicast forwarding method embodiment of the invention;

[0084] 图5〜8为本发明实施例提供的Trill组网内故障场景示意图; [0084] FIG. 5 ~ 8 provided in the Trill network failure scenarios schematic embodiment of the present invention;

[0085] 图9为本发明实施例提供的一种网络设备的结构示意图; [0085] FIG. 9 is a schematic structure of a network device according to an embodiment of the present invention;

[0086] 图10为本发明实施例提供的另一种网络设备的结构示意图。 [0086] FIG. 10 is a schematic structural diagram of another network device according to an embodiment of the present invention.

具体实施方式 Detailed ways

[0087] 本发明实施例提供了一种报文转发的技术方案,应用于多个节点设备基于大二层网络协议组建的大二层网络中,该多个节点设备包括作为网关节点的网关节点和作为接入设备的非网关节点,同一VLAN的各网关节点都配置有该VLAN的三层(L3)接口,同一VLAN的各网关节点的三层接口具有互不相同的IP地址和MAC地址,即实IP地址和实MAC地址。 [0087] Example embodiments provide a technical solution of the present invention is forwarding packets, is applied to a large plurality of node devices based on the formation of a large network protocol Layer Layer 2 network, the node apparatus includes a plurality of gateway node gateway node and a non-gateway node as an access device, each of the same VLAN gateway node is configured with a VLAN of the three (L3) interfaces, three of each gateway node interfaces have the same VLAN mutually different IP address and MAC address, i.e. real IP address and the real MAC address. 本发明实施例中,针对同一VLAN的各网关节点的该三层接口,还配置相同的虚IP地址和虚MAC地址,即同一VLAN下的不同网关节点,具有相同的虚IP地址和虚MAC地址。 Embodiments of the present invention, for each of the three interfaces of the gateway nodes of the same VLAN, but also configure the same virtual IP address and the virtual MAC address, i.e. different nodes in the same VLAN gateway, with the same virtual IP address and the virtual MAC address .

[0088] 本发明实施例中,网关节点可以通过一可达大二层网络中的所有节点设备的广播树,通告本节点设备配置的虚MAC地址以及所属VLAN。 [0088] embodiment, the gateway node can reach all broadcast by a large tree nodes Layer 2 network devices, the configuration of a node device according to the present advertisement virtual MAC address and the VLAN embodiment of the present invention. 这里,通告的虚MAC地址是三层(L3)接口虚IP地址对应的虚MAC地址。 Here, the virtual MAC address is advertised Layer (L3) IP address of virtual interfaces corresponding to the virtual MAC address.

[0089] 非网关节点设备通过该广播树接收到各个网关节点通告的该网关节点配置的虚MAC地址以及所属VLAN后,可以根据大二层网络协议和接收到的同一虚MAC地址对应的各个网关节点的大二层网络标识,计算去往同一VLAN的各网关节点的单播路由,进而生成该虚MAC地址对应的MAC地址转发表项(以下称为第一MAC地址转发表项),其中,该第一MAC地址转发表项中的VLAN为该MAC地址所属VLAN,目的MAC地址为该虚MAC地址、出接口为该单播路由中代价最小的路由对应的出接口; [0089] Non gateway node device arranged to each gateway node is received by the broadcast advertisement gateway node of the tree, and the virtual MAC address belongs Through VLAN, in accordance with various network protocols and gateways large Layer same virtual MAC address corresponding to the received Layer network node identified large, the calculation of each unicast route to the gateway node of the same VLAN, thereby generating the virtual MAC address corresponding to the MAC address forwarding table entry (hereinafter, referred to as a first MAC address entry), wherein the first VLAN MAC address forwarding entry that belongs to VLAN MAC address, destination MAC address for the virtual MAC address, the interface for a unicast route cost of the route with the smallest corresponding to the interfaces;

[0090] 当非网关节点接收到数据报文,且确定该数据报文的目的MAC地址为网关节点的虚MAC地址时,该非网关节点可以根据该数据报文的目的MAC地址,在第一MAC地址转发表中查询匹配的表项,并根据查询到的匹配表项,直接转发该数据报文而不进行大二层网络协议封装; [0090] When a non-gateway node receives the data packet and determine the destination MAC address of the data packet is the MAC address of the virtual gateway node, the non-gateway node according to the destination MAC address may be the data packet, the first MAC address forwarding table entry matches the query, and the query matches the entry and forwards the packet data network without large Layer protocol encapsulation;

[0091] 网关节点接收到该数据报文,且确定该数据报文的目的MAC地址为自身虚MAC地址时,网关节点根据该数据报文的目的IP地址,查找路由表转发该数据报文。 When [0091] the gateway node receives the data packet and determine the destination MAC address of the data packet to its own virtual MAC address, the gateway node according to the destination IP address of the data packet, the routing table to forward the data packet.

[0092] 其中,当非网关节点根据大二层网络协议和接收到的同一虚MAC地址对应的各个网关节点的大二层网络标识,计算得到的去往同一VLAN的单播路由中存在多条代价(cost) 最小的路由时,则需要形成等价路由,相应地,该非网关节点生成的该虚MAC地址对应的第一MC地址转发表项中的出接口为该等价路由中各路由对应的出接口形成的聚合端口。 [0092] wherein, when a non-gateway node unicast routing Layer The large network gateway node identifiers of network protocol and the same Layer large virtual MAC address corresponding to the received, calculated in the same VLAN destined there are multiple when the cost (cost) minimum route cost of the route is required, and accordingly, the MC address of the first non-gateway node generates the virtual MAC address corresponding to the interface for forwarding the equivalent routing table entry in each of the routes corresponding to the aggregated ports formed in the interface.

[0093] 相应地,非网关节点在转发目的MAC地址为上述虚MAC地址的数据报文时,由于根据该虚MAC地址查找到的匹配的第一MAC地址转发表项的出接口为聚合端口,非网关节点可以通过hash (哈希)算法从该聚合端口中选择一个端口作为该数据报文转发的出接口。 [0093] Accordingly, a non-gateway node as the destination MAC address forwarding data packets of the virtual MAC address, since the found based on the virtual MAC address matches the MAC address of the first forwarding entry port of the interface polymerization, non gateway node may be selected from the port by the polymerization hash (hash) algorithm as a port to forward the data packet out of the interface.

[0094] 在本发明实施例中,各个节点设备还需要通过上述广播树,将新学习到的本地MAC 地址以及所属VLAN通告给其他节点设备,该本地MAC地址包括从本节点设备连接的用户网络侧学习到的MAC地址; [0094] In the embodiment of the present invention, each node device also requires via the multicast tree, the newly learned local MAC address and the VLAN advertisement to another node apparatus, the local MAC address includes a user network connection from the local node device side MAC address learning;

[0095] 网关节点可以从该广播树接收其他节点设备通过的该其他节点设备新学习到的本地MAC地址以及所属VLAN,并根据预先基于该大二层网络协议生成的大二层网络转发表项,生成该其他节点设备的本地MAC地址对应的MAC地址转发表项(以下称为第二MAC地址转发表项),该第二MAC地址转发表项中的目的MAC地址为上述其他节点设备的本地MAC地址, 出接口为该大二层网络转发表项中去往该其他节点的出接口。 [0095] The gateway node may receive other node apparatus newly learned MAC address to the local node, and the VLAN other devices from the multicast tree, and according to entries based on the previously generated large network protocol Layer Layer large network transpose , the other node device generates local MAC address corresponding to the MAC address forwarding entries (hereinafter referred to as the second MAC address entry), the second MAC address destination MAC address forwarding table entry for the another device local node MAC address, network interfaces for forwarding the large Layer items destined for other nodes in the interface.

[0096] 当网关节点设备接收到单播数据报文,且确定该单播数据报文不是自身端口MAC 地址或虚MAC地址时,该网关节点可以根据该单播数据报文的目的MAC地址及所属VLAN,在上述第二MAC地址转发表项中进行匹配,并通过匹配表项的出接口直接转发该单播数据报文而不进行大二层网络协议。 [0096] When the gateway node device receives a unicast data packet, and determining that the unicast data packet is not self-port MAC address or virtual MAC address, the gateway node according to the destination MAC address of the unicast data packets and belongs to the VLAN, MAC address in the second matching entry published, and forwards the unicast data packet network without large Layer protocol matching entry directly through the outgoing interface.

[0097] 进一步地,在本发明实施例中,上述网关节点还与大二层网络外部的路由设备连接,当网关节点接收到该路由设备发送的数据报文时,在数据报文的目的MAC地址为自身MAC地址时,检查数据报文的目的IP地址,若发现该目的IP地址在自己的直联网段,则可以根据该数据报文的目的IP地址查找匹配的ARP表项,并根据查找到的ARP表项的目的MAC地址以及出接口,直接转发该数据报文而不进行大二层网络协议封装。 [0097] Further, in the embodiment of the present invention, the gateway routing node further device connected to an external large layer 2 network, the gateway node when receiving a data packet transmitted from the routing device, the destination MAC address in the data message when the address is its own MAC address, check the data packet destination IP address, if found that the destination IP address in their direct network segment, you can find the ARP entry matching the destination IP address of the data packet, and according to the search the destination MAC address of the ARP entry, and interface directly forwards the data packet network without large Layer protocol encapsulation.

[0098] 其中,当网关节点根据接收到的数据报文的目的IP地址未查找到匹配该数据报文的目的IP地址的ARP表项时,该网关节点需要在该数据报文对应的VLAN的广播域内,直接广播发送用于请求该数据报文的目的IP地址对应的MAC地址的ARP请求报文(以下称为第一ARP请求报文)而不进行大二层网络封装,该第一ARP请求报文的sender (发送端)IP地址为该网关节点的三层发送端口的实IP地址,源MAC地址及sender MAC地址为该网关节点的三层发送端口的实MAC地址。 [0098] wherein, when the gateway node according to the received data packet destination IP address is not found to match the ARP entry for the data packet destination IP address, the gateway node needs in the data packets corresponding to the VLAN broadcast domain, a direct broadcast ARP transmission destination IP address of the data packet corresponding to the MAC address request message for requesting (hereinafter referred to as a first ARP request packet) Layer 2 network without large package, the first ARP SENDER request message (transmission side) transmits the IP address of the gateway port for the three-node real IP address, source MAC address and the MAC address transmitted SENDER real MAC address for the three-port gateway node.

[0099] 当该网关节点接收到针对第一ARP请求报文的ARP应答报文(以下称为第一ARP应答报文)时,该网关节点可以根据该第一ARP应答报文建立该数据报文的目的IP地址对应的ARP表项,并根据该建立的ARP表项,直接转发该数据报文而不进行大二层网络协议封装。 [0099] When the first gateway node is received for the ARP request packet is ARP response packet (hereinafter referred to as a first ARP response packet), the gateway node may establish a response message of the data packet according to the first ARP packet destination IP address corresponding ARP entry, and in accordance with the established ARP entry, directly forwards the data packet network without large Layer protocol encapsulation.

[0100] 其中,网关节点在VLAN的广播域内广播ARP请求报文时,需要根据预先设置的VLAN 的广播域范围进行广播。 [0100] wherein the gateway node broadcasts an ARP request packet in the broadcast domain VLAN, broadcast domain VLAN range of the broadcast set in advance.

[0101] 具体的,在本发明实施例中,可以在各个节点设备上设置VLAN的广播域范围,该广播域范围包括该节点设备上允许该VLAN通过的用户侧端口和该节点设备上的位置该VLAN 裁剪树上的大二层网络侧接口,该裁剪树是基于VLAN对上述广播树进行分发树的剪枝得到的。 [0101] Specifically, in the embodiment of the present invention may be provided broadcast domain VLAN range on the respective node devices, the broadcast domain range includes a position allows the VLAN specified by the node device user-side port and the node apparatus Layer large network side interface of the VLAN cut tree, cut the tree is based on the above-described VLAN pruning tree broadcast distribution tree obtained.

[0102] 进一步地,在本发明实施例中,同一VLAN中的各网关节点还需要进行主网关节点的选举,以选举一个网关节点作为该VLAN的主网关节点;当网关节点接收到主机发送的用于请求网关节点的MAC地址的ARP请求报文(以下称为第二ARP请求报文)时,只有主网关节点进行响应。 [0102] Further, in the embodiments of the invention, each of the gateway nodes in the same VLAN also need RO gateway node, in order to elect a gateway node as the primary gateway node that VLAN; When the gateway node receives transmitted by the host an ARP request for the MAC address of the gateway node of the request packet (hereinafter, referred to as second ARP request message), only the master node responds to the gateway.

[0103] 具体的,在本发明实施例中,同一VLAN中的各个网关节点可以在该VLAN内广播携带有自身三层端口的实IP地址、三层端口的实MAC地址、虚IP地址及虚MAC地址的邻居发现及主网关节点选举协议报文;当同一VLAN中的各个网关节点接收到其他网关节点发送的上述协议报文,选举端口实IP地址和/或端口实MAC地址满足预设规则的网关节点,作为主网关节点。 [0103] Specifically, in the embodiment of the present invention, each of the gateway node may broadcast the same VLAN carried in the VLAN has its own real IP address of the three-port, three-port real MAC address, and the virtual address of the virtual IP neighbor discovery and MAC address of primary gateway node election protocol packets; when the respective gateway nodes in the same VLAN receiving the protocol message sent by the node to other gateway, a real election port IP address and / or port to meet the real MAC address preset rule the gateway node, as the primary gateway node. 例如,可以选举端口IP地址+端口MAC地址最大的网关节点作为主网关节点,或者, 选举端口IP地址或端口MC地址最大的网关节点作为主网关节点。 For example, IP address + port election port MAC address of the gateway node as the largest primary gateway node, or port election MC IP address or port address of the gateway node as the largest primary gateway node.

[0104] 当网关节点接收到主机发送的用于请求网关的MAC地址的第二ARP请求报文时,网关节点需要确定自身是否为对应VLAN的主网关节点,若确定自身为对应VLAN的主网关节点,则生成ARP应答报文(以下称为第二ARP应答报文),该第二ARP应答报文的目的MAC地址为该主机的MAC地址,源MAC地址和sender MAC地址均为该网关节点的虚MAC地址,sender IP地址为所述网关节点的虚IP地址,以使主机学习并建立网关节点的虚IP地址与虚MAC地址之间的关联关系;若确定自身不是对应VLAN的主网关节点,则该网关节点不对该ARP应答报文进行响应。 [0104] When the second gateway node receives ARP request for the MAC address of the gateway host sends a request message to the gateway node needs to determine whether it is a master gateway node corresponding to the VLAN, to determine if the VLAN itself as primary gateway node, is generated ARP response message (hereinafter, referred to as second ARP response packet), the second ARP response packet for the destination MAC address of the host's MAC address, source MAC address and the MAC address are the SENDER gateway node the virtual MAC address, the IP address of the virtual SENDER IP address of the gateway node, so that the host and establish learning the relationships between the virtual IP address of the gateway node virtual MAC address; determining if the gateway node itself is not the master of the VLAN , the gateway node does not answer the ARP packet response.

[0105] 其中,作为主网关节点的网关节点生成ARP应答报文后,根据该ARP应答报文的目的MAC地址,在第二MAC地址转发表项中进行匹配,并通过匹配表项的出接口直接转发该ARP 应答报文而不进行大二层网络协议封装。 After [0105] wherein, as a gateway node of the primary gateway node generates an ARP response packet according to the destination MAC address in the ARP response packet, the second MAC address forwarding matching entry, and the entry by matching interfaces directly forward the ARP reply without large Layer 2 network protocol encapsulation.

[0106] 需要注意的是,通过以上描述可以看出,在本发明实施例提供的技术方案中,非网关节点向网关节点转发数据报文、大二层网络节点设备对ARP请求报文、ARP应答报文的发送/转发,采取的均为不进行大二层封装,而直接查询匹配MAC地址转发表项的方式进行, 因此,为了使本领域技术人员更好地理解本发明实施例提供的技术方案,下面对大二层网络中二层单播/广播流量转发的技术方案进行详细描述。 [0106] It should be noted, can be seen from the above description, the technical solutions provided in embodiments of the present invention, the non-gateway node forwards the data packet to the gateway node, the network node apparatus Layer large ARP request packets, ARP response packet send / forward, are taken Layer 2 encapsulation is not large, but query matches the MAC address forwarding entry way, and therefore, in order to enable those skilled in the art better understand the embodiment of the present invention provides aspect, following large Layer network Layer unicast / broadcast traffic forwarded to the technical solutions described in detail.

