文章目录
随着网络中部署的业务不断增长，对于全双工点对点链路，单条物理链路的带宽可能已经不能满足正常业务流量的需求，而且单条链路没有冗余备份功能，发生故障可能影响整个网络。
想要升级带宽的最直接粗暴的方式就是：换性能更高的设备，或者是具备更高带宽的接口板，但是这个只有土豪老板才能闭眼考虑的事情，不适用于普通企业，而且这样做也比较浪费现有的资源。
还有一种方法就是直接增加设备间链路的数量，但添加三层接口，则需要在每个接口上配置IP地址，导致IP地址浪费，在二层上，则可能会出现环路，如果使用生成树协议，则会被堵塞端口。
xff0c; xff0c for a full two-point interface xff0c; xff0c for a single physical link; xff0c; and xff0c for a single link without redundant backup functions xff0c; which may affect the entire network as a result of a failure.
The most straightforward way to upgrade the bandwidth is xff1a; xff0c for a higher-performance device; or xff0c for a higher-bandwidth interface; xff0c for a single owner only; xff0c for an ordinary enterprise xff0c; and it is more wasteful of existing resources.
There is also a direct increase in the number of inter-equipment links of equipment xff0c; but the addition of three layers of interfaces xff0c; the need to have an IP address xff0c for each interface; leading to waste the IP address xff0c; xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxfffffffffpp; etc.; etc.; etc.; etc.; p.; ffffffffp.; etc.
上述两种方法都不够合理，那么我们可以使用：Eth-Trunk（链路聚合技术）
None of the above methods is reasonable xff0c; then we can use xff1a; Eth-Trunk（ connection polymerization technology xff09;
可以把多个独立的物理接口绑定在一起作为一个大带宽的逻辑接口使用。
Multiple separate physical interfaces can be tied together as a logic interface with a wide bandwidth.
Eth-Trunk可以用于二层的链路聚合，也可以用于三层的链路聚合。缺省情况下，以太网接口工作在二层模式。如果需要配置二层Eth-Trunk接口，可以通过命令将该接口切换成二层接口；如果需要配置三层Eth-Trunk接口，可以通过命令将该接口切换成三层接口。
Eth-Trunk can be used for two-storey connection polymer xff0c; or for three-storey connection polymers. xff0c in default; working in two-storey mode with the Tainet interface. If you need to configure the two-storey Eth-Trunk interface xff0c; you can switch the interface to two-storey interfaces xff1b by command; you need to configure the three-storey Eth-Trunk interface xff0c; you can switch the interface to three-storey interfaces by command.
根据不同的链路聚合模式，Eth-Trunk接口可以实现增加带宽、负载分担等功能，帮助提高网络的可靠性
The Eth-Trunk interface enables increased bandwidth, load-sharing, etc. & #xff0c; helps improve network reliability
Eth-Trunk链路聚合模式分为：
The Eth-Trunk chain polymerization pattern is divided into xff1a;
当两台设备中至少有一台设备不支持LACP协议时，我们可以使用手工负载分担模式来增加设备间的带宽以及可靠性。
在手工负载模式下，加入Eth-Trunk的链路都进行数据的转发。
br/> When at least one of the two devices does not support the LACP agreement & #xff0c; we can increase bandwidth and reliability between the devices using a manual load-sharing mode.
under the manual load mode xff0c; add the Eth-Trunk chain to all data relays.
手工负载分担模式配置
创建手工负载分担模式Eth-Trunk：
Create manual load-sharing mode Eth-Trunk：
-
在系统视图中执行命令，创建Eth-Trunk接口并进入Eth-Trunk接口视图。
Execute command & #xff0c in the system view; create the Eth-Trunk interface and enter the Eth-Trunk interface view.
-
执行命令，将Eth-Trunk接口切换为二层模式。（默认为二层模式）
执行命令，将Eth-Trunk接口切换为三层模式。Executing command & #xff0c; Switching the Eth-Trunk interface to 2-storey mode. & #xff08; Default to 2-storey mode & #xff09;
Executing command & #xff0c; Switching the Eth-Trunk interface to 3-storey mode. -
配置Eth-Trunk的工作模式：配置当前Eth-Trunk工作模式为手工负载分担模式。 （默认为手工负载分担模式）
Configure Eth-Trunk work mode & #xff1a; configure the current Eth-Trunk work mode for manual load-sharing mode. & #xff08; default for manual load-sharing mode xff09;
Eth-Trunk中加入成员接口：
Add a membership interface to Eth-Trunk & #xff1a;
方法一：在Eth-Trunk视图下，
Method I & #xff1a; in Eth-Trunk view & #xff0c;
- 执行命令，批量增加成员接口。
- 执行命令，增加一个成员接口。
在成员接口视图下：
Under the membership interface viewxff1a;
- 进入想要加入Eth-Trunk的接口视图，执行命令，将当前接口加入Eth-Trunk。
LACP模式也称为M:N模式，其中M条链路处于活动状态转发数据，N条链路处于非活动状态作为备份链路。
>LACP model also known as M: N-mode & xffoc; M-links are transmitting data #xff0c; N-links are not active as backup links.
