October 08, 2023
Link aggregation, redundancy, stacking, hot backup and other functions of main switch are very important, they determine performance, efficiency, stability, etc. of main switch in practical applications. Let's figure this out together.
1. Link Aggregation
is for combining two or more data channels into a single channel, which is displayed as one higher bandwidth logical channel. Link aggregation is typically used to connect one or more devices with high bandwidth requirements, such as servers or groups of servers connected to a backbone network. It can be used to increase bandwidth of channel and improve reliability of connection.
1. Example
The company consists of two floors, each of which has different businesses. Initially, networks of two floors are separated, but business contacts between same company are inevitable. At present, we can open network up to two floors, so that high-speed communication between departments with interconnectedness.
As shown below:
As shown in figure above, SwitchA and SwitchB are connected to VLAN10 and VLAN20, respectively, via Ethernet links, and relatively large data traffic flows between SwitchA and SwitchB. Users hope that more bandwidth can be provided between SwitchA and SwitchB so that same VLAN can communicate with each other. At same time, users also hope to provide a certain degree of redundancy to ensure reliability of data transmission and communication channels.
Create an Eth-Trunk interface and add member interfaces to increase throughput of link. Configure Eth-Trunk1 on two switches respectively. Add three line ports that need to communicate with Eth-Trunk1, set trunk port, and enable corresponding vlan pass; In this way, network on two floors can communicate normally.
2. Follow setup steps:
Create Eth-Trunk1 on SwitchA and configure it in LACP mode. SwitchB configuration process is same as SwitchA configuration process and does not repeat system-view [HUAWEI] sysname SwitchA [SwitchA] interface eth-trunk 1 [SwitchA -Eth-Trunk1] mode lacp [SwitchA-Eth-Trunk1] quit Configure member interfaces on SwitchA to attach to Eth-Trunk. The setup process for SwitchB is similar to that for SwitchA, so we won't go into details [SwitchA] interface gigabitethernet 0/0/1 [ SwitchA-GigabitEthernet0/0/ 1] eth-trunk 1 [ SwitchA-GigabitEthernet0/0/1] quit [SwitchA] interface gigabitethernet 0/0/2 [ SwitchA-GigabitEthernet0/0/2] eth-trunk 1< br>[ >SwitchA-GigabitEthernet0/0/2] quit [SwitchA] interface gigabitethernet 0/0/3 [ SwitchA-GigabitEthernet0/0/3] eth -trunk 1 [ SwitchA -GigabitEthernet0/0/3] quit Set SwitchA's system priority to 100, making it active end of LACP [SwitchA ] lacp priority 100 In Set active interface high threshold on SwitchA to 2 [SwitchA] interface eth-trunk 1 [SwitchA-Eth-Trunk1] max active-linknumber 2 [SwitchA-Eth-Trunk1 ] quit Set interface priority on SwitchA to determine active link [SwitchA] interface gigabitethernet 0/0/1 [ SwitchA-GigabitEthernet0/0/1] lacp priority 100 [ SwitchA-GigabitEthernet0/0/1] quit [SwitchA] interface gigabitethernet 0/0/2 [< br>SwitchA-GigabitEthernet0/0/2] lacp, priority 100< br>[ SwitchA-GigabitEthernet0/0/2] exit
Second, link redundancy
In order to maintain network stability, in a multi-switch network environment, some redundant connections are generally used to improve network efficiency and stability. Redundant connections are also referred to here as redundant or redundant connections. road.
3. Switch Stack
Connected with a special stacking cable, multiple switches can be combined into a logical switch. All switches in a logical switch use same configuration and routing information. Adding and removing individual switches from a logical switch will not affect its performance.
Stacked switches are connected by two loops. The switch hardware is responsible for load balancing data packets on dual rings. The loop acts as backplane of this large logical switch. When both loops are operating normally, data packet rate on this logical switch is 32Gbps.
When a data frame needs to be sent, switch software will calculate which ring is more available and then data frame will be sent to ring. If stack cable fails, switches at both ends of fault will detect fault and shut down affected ring, while logical switch can still operate in a single ring state and data packet rate is 16Gbps at that time. The switch stacking method uses daisy chain method, and connection method is shown in following figure.
Stacking improves switch port stability and throughput.
Fourth, hot backup (HSRP)
The main switch is core and heart of entire network. If a fatal failure occurs in main switch, local network will be paralyzed, and damage caused is incalculable. Therefore, when we choose core switches, we often see that some core switches have features such as stacking or hot backup.
Using hot standby for core switches is an inevitable choice to improve network reliability. When primary switch fails altogether, another backup router in system takes over all of its functionality until problematic router returns to normal. This is a hot backup routing protocol.
The condition for implementing HSRP is that there are several primary switches in system that form a "hot standby group" that forms a virtual router. At any given time, only one router in group is active, and it sends data packets. If active router fails, backup router will be selected instead of active router, but from point of view of hosts on this network, virtual router is unchanged. In this way, host stays connected and is not prone to failure, which better solves problem of switching main switch.
To reduce network traffic, after configuring Active Primary Switch and Standby Primary Switch, only Active Primary Switch and Standby Primary Switch will send HSRP packets regularly. If active primary switch fails, standby primary switch becomes active primary switch. If standby main switch fails or becomes active main switch, another main switch will be selected as standby main switch.
1. If link from access layer switch to main core switch fails, switch to backup machine and direction of data flow
When data link connecting access layer 1 switch to core switch A fails, data link of access layer 1 switch switches to core switch B, but access layer 1 switch loses 6 packets during switch, as shown in figure higher.
If main communication link between server and main switch A fails (for example, lines, NICs, etc.) and server's main NIC switches to backup NIC, 6 data packets will be lost. , but when primary link is restored, server will automatically switch from backup NIC to primary NIC, and no data packets will be lost during this switchover. The data flow direction of a particular terminal access server is shown in figure below.