[0107] 在大二层网络的控制平面,各个节点设备通过一可达大二层网络中所有节点设备的广播树,将新学习到的本地MAC地址以及所属VLAN通告给其他节点设备(其中,需要注意的是,在本发明实施例中,对于作为网关节点的节点设备,其通过上述广播树通告给其他节点设备的MAC地址为该网关节点配置的虚MAC地址);节点设备从该广播树接收其他节点设备通告的其他节点设备的本地MAC地址,并通过预先基于该大二层网络协议生成的大二层网络转发表项,生成原始MAC地址转发表项(在本发明实施例中,为第一MAC地址转发表项和第二MAC地址转发表项);在大二层网络的数据平面,节点设备接收到单播数据报文后,若该单播数据报文的目的MAC地址不是自身的MAC地址,则根据该单播数据报文的目的MAC地址及所属VLAN,在该原始MAC地址转发表项中进行匹配,并通过匹 [0107] In the control plane large layer 2 network, each node device via a Layer up large tree nodes all broadcast network devices, a new learned the MAC address of the local device and the other node belongs to a VLAN advertisement (wherein, Note that, in the embodiment of the present invention, the device for the gateway node as the node that advertised via the broadcast tree to the virtual MAC addresses of other node devices configured for the gateway node); node apparatus from the broadcast tree local MAC address of other node apparatus receiving apparatus advertised to other nodes, and forwarding entries in advance based on the large-Layer network Layer protocol generated large network, generating an original MAC address forwarding entries (in the embodiment of the present invention, is a first MAC address forwarding entries and the second MAC address entry); after receiving unicast data packets in the data plane, a large layer 2 network node device, if the destination MAC address of the unicast data packet is not itself MAC address, according to the destination MAC address and the VLAN unicast data packets, the original MAC forwarding matching entry, and by matching 表项的出接口直接转发该单播数据报文而不进行大二层网络协议封装。 An entry interfaces directly forward the unicast data packet network without large Layer protocol encapsulation.

[0108] 具体的,当一节点设备接收到其他节点设备通告的该其他节点设备新学习到的第一MAC地址以及所属的第一VLAN时,根据该其他节点设备的大二层网络标识,查询预先基于大二层网络协议生成的大二层网络转发表,确定该其他节点设备对应的出接口,进而生成该第一MAC地址对应的原始MAC地址转发表项。 [0108] Specifically, when a new learned the MAC address of a first node device receiving the other node device and the other device nodes of the first VLAN advertisement belongs, according to the another node device Layer large network identifier, the query pre-forwarding network protocol based on a large Layer Layer generated large network, determines that the other node device corresponding to the interfaces, thereby generating an original MAC address of the first MAC address forwarding entry.

[0109] 其中,节点设备确定其他节点设备对应的出接口具体可以通过以下方式实现: [0109] wherein, the node device determines another node device corresponding to the specific interface may be implemented in the following ways:

[0110] 节点设备可以从大二层网络转发表中,查找到该其他节点设备的大二层网络标识对应的转发表项;若查找的转发表项中仅包括一个出接口,则将查找到的出接口,作为其他节点设备对应的出接口;若查找到的转发表项中包括两个以上的出接口,则将该两个以上的出接口形成一个聚合端口,并将该聚合端口作为其他节点设备对应的出接口。 [0110] Layer node device from a large network forwarding table, find the other large Layer 2 network node device identifier corresponding forwarding entry; if it can find a forwarding entry comprises only one exit port, will find outbound interface, as another node device corresponding to the outgoing interface; if the item comprises two or more interfaces to find the forwarding table, then the two or more interfaces to form a polymeric a port and the other port as a polymerization node device corresponding to the interface. 其中,上述转发表项中包括的两个以上的出接口可以是以等价多路径的形式存在。 Wherein said forwarding entry includes the two or more interfaces may be present in the form of ECMP.

[0111] 相应地,当节点设备需要通过匹配表项的出接口转发单播数据报文时,若匹配表项的出接口为聚合端口,则可以通过HASH (哈希)算法选择该聚合端口的一个成员端口,并通过所选择的成员端口,直接转发该单播数据报文而不进行大二层网络协议封装。 [0111] Accordingly, when the node devices need to forward unicast data packet, the matching entry if the matching entry outgoing interface port interface polymerization, the polymerization may be selected by HASH (hash) algorithm port a port member, the port member and selected by directly forwarding the unicast data packet network without large Layer protocol encapsulation.

[0112] 其中,需要注意的是,在本发明实施例中,由于网关节点通过广播树通告的是虚MAC地址,且同一VLAN内各网关节点配置的虚MAC地址相同,因此,非网关节点设备在生成该虚MAC地址对应的MAC地址转发表项时,需要根据大二层网络协议和接收到的同一虚MAC地址对应的各个网关节点的大二层网络标识,计算去往同一VLAN的各网关节点的单播路由, 进而生成所述虚MC地址对应的第一MC地址转发表项。 [0112] wherein, be noted that, in the embodiment of the present invention, since a gateway node through a broadcast tree advertised the virtual MAC address, and the same identical configuration of each gateway node the VLAN virtual MAC address, therefore, a non-gateway node apparatus in generating the virtual MAC address corresponding to the MAC address forwarding entries, each need to large Layer Layer gateway node network protocols and large the same virtual MAC address corresponding to the received network identification, the gateway Flights calculated for each of the same VLAN unicast routing node, thereby generating a first virtual address corresponding to the MC MC address forwarding entry.

[0113] 进一步地,在大二层网络中二层单播/广播流量转发的技术方案中,当节点设备的本地MAC地址老化时,节点设备也需要将发生老化的本地MAC地址以及所属VLAN通告给其他节点设备,由其他节点设备删除发生老化的MAC地址以及所属VLAN所对应的转发表项。 [0113] Further, in a large network Layer Layer unicast / broadcast traffic forwarded aspect, when the local MAC address aging node device, the node device also needs to local MAC address and the VLAN advertisement aging occurs to other node devices by MAC address aging equipment other node deletion occurs and the VLAN corresponding forwarding entry.

[0114] 具体的,各个节点设备可以通过广播树,将发生老化的本地MAC地址以及所属VLAN 通告给其他节点设备;当节点设备从广播树上接收到其他节点设备处发生老化的MAC地址(第二MAC地址)以及所属VLAN时,从原始MAC地址转发表中,删除该第二MAC地址以及所属VLAN所对应的转发表项。 [0114] Specifically, each device may broadcast the node tree, the aged and the local MAC address belongs to another VLAN announcing the occurrence of a node device; occurs when the MAC address aging tree node from the broadcast receiving apparatus to the other apparatus at a node (p two MAC address) and belongs to a VLAN, MAC address forwarding table from the original, delete the second address and the MAC forwarding entry corresponding to the VLAN.

[0115] 本发明实施例中,节点设备在MAC地址老化时,可以通过通告报文在节点之间进行MAC地址同步。 Embodiment [0115] of the present invention, when the node device MAC address aging, a MAC address can be synchronized between nodes notification packet. 实际应用中有可能出现MAC地址假老化的情形。 Practical applications possible scenarios fake MAC address aging appear. 所谓MAC地址假老化,是指终端数据发送报文后,间隔一段较长时间(大于MAC地址老化时间)而不再发送报文,然后再次发送报文,反复如此,导致该终端的MAC地址在节点设备的端口上被学习、老化、再学习、再老化……,其中每一次老化或学习过程都会引发通告报文的发送。 The so-called false MAC address aging, refers to the terminal data packets sent, a longer interval of time (greater than the MAC address aging time) no longer send packets, and then transmits the message again, so repeatedly, causing the MAC address of the terminal port device nodes are learning, aging, re-learning, aging ...... again, each time learning process will lead to aging or advertisement packet sent.

[0116] 为了降低MAC地址频繁假老化所引发的通告报文发送数量,减轻节点设备的CPU处理负担,本实施例中各个节点设备可以针对自身不同的接口与VLAN的组合,分别配置一对应的MAC地址池,用于维护在该对应接口上学习到的归属于对应VLAN的本地MAC地址; [0116] number of sent notification packet in order to reduce the MAC address frequent false aging caused, reduce the CPU processing load node apparatus, for example, each node apparatus of the present embodiment may be directed to a combination of their different interfaces and the VLAN, are arranged a corresponding MAC address pool for maintaining learned in the interface corresponding to the local MAC address assigned to the VLAN;

[0117] 各个节点设备在新学习到本地MAC地址,例如在MAC转发表项中记录新学习到的MAC地址并建立对应的老化定时器后,若对应MAC地址池中未包含有该新学习到的本地MAC 地址时,则将新学习到的本地MAC地址增加到MAC地址池中。 [0117] Each node in the new device to the local MAC address learning, for example, to record a new MAC forwarding entry learned MAC address is established, and a corresponding aging timer, if the MAC address pool does not correspond with the new learning to leave when the local MAC address, then the new learned local MAC address MAC address added to the pool. 在本地MAC地址发生老化的一段预设时间后,才将老化的本地MAC地址从对应MAC地址池中删除,也就是说,在MAC转发表项中的MAC地址因老化定时器超时而老化后,再过一段预设时间,才将该MAC地址从对应MC地址池中删除。 After a preset time the local MAC address aging occurs, it will be the aging of the local MAC addresses from the MAC address pool to delete, that is to say, after the MAC address in the MAC forwarding entry due to aging and the aging timer expires published, after a period of predetermined time, it deletes the corresponding MAC address of the MC address from the pool.

[0118] 相应的,各个节点设备,进一步在MAC地址池中增加了新学习到的本地MAC地址时, 通过所述广播树,将新学习到的本地MAC地址以及所属VLAN通告给其他节点设备;以及,在MAC地址池中删除了老化的本地MAC地址时,通过所述广播树,将老化的本地MAC地址以及所属VLAN通告给其他节点设备。 When [0118] Accordingly, each node device, further increasing the learned new local MAC address in the MAC address pool, by the broadcast tree, the newly learned MAC addresses and other local node device belongs to a VLAN advertisement; and, when the MAC address pool local MAC address aging deleted by the broadcast tree, local MAC addresses and the VLAN aging advertised to other device nodes. 也就是说,MAC转发表项中的MAC地址老化并不会直接触发通告报文的发送,而是等待该MAC地址从MAC地址池中删除后,才去发送对应的通告报文。 In other words, MAC forwarding MAC address entry aging does not directly trigger notification message is sent, but wait for the MAC address from the MAC address pool to delete, before going to send corresponding notification packets.

[0119] 与之相对应的,在本发明实施例中,非网关节点也可以按照上述大二层网络中二层单播/广播流量转发的技术方案中的MAC地址老化方式对本地MC地址进行老化处理。 [0119] Correspondingly, in the embodiment of the present invention, a non-gateway nodes may also be carried out according to the MC address for the local MAC address of the Network Layer Layer large unicast / broadcast traffic forwarded aspect aging manner aging treatment.

[0120] 对于广播报文,在大二层网络中二层单播/广播流量转发的技术方案中,可以采用按照芯片vlan广播域范围(bitmap)直接广播的方式进行转发。 [0120] For broadcast packets, in a large network Layer Layer unicast / broadcast traffic forwarded aspect may be employed in accordance with the manner chip vlan broadcast domain range (Bitmap) direct broadcast forwarding.

[0121] 具体的,可以在各节点设备上设置VLAN的广播域范围,其中,该广播域范围包括该节点设备上允许该VLAN通过的用户侧端口和该节点设备上的位于该VLAN裁剪树上的大二层网络侧接口,该VLAN裁剪树是基于VLAN对该广播树进行分发树的剪枝得到的; [0121] Specifically, the broadcast domain VLAN range may be provided at each node device, wherein the range includes the broadcast domain VLAN cut on the tree allows the VLAN specified by the user-side port of the node device and the node device Layer large network-side interface, the tree is pruned VLAN crop obtained based VLAN distribution tree of the multicast tree;

[0122] 节点设备接收到待广播数据报文后,确定该待广播数据报文所属VLAN,并在该VLAN的广播域范围所包括的除接收端口外的其他端口上,直接转发该待广播数据报文而不进行大二层网络协议封装,其中,该待广播数据报文包括未知单播报文、广播数据报文和未知组播报文。 After the [0122] node device receives a to be broadcast data packet, determines that the to be broadcast packet belongs VLAN, and other ports except the receiving port broadcast domain range of the VLAN included, directly forwarding the pending broadcast data Layer packet network protocol rather large package, wherein the message comprises data to be broadcast or unknown unicast, broadcast data packets and unknown multicast packets.

[0123] 与之相对应的,在本发明实施例中,对于广播报文的处理,可以采用上述大二层网络中二层单播/广播流量转发的技术方案中进行广播流量转发方式对本发明实施例中的广播报文进行转发。 [0123] Correspondingly, in the embodiment of the present invention, for processing a broadcast packet, broadcast traffic forwarded manner in the large-Layer Network Layer unicast / broadcast traffic forwarded aspect of the present invention may be employed forwarding broadcast packets in the embodiment.

[0124]以数据中心内的Trill技术为例,在本发明实施例中,可以根据trill域的vlan裁剪树来设置硬件芯片中各个vlan的广播域范围,用户侧端口根据端口上允许通过的vlan来对应设置;对于trill 口,不能根据允许通过的vlan来设置,而是需要根据trill计算出来的vlan裁剪树来对应设置(需要去掉设备上端口允许通过的vlan和该vlan的广播bitmap内包含该端口的一致性检查)。 [0124] In Trill technology within the data center, for example, in the embodiment of the present invention, a domain-wide broadcast may be provided in various hardware chips cut tree vlan vlan according trill domain, according to user-side port on the port allows vlan to the corresponding set; trill for opening, can be set according to the vlan allow, but need to be provided corresponding (ports need to be removed to allow the device within the broadcast and the Bitmap vlan vlan by comprising the trill calculated in accordance with the tree crop vlan consistency check port).

[0125] 对于不同的vlan,可以使用不同的vlan裁剪树来实现广播流量的负载分担。 [0125] For different vlan, load balancing may be implemented using different broadcast traffic vlan cutting trees. 如图2 所示,实线部分的链路为vlanl的广播域,按照此链路拓扑,在nickname-1上,设置vlanl的bi tmap包括用户侧端口P5 (ni ckname-1到该VLAN内下挂主机的端口)和上行口PI-P4 (分别^nickname-l|?IJnickname-gl~nickname-g4|3tl^M □) ; iS:nickname-gl Ji, T^Svlanl |3tl bitmap 仅包括Pl、P4 和Pn (分别为nickname-gl 到nickname-1、nickname-4、nickname_n 的端口),在nickname-4上仅包括用户侧端口P5 (nickname-4到该VLAN内下挂主机的端口)及上行口Pl (nickname-4到nickname-gl 的端口);在nickname-n上,仅包括用户侧端口P5 (nickname-n到该VLAN内下挂主机的端口)和上行口Pl (nickname-n到nickname-gl的端口)。 2, the solid line is the link vlanl broadcast domains, follow this link topology in the nickname-1, bi tmap vlanl set comprises a user side port P5 (ni ckname-1 to lower the VLAN port linked host) and uplink port PI-P4 (respectively ^ nickname-l | IJnickname-gl ~ nickname-g4 | 3tl ^ M □); iS:? nickname-gl Ji, T ^ Svlanl | 3tl bitmap includes only Pl, P4 and Pn (respectively nickname-gl to the nickname-1, nickname-4, nickname_n port) in the nickname-4 comprising only (ports nickname-4 to the VLAN linked under host) user-side port P5 and the upstream port Pl (nickname-4 to the nickname-gl port); in the nickname-n, including only user-side port P5 (nickname-n to the VLAN linked under host port) and an uplink port Pl (nickname-n to the nickname -gl port). 对于vlan2在各RB上的bitmap的设置原理同vlanl (虚线部分的链路),在此不再重复描述。 For vlan2 principle bitmap is provided on each RB same vlanl (dotted line portion of the link), which is not repeatedly described. 对于从ni ckname-1用户侧端口接收到的报文需要在V Ian 1内广播的时候,只需要按照相应的vlan bitmap进行广播即可,该广播报文可以沿着实线部分的链路,到达所有vlanl相关的RB,且不会产生环路。 For received ni ckname-1 from the user-side port when the packets need to broadcast within 1 V Ian, only needs to be broadcast according to the corresponding vlan bitmap, the broadcast packet may be along the solid line of the link, reaches All vlanl related to RB, and does not produce loops.

[0126] 下面以数据中心内的Trill技术为例,对本发明实施例所提供的报文转发技术方案进行详细描述,相关原理同样适用于其他的大二层组网技术,如Vxlan (virtual Extended VLAN,虚拟扩展VLAN(Virtual Local Area Network,虚拟局域网),IETF的draft (草案))、SPB技术等。 [0126] In the following Trill technology within the data center as an example, the present invention will be described in detail with packet forwarding technical solutions provided in the embodiments, the relevant principles are equally applicable to other networking technologies Layer large as Vxlan (virtual Extended VLAN , a virtual extension of VLAN (virtual Local Area Network, a virtual local area network), IETF's draft (draft)), SPB technology.