图中设置的活跃链路数为2，即2条链路处于转发状态，1条链路处于备份状态，不转发数据，只有当活跃的链路出现故障时，备份链路才进行转发。
The number of active links set in the figure is 2， that is, 2 links are in re-transmission status xff0c; 1 link is back-up status xff0c; data xff0c is not re-transmitted; xff0c is re-transmitted only if the active links fail; and the backup links are re-transmitted.
LACP模式活动链路选取
如上图：设备之间相连的链路为3条，假如要求最大活跃链路为2，1条链路为备份状态。
br/> as above xff1a; chain links between equipment are 3 #xff0c; if the maximum active link is required to be 2xff0c; 1 link is a backup state.
在LACP模式的Eth-Trunk中加入成员接口后，这些接口将向对端通告自己的系统优先级、MAC地址、接口优先级、接口号等信息。对端接收到这些信息后，将这些信息与自身接口所保存的信息比较以选择能够聚合的接口，双方对哪些接口能够成为活动接口达成一致，确定活动链路。
xff0c after adding a member interface to the LACP model Eth-Trunk; these interfaces will inform each end of information about its system priorities, MAC addresses, interface priorities, interface numbers, etc. After receiving this information at the end xff0c; comparing this information with information stored in its own interfaces to select interfaces that can be aggregated xff0c; agreeing on which interfaces can be an activity interface xff0c; determining the activity links.
两端设备会选取主动端，根据以下条件选举：
The two ends of the device will select the
交换机接口会对接口进行排序，根据以下条件选举：
The switch interface will sort the interfaces xff0c; xff1a is selected in accordance with the following conditions;
接口LACP优先级是为了区别同一个Eth-Trunk中的不同接口被选为活动接口的优先程度，优先级高的接口将优先被选为活动接口。
注意：只通过比较主动端的各接口接口优先级确定活跃链路
The interface LACP priority is intended to distinguish between different interfaces in the same Eth-Trunk that are selected as activity interface priority xff0c; high priority interfaces are selected as activity interfaces.
Note xff1a; active links are determined only by comparing the active interface priority of each interface at the active end
LACP抢占机制
如上图：若LACP开启抢占机制，假如一条主链路发生故障，此时备用链路就会被启用，进行数据转发。假设过了一段时间之后发生故障的主链路回复正常，则端口开始抢占，在抢占延时超时后，成为活跃端口，进入转发状态。备份链路不再转发数据。
br/> as above xff1a; if LACP opens up the mechanism #xff0c; if the transmission of the main link fails #x0; #xxxpwpwpwppppppppppppt ; #ffpwpwpwpwpppppppppppppt ; #0xxfffffppppppppppppppppppppppppppppppppppt.; #fffffffffffpppppppppppppppppppppppppttttttttttttt.x0; fffffffffftppppptpppppppppppppppppt.x.fffffffffft.tppppppppppptt.t. ffffffffffffft.t.
为什么要设置抢占延时：
避免由于某些原因链路状态频繁变化而导致的链路聚合数据传输不稳定。（主动链路恢复后，立马抢占，然后突然又故障，有恢复。在恢复和故障之间来回徘徊）
& #xff0c; #xff0c; then suddenly fail xff0c; there is recovery. br/> Avoids instability in the transmission of chain polymer data due to frequent changes in chain status for some reasons. & #xff08; xff0c when active links are restored; xff0c; xff0c immediately; and xff0c suddenly; and xff09;
注意：
Note #xff1a;
- 若未开启抢占机制，则活跃端口故障恢复之后，不抢占，不会重新成为活跃端口，而是成为备份端口。
- 若希望手动修改端口优先级来指定原本优先级低的端口成为活跃端口，则需要先开启抢占功能，不然调高了优先级也不会切换成活跃端口。
LACP模式配置
创建LACP模式Eth-Trunk：
Create the LACP model Eth-Trunk：
- 执行命令，创建Eth-Trunk。
- 执行命令，将Eth-Trunk接口切换为二层模式。（默认为二层）
配置Eth-Trunk的工作模式：
Configure the Eth-Trunk mode of workxff1a;
- 在Eth-Trunk接口视图。执行命令，配置Eth-Trunk的工作模式为LACP模式。
Eth-Trunk中加入成员接口：
Add a membership interface to Eth-Trunk & #xff1a;
- 与上文手工负载分担相同。
Eth-Trunk接口进行负载分担时，可以选择IP地址或者包作为负载分担的散列依据；同时还可以设置成员接口的负载分担权重。
At load sharing for the Eth-Trunk interface & #xff0c; you can select an IP address or package as the hash basis for load sharing & #xff1b; and you can also set load sharing weights for the member interface.
