The Requirement of EtherChannel
Mission-critical apps need network structures that are strong. Now that there are a lot of faster switched Ethernet links on campus, users need to either combine the resources they already have or speed up their uplinks and core to make the network backbone work better.
See Below diagram, At 100 Mbps, traffic from several VLANs gathers on the access switches at the bottom and needs to be sent to the distribution switches in the middle. Obviously, the link between two switches needs to have bandwidth greater than 100 Mbps so that it can handle all the data from the VLANs. Use a faster port speed, like 1 or 10 Gbps, as a first step. As the speeds of the VLAN links get faster, this solution can only handle data from the fastest port for a certain amount of time. It then stops working. This can be fixed by adding more physical links between the two switches. This will make the connection between the switches faster overall. The only bad thing about this method is that each physical link setup has to be exact. That being said, spanninfg tree might get in the way & can block one of the links, as seen in below diagram:-
Originally created by Cisco as a way for LAN switches to communicate with one another, EtherChannel allows several Fast or Gigabit Ethernet ports to be combined into a single logical channel. The advantages of this technique are numerous:
- It uses the ports that are already on the switch. Investing in a faster and more costly switch-to-switch connectivity is unnecessary.
- For consistency in configuration across switch-to-switch links, most configuration operations can be done on the EtherChannel interface rather than on each individual port.
- The links that make up an EtherChannel can work together to distribute the load evenly. One or more approaches, such as source-IP to destination-IP or source-MAC to destination-MAC load balancing across physical links, can be implemented, depending on the hardware platform.
It is important to keep in mind that the purpose of EtherChannel is to boost the speed between switches. As the EtherChannel technology gained popularity, this notion was expanded upon. Additionally, many hardware devices that do not have switch capabilities facilitate the aggregation of links into an EtherChannel link setting. Whatever the case may be, EtherChannel establishes a relationship that is one-to-one. Creating an EtherChannel link between two switches or between an EtherChannel-enabled server and a switch is possible; however, it is not possible to deliver traffic to two separate switches using the same EtherChannel link. Whenever there is only one EtherChannel link, it will only connect the same two devices. The port configuration of each individual member of the EtherChannel group must be same on each of the devices. The technology known as EtherChannel only bundles ports that are of the same type. When access ports or trunks are aggregated on a Layer 2 switch, EtherChannel is the protocol that is utilized. One example would be that if the physical ports on one side are configured as trunks, then the physical ports on the other side must also be configured as trunks. There is a logical port channel interface associated with each EtherChannel. This means that any configuration that is done to the port channel interface will have an effect on all of the physical interfaces that are associated with that interface. STP commands or commands to setup a Layer 2 EtherChannel as a trunk or an access port are examples of the kind of commands that fall under this category.
Better redundancy can be achieved by creating a port channel between two aggregation switches from the same access layer using modern technologies like Virtual Switching System (VSS) and Virtual Port Channel (vPC).
Remember that EtherChannel forms what is viewed as a single logical connection in the aggregation. In order to avoid redundant links, spanning trees may block one of the multiple EtherChannel bundles that are present between two switches. All of the ports associated with a particular EtherChannel connection are blocked when a spanning tree blocks one of the redundant links. All physical links in an EtherChannel with a single EtherChannel link are active since the spanning tree only perceives the single (logical) link in such cases. The bandwidth of the EtherChannel and, thus, the STP cost, will automatically adjust in the event that one of its links fails.
Interactions between EtherChannel Modes
One of the three ways shown in below figure can be used to set up an EtherChannel:
- LACP: The Protocol for Negotiation by the IEEE
- PAgP: denotes Cisco’s protocol for negotiations.
- Static persistence: There is no negotiation protocol for static persistence.
LACP
Another part of the IEEE standard (802.3ad) is the Link Aggregation Control Protocol (LACP), which lets several physical ports work together as one virtual channel. When a switch sends LACP packets to a peer, it can arrange an automatic bundle. You can use LACP to make EtherChannels work in mixed-switch settings since it is an IEEE standard. LACP handles link additions and failures between two switches and makes sure that the configuration stays the same. It makes sure that all ports have the same speed, duplex setting, and VLAN information when EtherChannel is made. If you change one port after the channel has been created, it will also change all the other ports in the channel.
Switches send and receive LACP messages over ports that can handle Ethernet. The powers of each port are learned and compared to those of the local switch. LACP gives EtherChannel’s ports roles. That choice about which ports are active in EtherChannel is made by the switch with the lowest system priority. Based on their port priority, ports become active. A lower number means it’s more important. Of course, an EtherChannel can have up to 16 links, but only 8 can be active at any given time. Links that aren’t being used are put into a standby state. If a live link goes down, the standby links become active again.
Each switch has a different maximum number of active links in an EtherChannel.
The following are the modes of operation for the LACP:
- Active: Enable LACP
- Passive: Enable LACP only if an LACP device is detected
You can utilize the additional options listed below when establishing LACP:
System priority: The priority of the system is something that every switch that is running LACP needs to comply with. The CLI can be used to specify the system priority. In order to generate the system ID, the switch takes into account both the MAC address and the system priority.
Port priority: Each port in the switch is required to have a port priority in order to function properly. The CLI can be used to specify the port priority. The identification of the port is comprised of the port number as well as the port priority. When there is a limitation in the hardware that prohibits all compatible ports from aggregating, the switch will use the port priority to determine which ports will be placed in standby mode.
Administrative key: The administration key is a value that must be assigned to each port in the switch. This value can be given automatically or through the command line interface using the administrative key. The administrative key is what determines whether or not a port is able to aggregate with other ports. This capability is defined by the physical features of the port, which include parameters such as data rate, duplex capability, and whether or not the medium is shared or point-to-point.