[0127]采用Trill技术构建的数据中心大二层网络,通常分为核心层、接入层。 [0127] The Network Layer Trill large data center technology to build, usually divided into the core layer, the access layer. 通常,运行Tri 11协议的设备称为RBridge (即Routing Bridge,具有路由转发特性的网桥设备,也称为路由桥,以下简称RB),RB构成Tri11域(S卩Tri11网络)。 Typically, the device is called a protocol run RBridge Tri 11 (i.e. Routing Bridge, bridge device having a routing forwarding feature, also called routing bridge, hereinafter referred to as RB), RB configuration Tri11 domain (S Jie Tri11 network). RB通过运行链路状态协议认知Tri11 域的拓扑,并使用最短路径树算法生成从该RB到达Tr i 11网络内的其它各个RB的路由转发表(称之为Trill路由表)。 RB topology state protocol link running Tri11 cognitive domains, using shortest path tree algorithm generates a route to each other within the RB Tr i 11 from the network forwarding RB (Trill called routing table).

[0128] 在本发明实施例中,Tr ill域中各RB使用Tr ill控制协议进行nickname路由计算时,生成的nickname单播路由表和组播表仅在软件上保留,而不下发硬件。 [0128] In an embodiment of the present invention, each RB Tr ill Tr ill domain using control protocol performs route calculation nickname, nickname generated unicast routing table and multicast table remains only in the software, rather than hardware issued.

[0129] 相应地,由于Tri 11域中各RB的硬件转发表中没有nickname转发表项,因此,当Trill域中各RB使用ESADI (End Station Address Information,用于终端MAC地址在各个RB之间主动同步的协议)协议报文对各RB的本地MAC地址进行相互通告和同步时,ESADI协议报文需要被送往CPU进行软件转发和处理。 [0129] Accordingly, since the RB field of Tri 11 no nickname hardware forwarding table forwarding entries, therefore, when each RB used ESADI (End Station Address Information, the MAC address for the terminal in the domain between each RB Trill, when the active synchronization protocol) protocol packets to each other and to synchronize the local advertisement MAC addresses of the RB, ESADI protocol packets to be sent to the CPU for software processing and forwarding. 其中,可以在软件上预先生成一棵Tri11的广播树,不做任何裁剪,所有节点(Tri 11域内各RB)可达,专门用来转发ESADI协议报文。 Among them, the software can be pre-generated broadcast on a tree Tri11, do not do any cutting, all nodes (each RB Tri 11 domain) up, dedicated to forwarding ESADI protocol packets.

[0130]需要注意的是,为了防止由于MAC地址自动学习导致的转发环路,在本发明实施例中,Trill域内所有用户侧端口需要禁止MAC自动学习(禁止硬件学习),改为软件学习或者手工配置或者协议自动生成的方式;Trill 口(hybrid(混杂)口)禁止学习(禁止软件、硬件学习),改为手工配置或者协议自动生成的方式。 [0130] It is noted that, in order to prevent the MAC address auto-learning forwarding loops caused in the embodiment of the invention, all user-side ports to be inhibited Trill art MAC automatic learning (learning prohibition hardware), the software to learn, or manual configuration or protocol mode automatically generated; Trill, port (Hybrid (Hybrid) port) prohibits learning (inhibiting software, hardware learning), to manual configuration or protocol mode automatically generated. 此外,使用ESADI协议同步MAC地址时,所有节点(包括传输节点(即transit节点))都需要处理ESADI报文。 Further, when using the synchronization protocol ESADI MAC address, all nodes (including the transit node (i.e., transit node)) need to be addressed ESADI packets.

[0131] 同时,Trill域内配置为网关节点的RB节点还需要运行EGRP(Enhanced Gateway Redundancy Protocol,增强网关备份协议),对于同一VLAN的多个网关节点,除了需要配置各不相同的实IP地址/实MAC地址之外,还需要配置相同的虚IP地址/虚MAC地址(即同一VLAN的各网关节点的实IP地址/实MAC地址不同,虚IP地址/虚MAC地址相同)。 [0131] Meanwhile, Trill arranged within an RB node to the gateway node needs to run EGRP (Enhanced Gateway Redundancy Protocol, enhanced backup Gateway Protocol), for a plurality of gateway nodes in the same VLAN, in addition to configurations vary the real IP address / outside the real MAC address, you configure the same IP address of the virtual / virtual MAC address (i.e., different embodiments of the IP address of each gateway node same VLAN / real MAC address, the same IP address of the virtual / virtual MAC address).

[0132] 启用EGRP协议后,配置为网关节点的RB节点在VLAN内发送EGRP协议报文,该EGRP 协议报文可采用VLAN内广播发送的方式,进行邻居发现和master选举。 After the [0132] agreement EGRP enabled, configured as a gateway node RB node sends in a VLAN EGRP protocol packets, which EGRP protocol message can be sent in broadcast mode VLAN, neighbor discovery and master election.

[0133] 其中,各网关节点利用EGRP协议进行邻居发现和master选举可以具体为: [0133] wherein each gateway node uses EGRP neighbor discovery protocol and master election may be:

[0134] 各网关节点在VLAN内发送EGRP协议报文,该EGRP协议报文中可以携带有各自设备标识信息、端口IP地址、端口MAC地址、虚IP及虚MAC地址;该VLAN内其他网关节点接收到该EGRP协议报文后,将报文中的端口IP地址和/或端口MAC地址与自身的端口IP地址和/或端口MAC地址作比较,通过比较,端口IP地址+端口MAC地址最大(或端口IP地址或端口MAC地址最大)的网关节点将成为master,其余网关节点则作为slave。 [0134] Each gateway node sends in a VLAN EGRP protocol packets, which EGRP protocol message may carry their device identification information, port IP address and port MAC address, the virtual IP and virtual MAC address; other gateway node in the VLAN after receiving the EGRP protocol message, the port IP address of the packets and / or port MAC address is compared with its own port IP address and / or port MAC address, the maximum comparison, port IP address + port MAC address ( IP address or port or ports highest MAC address) of the gateway node will become the master, the remaining nodes are the gateway as slave.

[0135] 其中,该EGRP协议报文可以周期性发送,且发送周期根据实际场景设置。 [0135] wherein the EGRP packets may be periodically transmitted, and the transmission cycle provided in accordance with the actual scene. 当slave 在一定时间内(如两倍的EGRP发送周期)未接收到master发送的EGRP协议报文(可以根据EGRP协议报文中携带的设备标识确定接收到的EGRP协议报文是否为master发送的),则认为作为master的网关节点down,需要重新在剩下的网关节点之间进行邻居发现和新master 的选举。 When the slave within a certain time (e.g., twice EGRP transmission cycle) is not received EGRP protocol packet sent by the master (which may be determined in accordance with the received device identifier carried in the protocol message EGRP EGRP whether protocol packets sent by the master ) is considered as a gateway node down master, the need to re-neighbor discovery and new master election between remaining gateway node.

[0136] 其中,对于同一VLAN的多个网关节点,在接收到host (主机)发送的ARP (Address Resolution Protocol,地址解析协议)请求报文时,仅有被选举为master的网关节点需要响应应答报文,且该应答报文为非Tri 11封装的ARP报文,且该ARP报文的外层MAC地址和sender MAC地址均为网关节点的虚MAC地址。 [0136] wherein, for a plurality of gateway nodes in the same VLAN, upon receiving the Host (host) transmitted ARP (Address Resolution Protocol, ARP) request packet, only the gateway node is elected as the master requires a response acknowledgment packet and the reply packet Tri 11 non-encapsulated ARP packets, ARP packets and the outer SENDER MAC address and MAC address are virtual MAC address of the gateway node. 而对于网关节点请求host的ARP的处理流程, 由网关节点使用自身的实MAC地址作为ARP请求报文的源MAC地址,而不使用虚MAC地址。 For gateway node ARP request processing flow of the host, the gateway node itself using the real MAC address as source MAC address in the ARP request message, without using the virtual MAC address.

[0137] 对于MAC地址的扩散和同步,在本发明实施例中,对于配置为网关节点的RB设备, 在通过ESADI协议同步MAC地址时,可以配置为同步虚MAC地址,即将自身的虚MAC地址携带在ESADI协议报文中进行扩散和同步。 [0137] For the diffusion MAC address and synchronization, in the embodiment of the present invention, for the configuration of RB device gateway node, when passing through ESADI protocol synchronization MAC address, it can be configured to synchronize the virtual MAC address, i.e. its virtual MAC address carry diffusion and synchronization ESADI protocol packet.

[0138] Tr i 11域内各RB收到ESADI协议报文后,提取其中携带的MAC地址、VLAN信息以及所属的ni ckname信息,并根据该ni ckname查询对应的ni ckname路由表,将该ni ckname路由表转换为MAC转发表下发转发平面。 After [0138] each RB Tr i 11 receives ESADI art protocol message, which carries the extracted MAC address, VLAN information, and ni ckname information belongs, and based on the corresponding query ni ckname ni ckname routing table, the ni ckname It is converted to the routing table issued MAC forwarding table forwarding plane.

[0139] 具体的,可以用ESADI协议报文中携带的MAC+VLAN替换对应的nickname路由表项中的ni ckname,若该ni ckname路由表项的出端口为单个出端口,则对应的MAC表项中出端口直接使用该nickname路由表项的出端口;若该nickname路由表项有等价路由,则对应的MAC 表项中的出端口可以为该nickname路由表项中各端口(各等价路由对应的端口)对应的聚合端口。 [0139] Specifically, can MAC + ni ckname nickname routing table entry a VLAN substituted for the corresponding ESADI protocol packet carries, if the port of the ni ckname routing table entry for a single output port, the corresponding MAC an entry port nickname used directly in the routing table entry port; nickname if the routing table entries have equivalent routes, the corresponding MAC table entries, the port may nickname for each routing table entry ports (each equivalent corresponding to the route port) corresponding to the aggregated ports.

[0140] 需要注意的是,由于同一VLAN的网关节点具有相同的虚IP地址/虚MAC地址相同, 对于非网关节点,为了生成到网关节点的虚MC地址的等价路径,非网关节点需要分别计算本节点到该虚MAC地址对应的各网关节点的单播路由,并将该单播路由中cost最小的路径在本节点的出端口作为到该虚MAC地址的出端口(即该虚MAC地址对应的MAC转发表项中的出端口);若存在多条cost最小的路径,就选取多条这种路径,把这些路径在本节点的出端口一起列出来,如果共享一个出端口,网关mac对应的出端口就是一个单独的物理端口;如果有多个,则将这几个物理端口聚合后作为该虚MAC地址的出端口。 [0140] It is noted that, due to the same VLAN gateway node having the same address in the same virtual IP / virtual MAC address for the non-gateway node, in order to generate a virtual path equivalent to the MC address of the gateway node, a non-gateway nodes need to separately each of the local node to calculate the virtual MAC address of the gateway node corresponding to the unicast routing, and the routing unicast minimum cost path as a local node interface to the port of the virtual MAC address (i.e., the virtual MAC address the corresponding MAC forwarding entry port); if a plurality of minimum cost paths exist, select a plurality of such paths, these paths in the list to the port with the node, if a shared outlet port, the gateway mac is the output port corresponding to a single physical port; if there are multiple, then after the polymerization these physical ports as a port of the virtual MAC address.

[0141] 当Trill域内各RB接收到数据报文时,不对该报文进行Trill封装,而根据对应的MAC转发表项对报文进行转发;若根据报文的目的MAC地址查询到的MAC转发表项的出端口为单端口时,则将该报文从该出端口转发出去;若根据报文的目的MAC地址查询到的MAC转发表项的出端口为聚合端口,则可以通过对应的聚合链路的hash分担,选择聚合链路的某条子链路对应的端口对该报文进行转发。 [0141] When the RB Trill each domain received data packet, no packet encapsulation performed Trill, and according to the corresponding MAC forwarding entry for packet forwarding; if queried according to the MAC address of the MAC forwarding packets the table entry is a single-port port, the packet is forwarded out from the out port; if queried according to the MAC addresses of the packets in the MAC forwarding entry port interface polymerization, the polymerization may be by a corresponding hash link sharing, selecting a sliver link aggregation port corresponding to the link packet forwarding.

[0142] 对于L3数据报文,当网关节点对该报文进行L3转发时,可以直接转发给外网路由器,而不涉及解Trill封装的处理;相应地,外网路由器将该报文转发给目的主机所在VLAN 的网关节点时,也不涉及对加Trill封装的处理。 [0142] For L3 data packet, the gateway node when a packet for L3 forwarding, can be forwarded directly to the external network router, without involving the processing solution Trill package; accordingly, the outer Router forwards the message to when the destination host where the gateway node the VLAN not involve handling of the package plus Trill.

[0143] 其中,由于同一VLAN内各网关节点配置有相同的虚IP地址/虚MAC地址,且网关节点进行MAC地址同步时同步的MAC地址为虚MAC地址,因此,Tri11域内作为接入层设备的RB 上生成的MAC转发表中对应该虚MAC地址的出端口通常为为聚合端口(包括该RB连接各网关节点的端口),因此,当作为接入层设备的RB需要将接收到host发送的L3数据报文发送给网关节点时,可以通过对该L3数据报文的目的MAC地址(网关节点的虚MAC地址)对应的MAC转发表项中出端口对应的聚合链路进行hash,选择聚合链路的某条子链路对应的端口对该报文进行转发,实现对流量的分担。 [0143] However, since each gateway nodes within the same VLAN configured with the same virtual IP address / virtual MAC address, and the gateway node MAC addresses for synchronized synchronous virtual MAC address, therefore, Tri11 art as an access layer device RB generated in the MAC forwarding table of the virtual MAC address to be a typically port aggregation port (RB including the port connecting the gateway node), and therefore, when it is desired that RB received access layer device to the host transmission L3 of the data packets to the gateway node when possible the forwarding table entry corresponding to the port in the aggregated link destination MAC address through L3 data packets (gateway node virtual MAC address) corresponding to the MAC for the hash, selecting a polymerization a sliver link into the corresponding port on the packet forwarding, to achieve share of traffic.

[0144] 需要注意的是,在本发明实施例中,MAC表项中的聚合端口和传统聚合端口的不同之处在于:传统的聚合链路两端的设备都需要配置聚合,而本发明实施例中的聚合链路只需要从一侧设备的角度看是聚合链路即可,对端不一定需要是聚合链路。 [0144] It is noted that, in the embodiment of the present invention, the polymerization port MAC table entry port and conventional polymerization except that: the conventional polymerization equipment need to be configured at both ends of the link aggregation and embodiments of the invention the polymerization link only see one side of the device from the angle of the link can be polymerized, the polymerization need not necessarily be the end of the link.

[0145] 此外,当Trill内作为网关节点的RB接收到access节点(非网关节点)返回的ARP应答报文时,需要根据该ARP应答报文的源MAC地址查询自身存储的MAC表,并将该源MAC地址对应的MAC表项中的出端口作为ARP表项中的出端口(该出端口可以是单端口或者聚合端口),其中,当网关节点到access节点存在等价路径时,该出端口为聚合端口。 [0145] Further, when RB receives the Trill gateway node to the access node (non-gateway node) returns the ARP response packet, the MAC address table stored in itself need, according to the source MAC address in the ARP response packet, and the source MAC address of the egress port corresponding to the MAC table entry is ARP entry as egress port (the egress port may be a single port or ports polymerization), wherein the gateway node when there are equal to the access path node, the port aggregation port.