散列依据
| 接口负载分担 | 特点 |
|---|---|
| 逐流负载分担 | 当报文的源IP地址、目的IP地址都相同或者报文的源MAC地址、目的MAC地址都相同时，这些报文从同一条成员链路上通过。 |
| 逐包负载分担 | 以报文为单位分别从不同的成员链路上发送。 |
两种散列依据造成的问题：
Problems arising from the two hash bases xff1a;
假如主机A向主机B发送一个数据比较大的100个包。
If host A sends a larger 100 packages of data to host B.
- 逐流负载分担的处理方式是：这100个包都从一条物理链路发送。这时可能会造成一条物理链路负载较大，一条物理链路空闲。
- 逐包负载分担的处理方式是：可能第1，3，5…个包从一条物理链路发送；第2，4，6…个包从另一条物理链路发送。此时数据包可能不能按顺序到达目的端，可能会造成数据乱码。
配置命令
Configuration Command
进入Eth-Trunk接口视图，执行命令，配置Eth-Trunk接口的散列依据。（缺省情况下，当Eth-Trunk接口根据IP进行散列。）
Enter the Eth-Trunk interface view & #xff0c; execute command & #xff0c; configure the hash basis for the Eth-Trunk interface. & #xff08; default & #xff0c; when the Eth-Trunk interface is based on IP hash. & #xff09;
负载分担权重
配置成员接口的负载分担权重，某成员接口的权重值占所有成员接口负载分担权重之和的比例越大，该成员接口承担的负载就越大。
Configure the load-sharing weight of the member interface xff0c; the greater the weight of a member interface as a proportion of the sum of the burden-sharing weight of all member interfaces xff0c; the greater the load of the member interface.
配置命令
Configuration Command
进入以太网接口视图。执行命令，配置Eth-Trunk成员接口的负载分担权重。（缺省情况下，成员接口的负载分担权重为1。）
Enter the Ethernet interface view. Execute command xff0c; configure load-sharing weights for Eth-Trunk member interfaces. xff08; default xff0c; load-sharing weights for member interfaces are 1.xff09;
将成员接口加入Eth-Trunk时，需要注意以下问题：
Adding the membership interface to Eth-Trunk & #xff0c; attention needs to be paid to the following issues & #xff1a;
- 成员接口不能有IP地址等三层配置项，也不可以配置任何业务；
- 成员接口不能配置静态MAC地址；
- Eth-Trunk接口不能嵌套，即成员接口不能是Eth-Trunk；
- 一个以太网接口只能加入到一个Eth-Trunk接口，如果需要加入其他Eth-Trunk接口，必须先退出原来的Eth-Trunk接口；
- 如果本地设备使用了Eth-Trunk，与成员接口直连的对端接口也必须捆绑为Eth-Trunk接口，两端才能正常通信；
- Eth-Trunk有两种工作模式：二层工作模式和三层工作模式。Eth-Trunk的工作模式不影响成员链路的加入，例如，以太网接口既可以加入二层模式的Eth-Trunk，也可以加入三层模式的Eth-Trunk。
实验要求：配置链路聚合，使该拓扑链路增加带宽以及提高网络可靠性。核心层通过三层配置实现，汇聚层与接入层通过二层配置实现。
Experimental requirements xff1a; configuration of chain polymers xff0c; increasing bandwidth and network reliability of the puncture links. Core layer xff0c through three-storey configurations; assembly layer and access layer through two-storey configurations.
.
步骤一
搭建拓扑，配置PC的IP地址、子网掩码、网关。
Set-up of #xff0c; configures PC's IP address, subnet mask, gateway.
步骤二
核心层配置，以AR1为例：
Core layer configuration xff0c; use AR1 as an example xff1a;
验证：我们可以通过 来查看链路聚合简略消息。也可以通过来查看具体信息。
Validate & #xff1a; we can look at the links aggregates brief messages. We can also look at specific information.
步骤三
核心层汇聚层交换机使用二层互联，所以不需要配置IP地址。核心层与汇聚层交换机以LSW5为例：
Nucleus congregate switches use two layers of interconnectivity & #xff0c; therefore, IP addresses are not required. The Nucleus and Convergence layer switches take LSW5 as an example & #xff1a;
配置验证：
Configure authentication & #xff1a;
注册有任何问题请添加 微信：MVIP619 拉你进入群
打开微信扫一扫
添加客服
进入交流群
发表评论