[0146] 通过以上流程可以看出,在本发明实施例中,当大二层网络中设备接收到数据报文时,不对该报文进行大二层协议封装,而是根据数据报文的目的MAC地址查询MAC表项,并根据查询到的MAC转发表项将该数据报文以native (原生)形式进行转发。 [0146] As can be seen by the above process, in the embodiment of the present invention, when a large Layer 2 network device receives a data packet, the packet not large Layer protocol encapsulation, but according to the purpose of the data packet MAC MAC address lookup table entries, and the forwarding entry to the data packet according to the query to the native MAC (native) form forward. 当网关节点确定该数据报文需要进行L3转发时,不需要进行解除大二层封装的处理,即可发送给外网路由器;相应地,外网路由器将该报文发送给其他VLAN的网关节点时,也不需要进行大二层封装的处理,避免了在一个pipeline处理L3转发和大二层协议加/解封装的问题,降低了对芯片的要求,增强了大二层网络的可部署性; When the gateway node determines that the data packet needs to be forwarded when L3 is not required for processing a large release encapsulated in Layer 2, to send to the external network router; accordingly, the external network packet sent by another device to the gateway node the VLAN when not necessary to process large floor package, avoiding a pipeline processing in the L3 forwarding and high layer protocol problem plus / decapsulation, reducing the requirements for the chip, a large layer 2 network enhances deployability ;

[0147] 同时,由于配置为网关节点的大二层设备进行MAC地址同步时,使用的为虚MAC地址,且同一VLAN的各网关节点的虚MAC地址相同,因此,大二层设备中生成的对应网关节点的MAC转发表项中的出端口为聚合端口,当大二层设备需要将数据报文转发到网关节点时, 可以对相应的MAC转发表项中出端口对应的聚合链路进行hash,选择聚合链路的某条子链路对应的端口对该报文进行转发,实现了对流量的负载分担。 [0147] Meanwhile, since the configuration synchronization, and the same virtual MAC address of each gateway node using the same MAC address VLAN for the virtual MAC address, Layer large gateway node apparatus, therefore, generated in a large apparatus Layer gateway node corresponding to a MAC forwarding entry port output ports of the polymerization, when a large Layer device needs to forward the data packet to the gateway node may be a port forwarding entry corresponding to the aggregated link for the corresponding MAC hash selecting a sliver of aggregated links corresponding to the link packet to the port forward, achieving load balancing of traffic.

[0148] 为了更清楚地说明本发明实施例,下面以图3所示的Trill组网架构为例,说明本发明实施例提供的报文转发技术的具体实现过程。 [0148] In order to more clearly illustrate the embodiments of the present invention, the following Trill networking architecture shown in FIG. 3 as an example, specific implementation described embodiments provide packet forwarding techniques of the present invention.

[0149] 如图3所示,为本发明实施例提供的一种的Trill组网的架构示意图,其中: [0149] FIG. 3, the present one Trill architecture diagram according to an embodiment of networking the invention, wherein:

[0150] 各接入层设备(即access节点)的nickname分别为nickname-1、nickname_2、......、 nickname—n; [0150] nickname each access layer devices (i.e., access node) are nickname-1, nickname_2, ......, nickname-n;

[0151] 各核心层设备(即core节点)的nickname分别为nickname-gl、nickname_g2、 nickname_g3和nickname_g4; [0151] nickname of each core layer devices (i.e., core nodes) were nickname-gl, nickname_g2, nickname_g3 and nickname_g4;

[0152] nickname-gl 卿nickname为nickname-gl的设备,下同)、nickname-g2、nickname_ g3 以及nickname_g4均为网关节点;nickname-gl、nickname_g2和nickname_g3为vlanl 的网关;nickname-g2、nickname_g3 和nickname_g4 为vlan2的网关。 [0152] nickname-gl State nickname for the nickname-gl device, same below), nickname-g2, nickname_ g3, and the gateway node are nickname_g4; nickname-gl, nickname_g2 and nickname_g3 to vlanl gateway; nickname-g2, nickname_g3 and nickname_g4 to vlan2 gateway.

[0153] nickname-gl、nickname_g2和nickname_g3上配置有vlanl的三层接口,且每个网关节点配置有各不相同的实IP地址和实MC地址,同时每个网关节点配置相同的虚IP地址: 1 · 1 · 1 · 1,以及虚MAC地址:1-1-1。 [0153] nickname-gl, and configuration nickname_g2 nickname_g3 have vlanl Layer 3 interface, and each gateway node is configured with different IP address and the real address of the real MC, each gateway node is configured the same virtual IP address: 1 · 1 · 1 · 1 and the virtual MAC address: 1-1-1. 同理,nickname-g2、nickname_g3和nickname_g4上配置有vlan2的三层接口,且每个网关节点配置有各不相同的实IP地址和实MAC地址,同时每个网关节点配置相同的虚IP地址:2.2.2.1,以及虚MC地址:2-2-2。 Similarly, nickname-g2, and configuration nickname_g3 vlan2 nickname_g4 has three interfaces, and each gateway node is configured with the real different real MAC address and IP address, each gateway node is configured the same virtual IP address: 2.2.2.1, as well as virtual MC address: 2-2-2.

[0154] 在本发明实施例中,nickname-gl、nickname-g2、nickname_g3 以及nickname-g4 上均配置有EGRP协议。 [0154] In an embodiment of the present invention, nickname-gl, the nickname-g2, nickname_g3 and the nickname-g4 are arranged EGRP protocol. 启用EGRP协议之后,同一VLAN的各网关节点之间在VLAN内发送EGRP 协议报文,该协议报文可采用VLAN内广播的发送方式,进行邻居发现和master选举。 After enabling EGRP agreement between the gateway nodes within the same VLAN VLAN EGRP send packets, the protocol packets within the VLAN way to send broadcast text can be used, neighbor discovery and master election.

[0155] EGRP协议报文可以为周期性发送,发送间隔可设置。 [0155] EGRP packets may be transmitted periodically, the transmission interval can be set. 当选举为master的网关节点(以下简称master) down时,重新在剩下的网关节点之间进行邻居发现和新的master选举。 When elected as the master node gateway (hereinafter referred to as master) down, neighbor discovery and re-elects a new master gateway between the remaining nodes. 以图3为例,假设通过选举,vlanl中nickname-gl成为master,vlan2中nickname_g4是master,其余网关节点为s lave 〇 3 as an example, it is assumed elected, vlanl nickname-gl becomes the master, the nickname_g4 VLAN2 is master, the gateway node to the remaining square s lave

[0Ί5ό] 在host请求网关的ARP的处理流程中,当host发送网关的ARP请求的时候,只有master负责回复,slave不处理。 [0Ί5ό] In the process flow of the ARP request host gateway, when the gateway host of the ARP request, only the master is responsible for reply, slave untreated. 其中,在本发明实施例中,master回复的ARP应答报文的外层MC地址和sender MC地址均为虚拟MC地址。 Wherein, in the embodiment of the present invention, Master reply ARP response message and the MC address outer SENDER MC MC addresses are virtual addresses.

[0157] 例如,假设hostl · 1 · 1 · 100请求1 · 1 · 1 · 1的ARP,由nickname-gl负责回复应答报文, 应答报文为非Tri11封装的普通ARP报文,外层MAC和内层的sender MAC都是1-1-1,该ARP应答报文的格式可以如表1所示: [0157] For example, assume hostl · 1 · 1 · 100 Request ARP 1 · 1 · 1 · 1, the nickname-gl responsible Echo Reply packet, the response packet normal ARP packet encapsulated non Tri11 outer MAC and an inner sender MAC is 1-1-1, the ARP response message format may be as shown in table 1:

[0158] 表1 [0158] TABLE 1

[0159] [0159]

Figure CN104378297BD00211

[0160] 在网关请求host的ARP的处理流程中,各个master使用各自的实MAC地址作为ARP 请求报文的源MAC地址,而不使用虚拟MAC地址。 [0160] In the processing flow of the host's ARP request gateway, each master using the respective real MAC address as source MAC address in the ARP request message, without using the virtual MAC address. 对于host来说,此时网关的实MAC地址相当于同网段的另外一台虚拟host。 For the host, the real MAC address of the gateway at this time corresponds to a virtual host with additional segment.

[0161] 其中,若由于芯片限制等原因支持不了ARP表项的出端口为聚合端口,则网关以接收到ARP应答报文的端口作为ARP表项的出端口;若芯片支持ARP表项的出端口为聚合端口, 则网关接收到ARP应答报文后,根据该ARP应答报文的源MAC地址查询对应的MAC表项,若找到对应的MAC表项,copy该MAC表项的出端口作为ARP表项的出端口,从而,对于网关节点到该host存在等价路径的场合,ARP表项的出端口可以为聚合端口,可以对流量进行更好的分担和备份。 [0161] wherein, when due to the chip and other reasons not support the port the ARP entry for the aggregated ports, the gateway port to receive the ARP response packet as the outgoing port of the ARP entry; out if the chip supports ARP entry after the port aggregation port, the gateway receives the ARP response packet, the query corresponding MAC entries according to the source MAC address in the ARP response packet, if the corresponding MAC entries to find out the port copy the MAC entry as ARP the entry port, whereby, for the case where there are equal to the gateway node of the path host, an ARP entry port may be better sharing and backup traffic aggregation port.

[0162] 在本发明实施例中,可以通过命令行配置,将vlanl和vlan2的虚MAC地址通过ESA DI协议进行扩散和同步,最终所有节点均获知。 [0162] In an embodiment of the present invention may be configured by the command line, and the vlanl vlan2 virtual MAC address ESA DI by diffusion and synchronization protocol, all nodes are eventually informed. 其中,虚MAC地址I -1 -1同时归属于nickname-gl、nickname_g2 以及nickname_g3,虚MAC 地址2_2_2同时归属于nickname_g2、 nickname-g3以及nickname-g4。 Wherein the virtual MAC address assigned to the same time I -1 -1 nickname-gl, nickname_g2 and nickname_g3, while virtual MAC address assigned to 2_2_2 nickname_g2, nickname-g3 and the nickname-g4. 其中,由于网关节点的实MAC地址在本发明实施例提供的技术方案中作用不大,因此,可以配置实MC地址不需要通过ESADI协议报文进行扩散和同步。 Wherein, due to the real MAC address of the gateway node in the embodiment of the present invention is not the technical solution provided in effect, therefore, the MC address does not need to be configured real packet synchronization and diffusion through ESADI protocol.

[0163] Trill域内各RB接收到ESADI协议报文后,对于非网关节点,为了获得到达网关节点的等价路径,需要分别计算本节点到达该VLAN内各网关节点(如vlanl内的nickname-gl、 nickname_g2和nickname_g3)的单播路由,并在这些单播路由中选择cost最小的路径进行合并,并当存在多条时,则形成等价路由。 [0163] each RB Trill ESADI art after receiving the protocol message, the non-gateway nodes in order to obtain the equivalent path to the gateway node, the node needs to calculate the arrival of each of the gateway node the VLAN (such as in the nickname-gl vlanl , nickname_g2 and nickname_g3) unicast routing, and selects the minimum cost paths are combined in such unicast routing, and, when there are multiple, the cost of the route.

[0164] 非网关节点获得到达VLAN内网关节点的等价路径后,用链路聚合来表示对应的等价路径,形成最终的转发表项,并下发到硬件。 After the [0164] non-gateway nodes obtain the equivalent path reaches the VLAN gateway node, it indicates that the corresponding link of polymerization equivalent path, to form the final forwarding entries, and to hardware. 例如,图3所示的场景中,nickname-1和nickname-4生成的转发表项可以分别如表2和表3所示: For example, a scene shown in FIG. 3, nickname-1 and the nickname-4 can be generated forwarding entries are shown in Table 2 and Table 3:

[0165] 表2 [0165] TABLE 2

Figure CN104378297BD00212

[0167]表3 [0167] TABLE 3

Figure CN104378297BD00221

W]在数据转发流程中,对于非网关节点,接收到host发送的L3数据报文后,不对该数胃据报文进行Trill封装,并根据控制层面下到硬件的转发表项进行转发。 After W] in the data forwarding process, for the non-gateway node, the received L3 data packets sent by host, not the number of gastric encapsulated data packet Trill, and the hardware forwarding entries to be forwarded in accordance with the control plane. 其中,若出端口为聚合端口,则对该聚合端口对应的聚合链路进行hash,选择聚合链路的某条子链路对应的端口对该报文进行转发。 Wherein, when the port aggregation port, the port corresponding to the link aggregation polymerization is the hash, selecting a sliver link aggregation port corresponding to the link packet forwarding.

[0170] 对于网关节点,同一VLAN中各网关节点接收到非网关节点转发的host主机发送的L3数据报文,均认为自身为网关,不对该数据报文进行Tr i 11封装,并进行L3转发处理流程。 [0170] For the gateway node, the same VLAN each gateway node receives the L3 data packets non-gateway node forwards the host sent by the host, are considered itself as a gateway, without Tr i 11 encapsulating the data packet, and L3 Forwarding processes.

[0171] 以图3所示场景中vlanl内hostl (IP地址为1 · 1 · 1 · 100)向vlan2内hsot2 (IP地址为2.2.2.100)的报文转发流程为例。 [0171] In the scenario shown in FIG. 3 hostl in vlanl (IP address 1 · 1 · 1 · 100) (the IP address 2.2.2.100) to hsot2 vlan2 within the packet forwarding process, for example.

[0172] 如图4所示,为本发明实施例提供的一种三层单播报文转发方法的流程示意图,可以包括以下步骤: [0172] As shown in FIG 4, a schematic flow chart of a three layer embodiment to provide a method of forwarding unicast packets embodiment of the present invention may comprise the steps of:

[0173] 步骤400, vlanl内hostl在需要向vlan2内host2发送报文时,发现host2与自身不在同一网段(即host2不在本地网络),于是将报文发送给网关节点。 [0173] Step 400, the vlanl hostl need to send packets when the VLAN2 host2, host2 find itself in different network segments (i.e., not on the local network host2), then sends the packet to the gateway node. 如果hostl当前没有网关节点的ARP表项,则广播发送ARP请求报文,该请求报文中的target的IP为网关节点的虚IP地址。 If the gateway node is not currently hostl ARP entry, sending the broadcast ARP request packet, the request packet of the target IP address is a virtual IP gateway node. 该ARP请求报文经过access节点,转发至网关节点。 The ARP request packet through the access node, forwarded to the gateway node. 其中,access节点接收到ARP请求报文后,通过设置的VLAN的广播域范围对该ARP请求报文进行广播转发。 Wherein, after receiving the Access Node ARP request packet by broadcast domain VLAN range provided the broadcast ARP request packet forwarding.

[0174] 网关节点收到上述ARP请求报文后,若自身为对应VLAN内的mas ter网关节点,则向主机单播返回ARP应答报文,该响应报文中的sender IP为网关节点的虚IP地址,sender MAC地址为网关的虚MAC地址。 After [0174] the gateway node receiving the ARP request packet, if the VLAN itself as mas ter corresponding to the gateway node, then returns to the host unicast ARP response packet, the response packet to the sender IP virtual gateway node IP address, sender MAC address for the virtual MAC address of the gateway. 该ARP应答报文经过access节点,由access节点根据该ARP应答报文的目的MAC地址(即主机的MAC地址)查询匹配的MAC表项,转发至主机,从而主机学习至Ij网关的ARP表项,建立网关的虚IP地址与虚MAC地址之间的关联。 The ARP reply packets pass through access node, the access node query MAC entry matching according to the MAC address of the ARP reply packet (ie host MAC address), forwarded to the host, which hosts learn to ARP entries Ij gateway the association between the virtual IP address of the gateway to establish virtual MAC address.

[0175] 步骤401、hostl向nickname-1发送报文,该报文的目的MAC地址为网关节点的虚MAC地址1-1-1,目的IP为host2的IP地址2.2.2.100。 [0175] Step 401, hostl to the nickname-1 packets transmitted, the packet destination MAC address of the virtual MAC address of the gateway node 1-1-1, destination IP is the IP address of host2 2.2.2.100.

[0176] 具体的,hostl发送报文给host2,查找网关ARP表项,封装的目的MAC为1-1-1,目的IP为2.2.2.100。 [0176] Specifically, hostl sends a packet to host2, the gateway ARP lookup table entry for the destination MAC encapsulation 1-1-1, destination IP 2.2.2.100.

[0177] 步骤402、nickname-l接收到报文后,根据该报文的外层MAC地址查找MAC转发表, 并根据查询到的表项将该报文转发给nickname-gl。 [0177] Step 402, nickname-l receives the packet, it searches the MAC address of the outer packet according to the MAC forwarding table, and forwards the message to the nickname-gl according to the found entry.

[0178] 具体的,nickname-1接收到hostl发送的报文后,根据该报文的外层MAC地址(1-1- 1)查询本地MAC转发表,命中表3所示表项,该表项中出端口为聚合端口(Trunk (PI,P2, P3)),nickname-1通过hash算法选取其中一个,假设为端口Pl,并将该报文从Pl发送出去。 [0178] Specifically, nickname-1 hostl after receiving the message transmitted, based on the outer MAC address of the packet (1-1-1) queries the local MAC forwarding table, the table entry hit in FIG. 3, the table an entry port aggregation port (Trunk (PI, P2, P3)), nickname-1 which is selected by a hash algorithm, is assumed to be the port Pl, and the packet transmitted from Pl. 其中,nickname-1将该报文从Pl转发出去时,不会对该报文进行Tri 11封装,而是以将该报文以native形式发送出去。 Wherein, nickname-1 when the packet forwarded from Pl, not proceed Tri 11 encapsulate the packet, but in the native form of a message sent out.

[0179] 步骤403、nickname-gl接收到报文后,将报文转发给外网路由器。 [0179] Step 403, nickname-gl after receiving the packet, the packet will be forwarded to the external routers.

[0180] 具体的,nickname-gl接收到nickname-1转发的报文后,根据发现该报文的目的MAC地址为自身的虚MAC地址,则认为该报文是发送给自己的,此时需要查询路由表进行三层转发。 After [0180] Specifically, nickname-gl received message forwarded nickname-1, according to the MAC address of the packet is found in its virtual MAC address, then that packet is addressed to itself, this time need query routing table Layer 3 forwarding. 具体实现时,可以预先在虚MAC地址的MAC转发表项中设置特定的指示标识,这样nickname-gl根据该报文的目的MAC地址(1-1-1)查询本地MAC转发表,根据查询到的MAC转发表项中设置的指示标识,确定需要对该报文进行L3转发。 In specific implementation, may be previously virtual MAC address in the MAC forwarding entry set a specific instruction identifier, so that the local nickname-gl MAC forwarding table according to the destination MAC address of the packet (1-1-1) query, the query to MAC forwarding entry set in the instruction identifier, to determine the need for L3 packet forwarding. 于是查询路由表,由于本地没有直连路由,因此,nickname-gl将该报文转发给外网路由器。 Thus routing table, since no local direct route, and therefore, nickname-gl forwards the packet to the external routers. 其中,nickname-gl也不对该报文进行Trill封装。 Wherein, nickname-gl Trill nor be encapsulated packet.

[0181] 步骤404、外网路由器接收到报文后,根据报文的目的IP地址查询路由表,将报文车专发给nickname_g3〇 [0181] Step 404, after the external network is received by the packet according to the destination address of the IP packet routing table, the packet will be sent to the vehicle designed nickname_g3〇

[0182] 具体的,外网路由器接收到nickname-gl转发的报文后,根据该报文的目的IP地址(2.2.2.100)查询路由表,确定为等价路由,下一跳分别为nickname-g2、nickname_g3和nickname-g4,夕卜网路由器可以通过hash算法,从中选择一个作为实际下一跳,假设选定为]1;^1〇^1116-83,贝11夕卜网路由器将该报文转发给11;^1〇^1116-83。 [0182] Specifically, the external network is received by the nickname-gl packets forwarded, the destination IP address of the packet (2.2.2.100) searches the routing table, the route is determined to be equivalent, respectively, to the next hop nickname- g2, nickname_g3 and the nickname-g4, Xi Bu Router by hash algorithm, select one as the actual next hop for the selected assumed] 1; ^ ^ 1〇 1116-83, by a shell 11 will evening newspaper web Bu forwards to 11; 1〇 ^ ^ 1116-83. 其中,该报文仍未非1'1';[11封装的报文。 Wherein the packet has not yet non-1'1 '; [11 encapsulated packet.

[0183] 步骤405、nickname-g3接收到报文后,根据报文的目的IP地址查询本地ARP表,并根据查询到的表项将报文转发给nickname-4。 [0183] Step 405, nickname-g3, after receiving the message, the destination IP address of the local ARP query packet table, and forwards the packet to the nickname-4 according to the found entry.

[0184] 其中,若nickname-g3上没有host2 (2 · 2 · 2 · 100)的ARP表项,则可以使用自身的实IP地址/实MAC地址作为源IP地址/源MAC地址发起ARP请求,以获取host2的ARP信息,并根据得到的ARP信息将该报文转发给nickname-4,其中,nickname-g3发送给nickname-4的报文为非Tr ill封装报文,且目的MAC地址为host2的目的MAC地址。 [0184] wherein, if no host2 the nickname-g3 (2 · 2 · 2 · 100) ARP entry, you can use its real IP address / real MAC address as the source IP address / source MAC address initiates an ARP request, for host2 ARP information, and according to the destination MAC address in the ARP information obtained forwards the message to the nickname-4, wherein, nickname-g3 packets sent nickname-4 Tr ill non-encapsulated packets, and to host2 destination MAC address.

[0185] 步骤406、nickname-4接收到报文后,根据报文的目的MAC地址查询本地MAC转发表,并将该报文转发给host2。 [0185] Step 406, nickname-4 receives the packet according to the destination MAC address of the packet queries the local MAC forwarding table, and forwards the packet to host2.

[0186] 进一步地,在本发明实施例中,考虑到大二层网络中节点设备或链路可能发生异常,为了提高系统可用性,本发明实施例还提供了异常处理机制。 [0186] Further, in the embodiment of the present invention, in consideration to large Layer 2 network node device or link error may occur, in order to improve system availability, embodiments of the present invention further provides an exception handling mechanism.

[0187] 具体的,当非网关节点与网关节点之间的TRILL链路中断时,该非网关节点可以按照更新后的拓扑,重新计算去往同一VLAN的各网关节点的单播路由,并更新重新计算得到的单播路由,更新第一MAC地址转发表项的出接口; [0187] Specifically, when the non-TRILL link between the gateway node and the gateway node an interrupt, the gateway node can follow the non-updated topology recalculated each unicast route to the gateway node same VLAN and update unicast routing recalculation obtained, updating the first MAC address forwarding entries of the interfaces;

[0188] 该网关节点更新达到该非网关节点下所有主机的ARP表项的出端口。 [0188] The gateway node reaches an updated ARP entry port for all hosts in this non-gateway node.

[0189] 在网关节点宕机时,非网关节点更新该网关节点配置的虚MAC地址对应的第一MAC 地址转发表项的出接口。 First MAC address [0189] In the gateway node goes down, the non-gateway node configured to update the gateway node virtual MAC address forwarding entries corresponding to the interface.

[0190] 当非网关节点发生宕机(down)时,网关节点删除该非网关节点下的所有主机的ARP表项; [0190] When the non-occurrence of the gateway node down (down), the gateway node deletes the ARP entry for all hosts in the non-gateway node;

[0191] VLAN内的主网关节点down时,该VLAN内其他网关节点重新选举该VLAN内的主网关节点; [0191] When the primary gateway node within the VLAN down, this VLAN other gateway node re-election of the master gateway node within the VLAN;

[0192] 当网关节点发现外部路由器不可达时,退出本网关节点所在VLAN的主网关节点的选举,以触发其他网关节点重新选举新的主网关节点,禁用自身所有的三层接口,并通过所述广播树通告虚MAC地址的撤销消息,以使得非网关节点设备更新所述虚MAC地址对应的所述第一MAC地址转发表项的出接口。 [0192] When the gateway node finds external router is unreachable, the main gateway node election exit VLAN Ben joint point where, in order to trigger a new primary gateway node other gateway node re-election, disable all three of its own interface and through the said multicast tree revocation message advertisement virtual MAC address, so that the non-gateway node device updates the virtual MAC address corresponding to the first MAC address forwarding entry outgoing interface.

[0193] 仍以数据中心内的Trill技术为例,对本发明实施例所提供的技术方案中的异常处理机制进行详细描述。 [0193] Trill art still within the data center, for example, the present invention is described in detail with exception handling mechanism technical solutions provided in the embodiments.

[0194] 场景一、Trill链路故障 [0194] a scene, Trill link failure

[0195] 参见图5,假设nickname-1与nickname-gl之间的Trill链路发生故障,贝Ij与该!'!;[11链路直接关联的节点设备(8卩11;[010^11161和11;[010^1116-81)需要确定该链路故障后的网络拓扑,并向其他节点设备通告该网络拓扑,如可以通过ISIS协议报文向其他节点设备通告链路故障后的网络拓扑。 [0195] Referring to Figure 5, assume that the link between Trill nickname-1 and the nickname-gl fails, the shell Ij '!; [11 direct link associated node device (8 Jie 11;! [010 ^ 11161 and 11; [010 ^ 1116-81) determined by the network topology after the link failure, and other node devices advertise the network topology, the network topology as can the other node device to advertise a link failure by the ISIS protocol packets .

[0196] 各非网关节点需要按照更新后的拓扑,重新计算去往同一VLAN的各网关节点的单播路由,并更新重新计算得到的单播路由,更新第一MAC地址转发表项的出接口。 [0196] non-gateway nodes need to follow the updated topology recalculated each unicast route to the gateway node same VLAN and update unicast route recalculation obtained, updating the first entry in the MAC address forwarding outbound interface .

[0197] 例如,图5所示场景中,到虚MAC地址1-1 -1的路径中,端口PI、P2、P3对应的三条cost最小的路径中,由于端口Pl对应的路径出现了绕行,cost加大,被排除在外,因此只剩下端口P2和P3对应的两条cost最小的路径,因此,nickname-1上到虚MAC地址1-1-1的出端口刷新为新的聚合端口,对应的表项由表4刷新为表5: [0197] For example, in the scenario shown in FIG. 5, virtual MAC address to the path 1-1-1, the minimum path cost three ports PI, P2, P3 corresponding, since the path corresponding to the port Pl appears detour , increase the cost, are excluded, thus leaving only two minimum cost paths corresponding ports P2 and P3, therefore, the nickname-1 1-1-1 virtual MAC address to a port of the new refresh aggregation port , the corresponding entry in table 4 table 5 refresh:

[0198] 表4 [0198] TABLE 4

Figure CN104378297BD00241

[0202] 同时,nickname-gl节点更新到达nickname-1下的所有主机的ARP表项的出端口, 如从端口Pl更新为端口P2(绕行路径)。 [0202] Meanwhile, nickname-gl node updates the nickname-1 reaches the egress port ARP entry for all hosts, such as updating the port from port Pl to P2 (bypass path).

[0203] 场景二、access节点故障 [0203] second scenario, access node failure

[0204] 参见图6,假设nickname-ldown,贝Ij与nickname-1相关联的节点设备需要确定该节点设备故障后的网络拓扑,并向其他节点设备通告该网络拓扑。 [0204] Referring to Figure 6, it is assumed nickname-ldown, shellfish Ij nickname-1 associated with node device needs to determine the network topology after the failure of the node device to advertise the topology other network node apparatus.

[0205] 各个网关节点删除nickname-1下的所有主机的ARP表项。 [0205] each gateway node deletes the ARP entry for all hosts in the nickname-1.

[0206] 场景三、core节点故障 [0206] Scenario three, core node failure

[0207] 参见图7、假设nickname-gl (vlanl内的master)down,贝Ijvlanl内其他网关节点需要根据EGRP协议重新选举新的master (假设为nickname_g2),down的nickname-gl之前通过ESADI协议报文通告的虚MAC地址1-1-1在后续的ESADI同步更新的过程中老化。 [0207] Referring to Figure 7, it is assumed nickname-gl (master in vlanl) down, the inner shell Ijvlanl other gateway nodes need to elect a new master according EGRP protocol (assumed to nickname_g2), before the down of the nickname-gl by ESADI protocol packet Wen advertised virtual MAC address aging process 1-1-1 ESADI subsequent synchronization update. 各个节点设备刷新虚MAC地址1-1-1的新的出端口。 Each node apparatus newly refreshed virtual MAC address of the egress port 1-1-1. 例如,nickname-1到1-1-1的出端口为P2和P3组成的聚合端口,端口Pl被删除。 For example, nickname-1 to the output port P2, and P3 for 1-1-1 composition aggregation port, port Pl is deleted.

[0208] 场景四、上行链路(网关节点与外部路由器之间的链路)故障 [0208] Scene four uplink (the link between the gateway node and the external routers) fault

[0209] 参见图8,假设ni ckname-g 1与外部路由之间的链路中断,贝Ijn i ckname-g 1发现外部路由器不可达,down掉自身所有vlan上的三层接口,并通过ESADI协议报文撤销该节点发布的虚MAC地址信息,以使各非网关节点重新计算到达该网关节点配置的虚MAC地址的路径, 更新该虚MAC地址对应的第一MAC地址转发表项的出端口。 [0209] Referring to FIG 8, assuming that the link between ni ckname-g 1 and the external interrupt routing, shellfish Ijn i ckname-g 1 unreachable external router discovery, fall Layer Down interfaces on all VLAN itself, and through ESADI packets revoked node virtual MAC address information of the release so that the non-gateway nodes to recalculate a route virtual MAC address of the gateway node disposed arrives, updating a first MAC address corresponding to the virtual MAC address of the forwarding entry port . 同时,ni ckname-g 1退出所在VLAN 的EGRP选举,其余网关节点重新选举新的master,各非网关节点 Meanwhile, ni ckname-g 1 exit EGRP election is VLAN, and the remaining gateway node re-election of a new master, each non-gateway node

[0210] 例如,在图8所示的场景中,经过重新选举后,nickname_g2为新的master, nickname-1到达虚MAC地址1-1-1的出端口为仅包含端口P2和P3的聚合端口,端口Pl被删除。 Aggregation port [0210] For example, in the scenario shown in FIG. 8, after the re-election, nickname_g2 new master, nickname-1 reaches the MAC address of the virtual port 1-1-1 containing only port P2 and P3 , port Pl is deleted.

[0211] 应该认识到,上述实施例以数据中心内的Trill技术为例对本发明提供的报文转发技术方案进行了详细描述,当相关原理应用到其他大二层组网技术(如Vxlan技术、SPB技术)等中时,相应描述也应该进行调整。 [0211] It should be appreciated that the embodiments described above Trill technology within the data center as an example of the present invention provides a packet forwarding aspect been described in detail, when the correlation principle is applied to other large Layer networking technologies (e.g. Vxlan technology, SPB technology) or the like, the corresponding description also should be adjusted.

[0212] 例如,对于大二层网络中的接入层设备和核心层设备,在Trill技术中,接入层设备和核心层设备均为RB;而在Vxlan技术中,接入层设备为VTEP(Vxlan Tunnel Endpoint, Vxlan隧道端点),而对核心层设备并未专门定义;而在SPB技术中,接入层设备为BEB (Backbone Edge Bridge,骨干边缘桥)设备,核心设备为BCB(Backbone Core Bridge,骨干核心桥)设备。 [0212] For example, for a large Layer 2 network access layer device and the core devices, Trill in the art, the access layer device and the RB are core devices; Vxlan in the art, the access layer device is a VTEP (Vxlan tunnel endpoint, Vxlan tunnel endpoint), while the core layer devices not specifically defined; and in the SPB technology, the access layer device is a BEB (backbone edge bridge, backbone edge bridge) equipment, core equipment for the BCB (backbone core bridge, backbone core bridges) equipment.

[0213] 对于大二层网络转发表,在Trill技术中,该大二层网络转发表具体为Trill路由表;而在Vxlan技术中,该大二层网络转发表具体为Vxlan中的骨干网L3路由表;而在SPB技术中,该大二层网络转发表具体为SPBM中的骨干网的Bmac转发表。 [0213] For large networks Layer forwarding, Trill in the art, the large layer 2 network is particularly Trill forwarding routing table; Vxlan in the art, the large-Layer Network forwarding specifically Vxlan the backbone L3 routing table; and in the SPB technology, the large Layer 2 network forwarding specifically Bmac SPBM the backbone of forwarding.

[0214] 对于用于进行MAC地址同步的控制报文,在Trill技术中,该用于进行MAC地址同步的控制报文具体为ESADI协议报文;而在Vxlan技术和SI3B技术中,由于其并未定义主动的MAC地址同步机制,因此,可以将Trill技术中的ESADI协议报文借用在Vxlan技术和SI3B技术中,其具体实现在此不再赘述。 [0214] For MAC addresses for synchronized control packets, Trill in the art, the MAC address used for the synchronization control packet to a specific protocol packets ESADI; SI3B in Vxlan technology and art, and because it undefined active MAC address synchronization mechanism, and therefore, it can be Trill art ESADI packets borrowed Vxlan SI3B art technology and, in the specific implementation will not be repeated herein.

[0215] 基于相同的技术构思,本发明实施例还提供了上述大二层网络中的设备。 [0215] Based on the same technical concept, embodiments of the present invention further provides a floor of the large devices in the network.

[0216] 参见图9,为本发明实施例提供的一种网络设备的结构示意图,该网络设备可以应用于多个节点设备基于大二层网络协议组建的大二层网络中,多个节点设备包括作为网关设备的网关节点和作为接入设备的非网关节点,同一VLAN的各网关节点配置有相同的虚IP 地址和虚MAC地址,所述网络设备作为非网关节点应用于所述大二层网络中,该网络设备包括: [0216] Referring to Figure 9, a schematic diagram of structure of a network device according to an embodiment of the invention, the network device can be applied to a large plurality of network node devices Layer Layer network protocol based on the formation of large, a plurality of node devices as the gateway node comprises a gateway device and a non-gateway node as an access device, each of the same VLAN gateway node configured with the same virtual IP address and the virtual MAC address, the network gateway device to said node as a large non-Layer network, the network device comprising:

[0217] 地址管理模块91,用于通过一可达所述大二层网络中的所有节点设备的广播树, 接收各个网关节点通告的该网关节点配置的虚MAC地址以及所属VLAN信息; [0217] The address management module 91, for broadcast by all the tree nodes of the device up to a large Layer 2 network, each gateway node arranged to receive the advertisement gateway node virtual MAC address and the VLAN information;

[0218] 表项维护模块92,用于根据大二层网络协议和所述地址管理模块91接收到的同一虚MAC地址对应的各个网关节点的大二层网络标识,计算去往同一VLAN的各网关节点的单播路由,进而生成所述虚MAC地址对应的第一MAC地址转发表项,其中,所述第一MAC地址转发表项中的VLAN为所述虚MAC所属VLAN、目的MAC地址为所述虚MAC地址、出接口为所述单播路由中代价最小的路由对应的出接口; [0218] module 92 maintains entries for each gateway node according to large Layer Layer large network protocol and the address management module of the same virtual MAC address 91 corresponding to the received network identifier, each destined for calculating the same VLAN unicast routing gateway node, and further generating a first MAC address corresponding to the virtual MAC address forwarding entry, wherein the first MAC address VLAN forwarding table entry belongs to the virtual MAC VLAN, destination MAC address the virtual MAC address, said unicast egress interface to minimize the cost of the route corresponding to the route out of the interface;

[0219] 报文转发模块93,用于当所述网络设备接收到目的MAC地址为一网关节点的虚MAC 地址的数据报文时,根据该数据报文的目的MAC地址,查找匹配的第一MAC地址转发表项,并根据查找到的匹配表项,直接转发所述数据报文而不进行大二层网络协议封装。 When [0219] the packet forwarding module 93, configured to, when the network device receives the destination MAC address is the MAC address of the virtual data packets of a gateway node, according to the MAC address of the data packet, first find a match MAC address forwarding table entries, according to find a matching entry and forwards the packet data network without large Layer protocol encapsulation.

[0220] 其中,所述表项维护模块92还可用于, [0220] wherein the entry module 92 can be used for maintenance,

[0221] 在所述网络设备与网关节点之间的链路中断时,按照更新后的拓扑,重新计算去往同一VLAN的各网关节点的单播路由,并更新重新计算得到的单播路由,更新第一MAC地址转发表项的出接口; [0221] When the link between the network device and the gateway node an interrupt, in accordance with the updated topology recalculated each unicast route to the gateway node same VLAN and update unicast route recalculated, updating the first MAC address forwarding entries of the interfaces;

[0222] 在网关节点宕机时,更新该网关节点配置的虚MAC地址对应的第一MAC地址转发表项的出接口。 First MAC address [0222] In the gateway node goes down, updating the configuration of the gateway node virtual MAC address forwarding entry corresponding to the outgoing interface.

[0223] 如图10所示,为本发明实施例提供的另一种网络设备的结构示意图,该网络设备可以应用于多个节点设备基于大二层网络协议组建的大二层网络中,多个节点设备包括作为网关设备的网关节点和作为接入设备的非网关节点,同一VLAN的各网关节点配置有相同的虚IP地址和虚MAC地址,所述网络设备作为网关节点应用于所述大二层网络中,该网络设备包括: [0223] As shown in FIG. 10, a schematic structural diagram of another network device according to an embodiment of the present invention, the network device can be applied to a large plurality of network node devices Layer Layer network protocol based on the formation of large, the multi- node apparatus includes a gateway node as an access gateway device and a non-gateway node devices, each of the same VLAN gateway node configured with the same virtual IP address and the virtual MAC address, the network device as a gateway node to said large Layer 2 network, the network device comprising:

[0224] 地址管理模块101,用于通过一可达所述大二层网络中的所有节点设备的广播树, 通告本节点设备配置的虚MAC地址以及所属VLAN信息,以使非网关节点设备根据大二层网络协议和接收到的同一虚MAC地址对应的各个网关节点的大二层网络标识,计算去往同一VLAN的各网关节点的单播路由,进而生成所述虚MAC地址对应的第一MAC地址转发表项,其中,所述第一MAC地址转发表项中的VLAN为所述虚MAC所属VLAN、目的MAC地址为所述虚MAC 地址、出接口为所述单播路由中代价最小的路由对应的出接口; [0224] The address management module 101, a multicast tree by all the node devices up to a large Layer 2 network, virtual MAC address and the VLAN configuration information advertisement node device, so that a non-gateway node apparatus according to the large Layer 2 network gateway node identifiers of network protocol and the same Layer large virtual MAC address corresponding to the received calculated for each unicast route to the gateway node of the same VLAN, thereby generating a first address corresponding to the virtual MAC MAC address forwarding entries, wherein the first MAC address VLAN forwarding table entry belongs to the virtual MAC VLAN, destination MAC address of the virtual MAC address, the unicast interface minimum cost route route outbound interface;

[0225] 报文转发模块102,用于当所述网络设备接收到目的MAC地址为自身虚MAC地址的数据报文,且判断出所述数据报文的目的MAC地址为自身虚MAC地址时,根据所述数据报文的目的IP地址,查找路由表转发所述数据报文;其中,所述数据报文是非网关节点根据报文的目的MAC地址,匹配所述第一MAC地址转发表项后直接转发的且没有进行大二层网络协议封装。 [0225] message forwarding module 102, when the network device for the destination MAC address receives the data packets of the virtual MAC address itself, and determines that the destination MAC address of the data packet when the virtual MAC address of its own, after wherein the data packet gateway node according to a non-packet destination MAC address matching the MAC address forwarding entries first; destination IP address of the data packet, the routing table to forward the data packet and straight forward without major protocol encapsulation layer 2 network.

[0226] 其中,该网络设备还包括:表项维护模块103; [0226] wherein, the network device further comprising: a maintenance module 103 entries;

[0227] 所述地址管理模块101还可用于,从所述广播树上接收其他节点设备通告的所述其他节点设备新学习到的本地MAC地址以及所属VLAN;所述本地MAC地址包括从所述其他节点设备连接的用户网络侧学习到的MAC地址; [0227] The address management module 101 may also be configured to receive the advertised another apparatus other nodes learn of new devices to the local MAC address and the VLAN ID from the broadcast tree; the local MAC address comprises from the users connected to the network node apparatus other side of learned MAC addresses;

[0228] 表项维护模块103,用于根据预先基于该大二层网络协议生成的大二层网络转发表项,生成所述其他节点设备的本地MAC地址对应的第二MAC地址转发表项,所述第二MAC地址转发表项中的目的MAC地址为所述其他节点设备的本地MAC地址,出接口为所述大二层网络转发表项中去往所述其他节点设备的出接口; [0228] entries maintenance module 103, according to predetermined network protocol based on the generation of large Layer Layer large network forwarding entries, generating the other node device corresponding to a second local MAC address MAC address forwarding entries, the second object of the MAC address forwarding table is a MAC address of the local MAC address of the other node device, a network interface of the large Layer forwarding entry device destined for the other nodes in the outgoing interface;

[0229] 所述报文转发模块102还可用于,当所述网络设备接收到目的MAC地址不是自身端口MAC地址或所述虚MAC地址的单播数据报文后,根据所述单播数据报文的目的MAC地址及所属VLAN,在所述第二MAC地址转发表项中进行匹配,并通过匹配表项的出接口直接转发所述单播数据报文而不进行大二层网络协议封装。 [0229] The message forwarding module 102 may also be used, when the network device receives the destination MAC address is not its own MAC address or port of the virtual MAC address unicast data packets, unicast data packets according to the destination MAC address and the VLAN packets, the second MAC address forwarding matching entry, and forward the unicast packet data network without large Layer protocol encapsulation matching entry directly through the interface of the publication.

[0230] 其中,所述网络设备还与所述大二层网络外部的路由设备连接; [0230] wherein the network device is further connected to the external network routing Layer large equipment;

[0231] 所述大二层网络中各节点设备上设置有VLAN的广播域范围,其中,所述广播域范围包括该节点设备上允许该VLAN通过的用户侧端口和该节点设备上的位于该VLAN裁剪树上的大二层网络侧接口,所述裁剪树是基于VLAN对所述广播树进行分发树的剪枝得到的; [0231] The Network Layer is provided with a large broadcast domain VLAN range in each node device, wherein the broadcast domain on the range comprises allowing the VLAN specified by the user-side port of the node device and the node device large Layer 2 network side interface VLAN cut tree, the tree was cut based VLAN pruning tree for the broadcast distribution tree;

[0232] 所述报文转发模块102还可用于,在接收到所述路由设备发送的数据报文时,查找匹配所述数据报文的目的IP地址的ARP表项,并根据查找到的ARP表项的目的MAC地址以及出接口,直接转发所述数据报文而不进行大二层网络协议封装。 [0232] The message forwarding module 102 may also be used, upon receiving the data packet sent by the routing device, the ARP lookup entry matching the destination IP address of the data packet, and according to the searched ARP the destination MAC address entry and egress interface to forward the data packet directly without large network Layer protocol encapsulation.

[0233] 其中,所述报文转发模块102还可用于,若未查找到匹配所述数据报文的目的IP地址的ARP表项,在该数据报文对应的VLAN的广播域内,直接广播发送用于请求所述数据报文的目的IP地址对应的MAC地址的第一ARP请求报文而不进行大二层网络协议封装,所述第一ARP请求报文的发送端sender IP地址是所述网络设备的发送端口IP地址,源MAC地址及发送端MAC地址是所述网络设备的发送端口MAC地址; [0233] wherein the message forwarding module 102 may also be used, if not found ARP entry matching the destination IP address of the data packet, the data in the broadcast domain VLAN corresponding to the packet directly broadcast transmission a first MAC address for the ARP request for the destination IP address of the data packet corresponding to a request packet network without large Layer protocol encapsulation, the first ARP request packet sender IP address is the SENDER transmission port of the network device IP address, MAC address and the transmission source MAC address is the MAC address of the transmission port of the network device;

[0234] 所述表项维护模块103还可用于,当所述网络设备接收到根据接收到针对第一ARP 请求报文的第一ARP应答报文时,建立所述数据报文的目的IP地址对应的ARP表项; [0234] maintenance of the entry module 103 may also be used, when the network device receives a response packet according to the received first ARP ARP request for the first packet, the destination IP address establishing the data packet corresponding ARP entry;

[0235] 所述报文转发模块102还可用于,根据所述表项维护模块建立的ARP表项,转发所述数据报文而不进行大二层网络协议封装。 [0235] The message forwarding module 102 may also be used, maintaining an ARP entry module established according to the entry, not to forward the data packet Network Layer protocol encapsulation large.

[0236] 其中,同一VLAN中的一个网关节点被选举为该VLAN的主网关节点, [0236] wherein a gateway node in the same VLAN is elected as the VLAN primary gateway node,

[0237] 所述报文转发模块102还可用于,当接收到主机发送的用于请求网关节点的MAC地址的第二ARP请求报文,且确定所述网络设备为对应VLAN的主网关节点时,生成第二ARP应答报文,该第二ARP应答报文的目的MAC地址为所述主机的MAC地址,源MAC地址和发送端sender MAC地址均为所述网络设备的虚MAC地址,sender IP地址为所述网络设备的虚IP地址,以使所述主机学习并建立所述网络设备的虚IP地址与虚MAC地址之间的关联关系;根据所述ARP应答报文的目的MAC地址,在第二MAC地址转发表项中进行匹配,并通过匹配表项的出接口直接转发所述ARP应答报文而不进行大二层网络协议封装 [0237] The message forwarding module 102 may also be used, upon receiving the second ARP request for the MAC address of the gateway node sent by the host request message, and determining the network device is a master gateway node corresponding to the VLAN generating a second ARP response message, the second ARP response packet to the destination MAC address of the host MAC address, source MAC address and the MAC address of both the end SENDER virtual MAC address of the network device, sender IP association between the network address of the virtual IP address of the device, so that the host device to learn and establishing the virtual network IP address and virtual MAC address; according to the destination MAC address of the ARP response packet, in second MAC address forwarding matching entry, and forwarded directly through the matching entry interface of the ARP response packet network without encapsulation leave large Layer

[0238] 其中,所述网络设备还包括: [0238] wherein, the network device further comprises:

[0239] 选举模块104,用于在所述网络设备所属的VLAN内广播携带有所述网络设备自身端口IP地址、端口MAC地址、虚IP及虚MAC地址的邻居发现及主网关节点选举协议报文;接收所述VLAN内其他网关节点发送的所述协议报文,选举端口IP地址和/或端口MC地址满足预设规则的网关节点,作为所述主网关节点。 [0239] module 104 Election configured within VLAN broadcast the network device belongs port carries its own neighbor IP address, port MAC address, IP address and the virtual MAC address of the virtual network device discovery and the main gateway node election protocol packets text; gateway node receiving the protocol message sent by the VLAN another gateway node, the IP address and port election / MC address or port satisfies a predetermined rule, as the primary gateway node.

[0240] 其中,所述表项维护模块103还可用于,在非网关节点与所述网络设备之间的链路中断时,更新达到该非网关节点下所有主机的ARP表项的出接口;在非网关节点发生宕机down时,删除该非网关节点下的所有主机的ARP表项; [0240] wherein the entry module 103 may also be used to maintain, when the link between the gateway node and the non-network device interrupt, update the outgoing interface to reach all hosts ARP entry in the non-gateway node; when a non-gateway node occurs downtime down, delete the ARP entries for all hosts in the non-gateway node;

[0241] 所述地址管理模块101还可用于,在所述网络设备发现外部路由器不可达时,禁用自身所有的三层接口,并通过所述广播树通告虚MAC地址的撤销消息,以使得非网关节点设备更新所述虚MAC地址对应的所述第一MC地址转发表项的出接口; [0241] The address management module 101 may also be used, when the external network device discovery unreachable router, disable all three interfaces itself, and notifies virtual MAC address withdraw message broadcast by the tree, so that the non a gateway node device updates the virtual MAC address corresponding to the first address entry MC outgoing interface;

[0242] 所述选举模块104还可用于,在所述网络设备发现外部路由器不可达时,退出本网关节点所在VLAN的主网关节点的选举,以触发其他网关节点重新选举主网关节点。 [0242] The election of block 104 may also be used, when the external network device discovery unreachable router, gateway node exit RO Ben articulation point is located VLAN to trigger the re-election of the master gateway node to other gateway nodes.

[0243] 通过以上的实施方式的描述,本领域的技术人员可以清楚地了解到本发明可借助软件加必需的通用硬件平台的方式来实现,当然也可以通过硬件,但很多情况下前者是更佳的实施方式。 [0243] By the above described embodiments, those skilled in the art may clearly understand that the present invention may be implemented by software plus a necessary universal hardware platform, also be implemented by hardware, but the former is in many cases more good embodiments. 基于这样的理解,本发明的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来,该计算机软件产品存储在一个存储介质中,包括若干指令用以使得一台终端设备(可以是手机,个人计算机,服务器,或者网络设备等)执行本发明各个实施例所述的方法。 Based on such understanding, the technical solutions of the present invention in essence or the part contributing to the prior art may be embodied in a software product, which computer software product is stored in a storage medium and includes several instructions to enable a terminal devices (which may be a mobile phone, a personal computer, a server, or network device) to execute the methods according to embodiments of the present invention.

[0244] 以上所述仅是本发明的优选实施方式,应当指出,对于本技术领域的普通技术人员来说,在不脱离本发明原理的前提下,还可以做出若干改进和润饰,这些改进和润饰也应视本发明的保护范围。 [0244] The above are only preferred embodiments of the present invention, it should be noted that those of ordinary skill in the art, in the present invention without departing from the principles of the premise, can make various improvements and modifications, such modifications and modifications should also depend on the scope of the present invention.

Claims (18)

1. 一种报文转发方法,应用于多个节点设备基于大二层网络协议组建的大二层网络中,其特征在于,所述多个节点设备包括作为网关设备的网关节点和作为接入设备的非网关节点,同一VLAN的各网关节点配置有相同的虚IP地址和虚MAC地址,该方法包括: 非网关节点通过一可达所述大二层网络中的所有节点设备的广播树,接收各个网关节点通告的该网关节点配置的虚MAC地址以及所属VLAN信息; 非网关节点根据大二层网络协议和接收到的同一虚MAC地址对应的各个网关节点的大二层网络标识,计算去往同一VLAN的各网关节点的单播路由,进而生成所述虚MAC地址对应的第一MAC地址转发表项,其中,所述第一MAC地址转发表项中的VLAN为所述虚MAC所属VLAN、目的MAC地址为所述虚MAC地址、出接口为所述单播路由中代价最小的路由对应的出接口; 非网关节点接收到目的MAC A packet forwarding method applied to a large plurality of node devices Layer Layer large networking based network protocol, wherein said apparatus comprises a plurality of nodes as a gateway node as an access gateway device and non gateway node devices, each of the same VLAN gateway node configured with the same virtual IP address and the virtual MAC address, the method comprising: a non-gateway node broadcasts to all tree nodes of the device up to a large Layer 2 networks, virtual MAC address and the VLAN configuration information of each gateway node receiving the advertisement gateway node; Layer large network identification corresponding to a non-gateway node according to a large network protocol and the received Layer same virtual MAC address of each gateway node, to calculate unicast route to the gateway node of the same VLAN, thereby generating the virtual MAC addresses corresponding to a first forwarding entry, wherein the first MAC address forwarding entry in the virtual MAC VLAN is the VLAN , destination MAC address is the virtual MAC address, the interface of said unicast-cost routing outbound interface minimum route; non-gateway node receives the destination MAC 地址为一网关节点的虚MAC地址的数据报文时,根据该数据报文的目的MAC地址,查找匹配的第一MAC地址转发表项,并根据查找到的匹配表项,直接转发所述数据报文而不进行大二层网络协议封装。 When the packet data address is virtual MAC address of a gateway node, according to the MAC address of the data packet, looks up the MAC address forwarding entries first, and according to find a matching entry, directly forward the data Layer packets without large network protocol encapsulation.
2. 如权利要求1所述的方法,其特征在于,该方法还包括: 在非网关节点与网关节点之间的链路中断时,非网关节点按照更新后的拓扑,重新计算去往同一VLAN的各网关节点的单播路由,并更新重新计算得到的单播路由,更新第一MAC 地址转发表项的出接口; 在网关节点宕机时,非网关节点更新该网关节点配置的虚MAC地址对应的第一MAC地址转发表项的出接口。 2. The method according to claim 1, wherein the method further comprises: the link between the non-gateway node and the gateway node an interrupt, a non-gateway node according to the updated topology, go to the same VLAN recalculation unicast routing each gateway node, and to update the unicast route recalculation obtained, updating the first entry of the MAC address forwarding outbound interface; down at the gateway node, a non-gateway node updates the virtual MAC address of the gateway node configuration corresponding to a first MAC address forwarding entries outgoing interface.
3. —种报文转发方法,应用于多个节点设备基于大二层网络协议组建的大二层网络中,其特征在于,所述多个节点设备包括作为网关节点的网关节点和作为接入设备的非网关节点,同一VLAN的各网关节点配置有相同的虚IP地址和虚MAC地址,该方法包括: 网关节点通过一可达所述大二层网络中的所有节点设备的广播树,通告本节点设备配置的虚MAC地址以及所属VLAN信息,以使非网关节点设备根据大二层网络协议和接收到的同一虚MAC地址对应的各个网关节点的大二层网络标识,计算去往同一VLAN的各网关节点的单播路由,进而生成所述虚MAC地址对应的第一MAC地址转发表项,其中,所述第一MAC地址转发表项中的VLAN为所述虚MAC所属VLAN、目的MAC地址为所述虚MAC地址、出接口为所述单播路由中代价最小的路由对应的出接口; 网关节点接收到目的MAC地址为自身虚MAC 3 - Species packet forwarding method applied to a large plurality of node devices Layer Layer large networking based on network protocols, wherein the plurality of node devices includes a gateway node as an access gateway node and non gateway node devices, each of the same VLAN gateway node configured with the same virtual IP address and the virtual MAC address, the method comprising: a gateway node broadcasts to all tree nodes of the device up to a large Layer 2 networks, advertised virtual MAC address and the VLAN configuration information of the local node device, so that a large non-gateway node apparatus according to the respective Layer Layer gateway node network protocols and large the same virtual MAC address corresponding to the received network identification, computing destined same VLAN unicast routing each gateway node, thereby generating a first virtual MAC addresses corresponding forwarding entry, wherein the first MAC address forwarding entry to the virtual MAC VLAN published belongs VLAN, destination MAC the virtual address is a MAC address, an interface unicasting the route corresponding to the route with the smallest cost of the outgoing interface; receiving the gateway node itself to the destination MAC address is a virtual MAC 址的数据报文,所述数据报文是非网关节点根据报文的目的MAC地址,匹配所述第一MAC地址转发表项后直接转发的且没有进行大二层网络协议封装; 所述网关节点在判断出所述数据报文的目的MAC地址为自身虚MAC地址时,根据所述数据报文的目的IP地址,查找路由表转发所述数据报文。 Packet data address, the data packet gateway node according to a non-packet destination MAC address matching the MAC address of the first entry is not straight forward and Network Layer protocol encapsulation large; the gateway node when it is determined that the data packet destination MAC address of its own virtual MAC address, the destination IP address of the data packet, the routing table to forward the data packet.
4. 如权利要求3所述的方法,其特征在于,该方法还包括: 各个节点设备通过所述广播树,将新学习到的本地MAC地址以及所属VLAN通告给其他节点设备,所述本地MAC地址包括从本节点设备连接的用户网络侧学习到的MAC地址; 网关节点从所述广播树上接收其他节点设备通告的所述其他节点设备新学习到的本地MAC地址以及所属VLAN,并根据预先基于该大二层网络协议生成的大二层网络转发表项, 生成所述其他节点设备的本地MAC地址对应的第二MAC地址转发表项,所述第二MAC地址转发表项中的目的MAC地址为所述其他节点设备的本地MAC地址,出接口为所述大二层网络转发表项中去往所述其他节点的出接口; 所述网关节点接收到目的MAC地址不是自身端口MAC地址或所述虚MAC地址的单播数据报文后,根据所述单播数据报文的目的MAC地址及所属VLAN,在所述第二MAC地址 4. The method according to claim 3, characterized in that, the method further comprising: each node of the multicast tree through the device, a new learned local MAC address and the VLAN advertisement to another node apparatus, the local MAC this address includes learning from the user side of the network node apparatus connected to the MAC address; gateway node receiving the advertisement of another apparatus other nodes learn of new devices to the local MAC address and the VLAN ID from the broadcast tree, and according to the pre forwarding entry based on the large-Layer network Layer protocol generated large network, the generated second MAC address of the other node device corresponding to the local MAC address forwarding entry, the second MAC address forwarding destination MAC entry address of the local MAC address of the other node device, the interface of the items destined for other nodes in the network the forwarding table Layer an interface; receiving at the gateway node to the destination MAC address is not its own MAC address or port after the unicast data packet the virtual MAC address, according to the destination MAC address and the VLAN of the unicast data packet in the second MAC address 发表项中进行匹配,并通过匹配表项的出接口直接转发所述单播数据报文而不进行大二层网络协议封装。 Matching table entry, and forward the unicast packet data network without large Layer protocol encapsulation matching entry directly through the interface.
5. 如权利要求4所述的方法,其特征在于, 所述网关节点还与所述大二层网络外部的路由设备连接; 在各个节点设备上设置VLAN的广播域范围,其中,所述广播域范围包括该节点设备上允许该VLAN通过的用户侧端口和该节点设备上的位于该VLAN裁剪树上的大二层网络侧接口,所述裁剪树是基于VLAN对所述广播树进行分发树的剪枝得到的; 网关节点在接收到所述路由设备发送的数据报文时,查找匹配所述数据报文的目的IP 地址的ARP表项,并根据查找到的ARP表项的目的MAC地址以及出接口,直接转发所述数据报文而不进行大二层网络协议封装。 5. The method according to claim 4, wherein said gateway node is further connected to the external network routing devices larger Layer; broadcast domain VLAN range is provided in each node device, wherein the broadcast the cut tree is located on VLAN domains include the node device by allowing the VLAN of the user-side port and the network node apparatus side interface Layer large, the cut tree is a tree-based VLAN for the broadcast distribution tree pruning obtained; the gateway node when packet data is routed to the receiving device, the ARP lookup entry matching the destination IP address of the data packet, and the destination MAC address of the found entry according to the ARP and an interface to directly forward the data packet network without large Layer protocol encapsulation.
6. 如权利要求5所述的方法,其特征在于, 若网关节点未查找到匹配所述数据报文的目的IP地址的ARP表项,所述方法还包括: 所述网关节点在该数据报文对应的VLAN的广播域内,直接广播发送用于请求所述数据报文的目的IP地址对应的MAC地址的第一ARP请求报文而不进行大二层网络协议封装,所述第一ARP请求报文的发送端sender IP地址是所述网关节点的发送端口IP地址,源MAC地址及发送端MC地址是所述网关节点的发送端口MC地址; 所述网关节点根据接收到的针对第一ARP请求报文的第一ARP应答报文,建立所述数据报文的目的IP地址对应的ARP表项,并根据该建立的ARP表项,转发所述数据报文而不进行大二层网络协议封装。 6. The method according to claim 5, characterized in that, if the gateway node is not found ARP entry matching the destination IP address of the data packet, the method further comprising: the gateway node in the data packet VLAN corresponding text broadcast domain, transmitting a first direct broadcast ARP request for the destination IP address of the data packet corresponding to the MAC address request message without a large network Layer protocol encapsulation, the first ARP request sending end packet sender IP address is the IP address of the transmission port of the gateway node, and sending the source MAC address of MC port address is transmitted to the MC address of the gateway node; the gateway node according to the received ARP for the first requesting the first ARP response packet, the ARP entry of the data packet corresponding to the destination IP address, and according to the ARP table entry is established, forward the data packet without performing a large network protocol Layer package.
7. 如权利要求4所述的方法,其特征在于,同一VLAN中的一个网关节点被选举为该VLAN 的主网关节点,该方法还包括: 网关节点接收到主机发送的用于请求网关节点的MAC地址的第二ARP请求报文,且确定自身为对应VLAN的主网关节点时,生成第二ARP应答报文,该第二ARP应答报文的目的MAC地址为所述主机的MAC地址,源MAC地址和发送端sender MAC地址均为所述网关节点的虚MAC 地址,sender IP地址为所述网关节点的虚IP地址,以使所述主机学习并建立所述网关节点的虚IP地址与虚MAC地址之间的关联关系; 所述网关节点根据所述ARP应答报文的目的MAC地址,在第二MAC地址转发表项中进行匹配,并通过匹配表项的出接口直接转发所述ARP应答报文而不进行大二层网络协议封装。 7. The method according to claim 4, wherein a gateway node in the same VLAN gateway is elected as the master node VLAN, the method further comprising: a gateway node receives a request for the host to send the gateway node MAC address of the second ARP request message, and determines its own corresponding primary VLAN gateway node generates a second ARP response message, the second ARP response packet destination MAC address is the MAC address of the host, the source MAC address and the MAC address of both the end SENDER gateway node virtual MAC address, the IP address of the virtual SENDER IP address of the gateway node, so that the virtual host IP address learning and with the establishment of the virtual gateway node association between the MAC address; the gateway node according to the destination MAC address of the ARP response packet, the second MAC address forwarding matching entry in the ARP reply and forwarding matching entry directly through the interface of the Layer packets without large network protocol encapsulation.
8. 如权利要求7所述的方法,其特征在于,所述主网关节点的选举包括: 同一VLAN中的各个网关节点在该VLAN内广播携带有各自端口IP地址、端口MAC地址、虚IP及虚MC地址的邻居发现及主网关节点选举协议报文; 同一VLAN中的各个网关节点接收其他网关节点发送的所述协议报文,选举端口IP地址和/或端口MC地址满足预设规则的网关节点,作为所述主网关节点。 Virtual IP gateway and each node in the same VLAN is the VLAN that carries broadcast IP address of each port, port MAC address: The method as claimed in claim 7, wherein the primary gateway node election comprises MC address neighbor discovery and virtual master gateway node election protocol packets; each gateway node in the same VLAN receiving the protocol message sent by the node to other gateway, election port IP address and / or port address satisfies preset rule MC gateway node as the primary gateway node.
9. 如权利要求3所述的方法,其特征在于,还包括: 在非网关节点与网关节点之间的链路中断时,该网关节点更新达到该非网关节点下所有主机的ARP表项的出接口; 在非网关节点发生宕机down时,网关节点删除该非网关节点下的所有主机的ARP表项; VLAN内的主网关节点down时,该VLAN内其他网关节点重新选举该VLAN内的主网关节占. 网关节点在发现外部路由器不可达时,退出本网关节点所在VLAN的主网关节点的选举,以触发其他网关节点重新选举主网关节点,禁用自身所有的三层接口,并通过所述广播树通告虚MAC地址的撤销消息,以使得非网关节点设备更新所述虚MAC地址对应的所述第一MAC地址转发表项的出接口。 9. The method according to claim 3, characterized in that, further comprising: when the link between the gateway node and the non-gateway node an interrupt, update the gateway node to reach all hosts ARP entry in the non-gateway node the Interface; down when downtime occurs in a non-gateway nodes, gateway node deletes the ARP entries for all hosts in the non-gateway node; the main gateway node within the VLAN down, the VLAN other gateway node re-election of the VLAN the main gateway section accounted gateway node when it finds an external router does not reach, withdraw from the election the main gateway node of the VLAN Ben joint point where, in order to trigger the re-election of other primary gateway node gateway node, disable all three of its own interface and through the said multicast tree revocation message advertisement virtual MAC address, so that the non-gateway node device updates the virtual MAC address corresponding to the first MAC address forwarding entry outgoing interface.
10. —种网络设备,应用于多个节点设备基于大二层网络协议组建的大二层网络中,其特征在于,所述多个节点设备包括作为网关设备的网关节点和作为接入设备的非网关节点,同一VLAN的各网关节点配置有相同的虚IP地址和虚MAC地址,所述网络设备作为非网关节点应用于所述大二层网络中,该网络设备包括: 地址管理模块,用于通过一可达所述大二层网络中的所有节点设备的广播树,接收各个网关节点通告的该网关节点配置的虚MAC地址以及所属VLAN信息; 表项维护模块,用于根据大二层网络协议和所述地址管理模块接收到的同一虚MAC地址对应的各个网关节点的大二层网络标识,计算去往同一VLAN的各网关节点的单播路由, 进而生成所述虚MAC地址对应的第一MAC地址转发表项,其中,所述第一MAC地址转发表项中的VLAN为所述虚MAC所属VLAN、目的MAC地址为所 10. - kind of network devices, a plurality of node devices applied to large Layer Layer large networking based on network protocols, wherein the plurality of node devices includes a gateway device and a gateway node as an access device a non-gateway nodes, the gateway nodes each of the same VLAN configured with the same virtual IP address and the virtual MAC address, the network device as a gateway node to said non-Layer large network, the network device comprising: address management module, with broadcast by all nodes in the tree of the device up to a large Layer 2 network, virtual MAC address and the VLAN configuration information of each gateway node receiving the advertisement gateway node; entry maintenance module for a large two-story large Layer 2 network gateway node identifiers of a network protocol and the address management module of the same virtual MAC address corresponding to the received calculated destined for each of the same VLAN gateway node unicast route, thereby generating the virtual MAC address corresponding to a first MAC address forwarding entries, wherein the first MAC address forwarding entry in a VLAN is the VLAN virtual MAC belongs, the destination MAC address 虚MAC地址、出接口为所述单播路由中代价最小的路由对应的出接口; 报文转发模块,用于当所述网络设备接收到目的MAC地址为一网关节点的虚MAC地址的数据报文时,根据该数据报文的目的MAC地址,查找匹配的第一MAC地址转发表项,并根据查找到的匹配表项,直接转发所述数据报文而不进行大二层网络协议封装。 The virtual MAC address, the unicast interface minimum route cost of routing outbound interface; message forwarding module, configured to, when the network device receives the destination MAC address is a virtual MAC address of a data packet gateway node when the text, according to the MAC address of the data packet, looks up the MAC address forwarding entries first, and according to find a matching entry and forwards the packet data network without large Layer protocol encapsulation.
11. 如权利要求10所述的网络设备,其特征在于,所述表项维护模块还用于, 在所述网络设备与网关节点之间的链路中断时,按照更新后的拓扑,重新计算去往同一VLAN的各网关节点的单播路由,并更新重新计算得到的单播路由,更新第一MAC地址转发表项的出接口; 在网关节点宕机时,更新该网关节点配置的虚MAC地址对应的第一MAC地址转发表项的出接口。 11. The network apparatus according to claim 10, wherein the entry module further configured to maintain the link between the network device and the gateway node is interrupted, the topology according to the updated recalculation Flights each gateway node unicast same VLAN routing update unicast route and re-calculated, updating a first MAC address forwarding entries outgoing interface; down at the gateway node, a gateway node is configured to update the virtual MAC a first address corresponding to the MAC address forwarding entry of the interface.
12. —种网络设备,应用于多个节点设备基于大二层网络协议组建的大二层网络中,其特征在于,所述多个节点设备包括作为网关设备的网关节点和作为接入设备的非网关节点,同一VLAN的各网关节点配置有相同的虚IP地址和虚MAC地址,所述网络设备作为网关节点应用于所述大二层网络中,该网络设备包括: 地址管理模块,用于通过一可达所述大二层网络中的所有节点设备的广播树,通告本节点设备配置的虚MAC地址以及所属VLAN信息,以使非网关节点设备根据大二层网络协议和接收到的同一虚MAC地址对应的各个网关节点的大二层网络标识,计算去往同一VLAN的各网关节点的单播路由,进而生成所述虚MAC地址对应的第一MAC地址转发表项,其中,所述第一MAC地址转发表项中的VLAN为所述虚MAC所属VLAN、目的MAC地址为所述虚MAC地址、出接口为所述单播路由中代价 12. - kind of network devices, a plurality of node devices applied to large Layer Layer large networking based on network protocols, wherein the plurality of node devices includes a gateway device and a gateway node as an access device a non-gateway nodes, the gateway nodes each of the same VLAN configured with the same virtual IP address and the virtual MAC address, the network device as a gateway node to said large Layer 2 network, the network device comprising: address management means for All tree nodes by broadcasting the device up to a large Layer 2 network, virtual MAC address and the VLAN configuration information advertisement node device, so that a large non-gateway node apparatus according to the network protocol and the received Layer same large Layer 2 network gateway node identifiers of the virtual MAC address, the unicast route to the calculation of each gateway node of the same VLAN, thereby generating a first MAC address of the virtual MAC address forwarding table entries, wherein said a first MAC address entry to the virtual MAC VLAN published belongs VLAN, destination MAC address of the virtual MAC address, the interface of said unicast-cost routing 小的路由对应的出接口; 报文转发模块,用于当所述网络设备接收到目的MAC地址为自身虚MAC地址的数据报文,且判断出所述数据报文的目的MAC地址为自身虚MAC地址时,根据所述数据报文的目的IP地址,查找路由表转发所述数据报文;其中,所述数据报文是非网关节点根据报文的目的MC地址,匹配所述第一MC地址转发表项后直接转发的且没有进行大二层网络协议封装。 Small outbound interface corresponding to the route; message forwarding module, configured to, when the network device receives the destination MAC address of the data packets across the virtual MAC address, destination MAC address and it is determined that the data packet is its own virtual MAC address, the destination IP address of the data packet, routing table look forward the data packet; wherein said non-packet data gateway node according to the destination address of the packet MC, the MC address matches the first forwarded directly after forwarding entry and without large Layer 2 network protocol encapsulation.
13. 如权利要求12所述的网络设备,其特征在于,该网络设备还包括:表项维护模块; 所述地址管理模块还用于,从所述广播树上接收其他节点设备通告的所述其他节点设备新学习到的本地MAC地址以及所属VLAN;所述本地MAC地址包括从所述其他节点设备连接的用户网络侧学习到的MAC地址; 表项维护模块,用于根据预先基于该大二层网络协议生成的大二层网络转发表项,生成所述其他节点设备的本地MAC地址对应的第二MAC地址转发表项,所述第二MAC地址转发表项中的目的MAC地址为所述其他节点设备的本地MAC地址,出接口为所述大二层网络转发表项中去往所述其他节点设备的出接口; 所述报文转发模块还用于,当所述网络设备接收到目的MAC地址不是自身端口MAC地址或所述虚MAC地址的单播数据报文后,根据所述单播数据报文的目的MAC地址及所属VLAN, 在所述第 13. The network apparatus of claim 12, wherein the network device further comprising: a maintenance module entries; said address management module is further configured to receive the other apparatus from the broadcast advertisement node tree other node apparatus newly learned belongs to the VLAN and the MAC address of the local; the local address comprises a MAC learning from said another network node apparatus connected to the user-side MAC address; maintenance module entry, based on the second year according to previously generating a large network protocol layer 2 network forwarding entry, the second MAC address of other nodes generating a local device address corresponding to the MAC forwarding entry, the second MAC address destination MAC address forwarding table items as the local MAC address of the other node device, a network interface of the large Layer forwarding entry addressed to the other node device that interfaces; the message forwarding module is further configured to, when the network device receives the object after the MAC address is not its own MAC address or port of the virtual MAC address unicast data packets, according to the destination MAC address and the VLAN said unicast data message, the first MAC地址转发表项中进行匹配,并通过匹配表项的出接口直接转发所述单播数据报文而不进行大二层网络协议封装。 MAC forwarding matching entry, and forward the unicast packet data network without large Layer protocol encapsulation matching entry directly through the interface of the publication.
14. 如权利要求13所述的网络设备,其特征在于,所述网络设备还与所述大二层网络外部的路由设备连接; 所述大二层网络中各节点设备上设置有VLAN的广播域范围,其中,所述广播域范围包括该节点设备上允许该VLAN通过的用户侧端口和该节点设备上的位于该VLAN裁剪树上的大二层网络侧接口,所述裁剪树是基于VLAN对所述广播树进行分发树的剪枝得到的; 所述报文转发模块还用于,在接收到所述路由设备发送的数据报文时,查找匹配所述数据报文的目的IP地址的ARP表项,并根据查找到的ARP表项的目的MAC地址以及出接口,直接转发所述数据报文而不进行大二层网络协议封装。 14. The network apparatus according to claim 13, wherein the network device is further connected to the external network routing devices Layer large; VLAN broadcast provided on each of the large network node apparatus Layer gamut, wherein the broadcast domain comprises a large range of Layer 2 network tree crop located on VLAN allows the VLAN specified by the user-side port of the node device of the node and the device-side interface, and the tree is cut based VLAN pruning the trees obtained broadcast distribution tree; the message forwarding module is further configured to, upon receiving the routing data packets sent by the device, searches for the destination IP address matches the data packet is ARP entry, and according to the obtained destination MAC address ARP entry and egress interface to forward the data packet directly without large network Layer protocol encapsulation.
15. 如权利要求14所述的网络设备,其特征在于, 所述报文转发模块还用于,若未查找到匹配所述数据报文的目的IP地址的ARP表项,在该数据报文对应的VLAN的广播域内,直接广播发送用于请求所述数据报文的目的IP地址对应的MAC地址的第一ARP请求报文而不进行大二层网络协议封装,所述第一ARP请求报文的发送端sender IP地址是所述网络设备的发送端口IP地址,源MAC地址及发送端MAC地址是所述网络设备的发送端口MAC地址; 所述表项维护模块还用于,当所述网络设备接收到根据接收到针对第一ARP请求报文的第一ARP应答报文时,建立所述数据报文的目的IP地址对应的ARP表项; 所述报文转发模块还用于,根据所述表项维护模块建立的ARP表项,转发所述数据报文而不进行大二层网络协议封装。 15. The network apparatus according to claim 14, wherein the message forwarding module is further configured to, if not found ARP entry matching the destination IP address of the data packet, the data packet VLAN corresponding broadcast domain, transmitting a first direct broadcast ARP request for the destination IP address of the data packet corresponding to the MAC address request message without a large network Layer protocol encapsulation, the first ARP request packet sending end packet sender IP address is the IP address of the transmission port of the network device, and a transmission source MAC address of the MAC address is the MAC address of the transmission port of the network device; the entry module is further configured to maintain, when the network device receives the ARP entry response message, establishing the packet data according to the received first ARP ARP request packet for a first destination IP address corresponding to; the message forwarding module is further configured to, in accordance with maintaining the entry module establishes the ARP entry, not to forward the data packet network Layer protocol encapsulation large.
16. 如权利要求13所述的网络设备,其特征在于,同一VLAN中的一个网关节点被选举为该VLAN的主网关节点, 所述报文转发模块还用于,当接收到主机发送的用于请求网关节点的MAC地址的第二ARP请求报文,且确定所述网络设备为对应VLAN的主网关节点时,生成第二ARP应答报文,该第二ARP应答报文的目的MAC地址为所述主机的MAC地址,源MAC地址和发送端sender MAC地址均为所述网络设备的虚MAC地址,sender IP地址为所述网络设备的虚IP地址,以使所述主机学习并建立所述网络设备的虚IP地址与虚MAC地址之间的关联关系;根据所述ARP应答报文的目的MAC地址,在第二MAC地址转发表项中进行匹配,并通过匹配表项的出接口直接转发所述ARP应答报文而不进行大二层网络协议封装。 16. The network apparatus according to claim 13, wherein a gateway node in the same VLAN is elected as master node the VLAN gateway, the message forwarding module is further configured to, when receiving transmitted by the host ARP request for the MAC address of a second gateway node to a request packet, and determining the network device is a gateway node of a corresponding main VLAN to generate a second ARP response message, the second destination MAC address for the ARP response packet the host MAC address, source MAC address and the MAC address of both the end SENDER virtual network device MAC address, IP address SENDER IP address of the virtual network device, so that the host and establish a learning virtual network device IP address and the association between the virtual MAC address; according to the destination MAC address of the ARP response packet, the second MAC address forwarding matching entry published, by directly forwarding matching entry outbound interface the ARP reply packet network without large Layer protocol encapsulation.
17. 如权利要求16所述的网络设备,其特征在于,所述网络设备还包括: 选举模块,用于在所述网络设备所属的VLAN内广播携带有所述网络设备自身端口IP地址、端口MAC地址、虚IP及虚MAC地址的邻居发现及主网关节点选举协议报文;接收所述VLAN 内其他网关节点发送的所述协议报文,选举端口IP地址和/或端口MC地址满足预设规则的网关节点,作为所述主网关节点。 17. The network device according to claim 16, wherein the network device further comprises: election means for VLAN within the network device itself belongs carrying broadcast IP address of the network device port, the port neighbor MAC address, IP address and the virtual MAC address discovery and virtual master gateway node election protocol packet; receiving said protocol within said other transmission VLAN packet gateway node election port IP address and / or port address satisfies a preset MC rule gateway node, as the primary gateway node.
18. 如权利要求17所述的网络设备,其特征在于, 所述表项维护模块还用于,在非网关节点与所述网络设备之间的链路中断时,更新达到该非网关节点下所有主机的ARP表项的出接口;在非网关节点发生宕机down时,删除该非网关节点下的所有主机的ARP表项; 所述地址管理模块还用于,在所述网络设备发现外部路由器不可达时,禁用自身所有的三层接口,并通过所述广播树通告虚MAC地址的撤销消息,以使得非网关节点设备更新所述虚MAC地址对应的所述第一MAC地址转发表项的出接口; 所述选举模块还用于,在所述网络设备发现外部路由器不可达时,退出本网关节点所在VLAN的主网关节点的选举,以触发其他网关节点重新选举主网关节点。 At 18. The network device according to claim 17, wherein the entry module is further configured to maintain, when the link between the gateway node and the non-network device interrupt, update the non-gateway node reaches All hosts an ARP entry interfaces; occurs when a non-gateway node down Down, ARP entry is deleted in all of the non-host gateway node; said address management module is further used, they were found in the external network device when the router is not reachable, disable all three interfaces itself, and notifies virtual MAC address withdraw message broadcast by the tree, so that the updating of the virtual MAC address corresponding to a non-gateway node apparatus a first MAC address forwarding entries the outgoing interface; the election module is further configured, when the external network device discovery unreachable router, gateway node exit RO Ben articulation point is located VLAN to trigger the re-election of the master gateway node to other gateway nodes.
CN201310357517.1A 2013-08-15 2013-08-15 OSPF packets forwarding method and apparatus CN104378297B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201310357517.1A CN104378297B (en) 2013-08-15 2013-08-15 OSPF packets forwarding method and apparatus

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201310357517.1A CN104378297B (en) 2013-08-15 2013-08-15 OSPF packets forwarding method and apparatus

Publications (2)

Publication Number Publication Date
CN104378297A CN104378297A (en) 2015-02-25
CN104378297B true CN104378297B (en) 2018-01-16

Family

ID=52556964

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201310357517.1A CN104378297B (en) 2013-08-15 2013-08-15 OSPF packets forwarding method and apparatus

Country Status (1)

Country Link
CN (1) CN104378297B (en)

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106209638B (en) 2015-05-04 2019-07-12 新华三技术有限公司 From virtual LAN to the message forwarding method and equipment of virtual expansible local area network
CN104869058B (en) * 2015-06-04 2018-10-19 北京京东尚科信息技术有限公司 A data packet forwarding apparatus and method
CN106302264B (en) * 2015-06-08 2019-06-14 新华三技术有限公司 Dune exchanges the unicast message interoperability methods and device between cable card board and flexible cable card board
CN106487677A (en) * 2015-09-02 2017-03-08 华为技术有限公司 Provider edge equipment and data forwarding method
CN107204907A (en) * 2016-03-16 2017-09-26 中兴通讯股份有限公司 Interconnection method and device for cloud data center
CN106059946A (en) * 2016-05-23 2016-10-26 杭州华三通信技术有限公司 Message forwarding method and message forwarding device
CN106411835A (en) * 2016-05-25 2017-02-15 海尔优家智能科技(北京)有限公司 Transmission method and apparatus based on AllJoyn Control Panel
CN106506378A (en) * 2016-11-10 2017-03-15 杭州迪普科技股份有限公司 MAC address table updating method and device
CN108075969A (en) * 2016-11-17 2018-05-25 新华三技术有限公司 Message forwarding method and device

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103078969A (en) * 2013-02-01 2013-05-01 杭州华三通信技术有限公司 Method and equipment for announcing media access control (MAC) address information
CN103227843A (en) * 2012-08-31 2013-07-31 杭州华三通信技术有限公司 Physical link address management method and device

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9363225B2 (en) * 2012-01-12 2016-06-07 Cisco Technology, Inc. Connecting layer-2 domains over layer-3 networks

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103227843A (en) * 2012-08-31 2013-07-31 杭州华三通信技术有限公司 Physical link address management method and device
CN103227757A (en) * 2012-08-31 2013-07-31 杭州华三通信技术有限公司 Message forwarding method and equipment
CN103078969A (en) * 2013-02-01 2013-05-01 杭州华三通信技术有限公司 Method and equipment for announcing media access control (MAC) address information

Also Published As

Publication number Publication date
CN104378297A (en) 2015-02-25

Similar Documents

Publication Publication Date Title
US8166205B2 (en) Overlay transport virtualization
EP2572476B1 (en) Method and apparatus for mpls label allocation for a bgp mac-vpn
US8462774B2 (en) Virtual IP interfaces on multi-chassis link aggregates
US8855117B2 (en) Scalable media access control protocol synchronization techniques for fabric extender based emulated switch deployments
US9628375B2 (en) N-node link aggregation group (LAG) systems that can support various topologies
US10225094B2 (en) SDN facilitated multicast in data center
US7558878B2 (en) Optimal sync for rapid spanning tree protocol
US9912612B2 (en) Extended ethernet fabric switches
US20120014386A1 (en) Delegate Gateways and Proxy for Target Hosts in Large Layer 2 and Address Resolution with Duplicated Internet Protocol Addresses
US20070036178A1 (en) Layer 2 virtual switching environment
US10097372B2 (en) Method for resource optimized network virtualization overlay transport in virtualized data center environments
CN103227843B (en) Kind of physical link address management method and apparatus
US8867367B2 (en) 802.1aq support over IETF EVPN
CN102986172B (en) Virtual Chassis
CN101047601B (en) Implementing method and system of double-attach network based on VPLS
US8472447B2 (en) IP multicast snooping and routing with multi-chassis link aggregation
US8953590B1 (en) Layer two virtual private network having control plane address learning supporting multi-homed customer networks
US9197583B2 (en) Signaling of attachment circuit status and automatic discovery of inter-chassis communication peers
JP5974092B2 (en) Packet Broadcast mechanism in the divided architecture network
JP2009510953A (en) Provider Link State Bridging
CN102546351B (en) System and method for interconnecting openflow network and conventional Internet protocol (IP) network
KR20140027455A (en) Centralized system for routing ethernet packets over an internet protocol network
US8694664B2 (en) Active-active multi-homing support for overlay transport protocol
CN101964746A (en) Routing frames in a shortest path computer network for a multi-homed legacy bridge node
CN101710877B (en) Pseudo wire based traffic processing method, apparatus and system for

Legal Events

Date Code Title Description
C06 Publication
EXSB Decision made by sipo to initiate substantive examination
CB02
GR01