Link Aggregation - Introduction to TRUNK Technology.

CCNA 200-301

CCNA 200-301

CCNP Enterprise

CCNP Enterprise

CCNP Security

CCNP Security

CCIE Enterprise Lab

CCIE Enterprise Lab

CCIE Security Lab

CCIE Security Lab

CCNP Service Provider

CCNP Service Provider

CCNP Data Center

CCNP Data Center

CCNP Collaboration

CCNP Collaboration

CCIE DC Lab

CCIE DC Lab

ic_r
ic_l
Link Aggregation - Introduction to TRUNK Technology.
images

I. Overview

With the continuous development and application of network technology, the speed of the network is getting faster and faster, and the application of the network is more and more complicated. Therefore, in many practical applications, the network speed becomes the bottleneck of various network applications. Increasing network speed through upgrades is an effective solution to the problem, such as from 10M Ethernet to 100M Ethernet to 1000M Ethernet or higher.

However, this method is expensive and requires more equipment. For enterprises and organizations with insufficient funds, how to use existing resources to improve network speed is also an urgent problem to be solved.

As the saying goes, everyone collects firewood high, and Trunk technology is to use multiple links together as a link to increase network speed and meet user needs. This is important in the application of small and medium-sized enterprise switches, and several links can be bundled to complete high-speed transmission tasks without additional investment. And you can also configure the transfer rate according to actual needs. Without the network upgrade, the rate can only be increased by an order of magnitude and cannot be flexibly configured as needed. This is of great value in flexible and efficient high-speed communication between switches and switches, between switches and servers.

Second, concepts and functions

Trunk technology, also known as port aggregation (port Aggregate), link aggregation (multiple link aggregation). The essence is to bind multiple Ethernet ports together as a logical link.

As the user uses this logical link, it seems to use a separate physical link, but the bandwidth of this logical link has increased linearly (for example, if there are four 100M physical ports converged into one trunk, then this The logical link bandwidth is 800M).

Pay attention when configuring the trunk:

1. Before forming a trunk, you must make the ports in the group the same type (that is, the same transmission rate is required to be 10M, 100M, and both work in full duplex, and the connected media should be the same).

2. The port members that become a trunk should be in the same spanning tree group. At this point, the spanning tree treats the trunk as a spanning tree port.

3. A port member of a trunk should belong to the same VLAN member.

The advantages of Trunk are:

1, can quickly and effectively linearly increase the bandwidth (when conditions permit, you can determine how fast the transmission rate needs to determine how many ports form a trunk according to the needs of the application)

2. The redundancy of the link can be provided. If one of the ports in the aggregation fails, the port can be modified. For example, when the port is the root port of the trunk, the port is disabled in the spanning tree and is in the remaining trunk port. Reselect a port as the root port to ensure the system is running properly. But in the user's opinion, the system has not changed.

3. Load balancing can be implemented on ports in a trunk group.

Third, an example

1, high-speed connection between switches

The use of trunk to form a link between the switch and the switch enables high-speed communication connections between workstations connected to the two switches, avoiding communication bottlenecks caused by connecting one switch to two switches. And using trunk can also configure trunk according to the requirements of the application.

For example, if the communication speed between the two switches is 600M, then you can select three links (or four) to become a trunk to meet the demand. If you use four links to form a trunk, 600M traffic will be in these four. Load balancing is achieved on the road. At the same time, if one link fails or fails, the link will be automatically deleted from the trunk, and the traffic will be load balanced in the remaining links, and the user will not notice it. Only at this time the communication rate will drop by 200M.

2. High-speed connection between the switch and the server

In the high-speed connection between the switch and the server, the application principle and method are the same as the application between the switches.

In this example, server 1 is configured with a multi-port NIC that is connected to the switch through a trunk. Server 2 connects to the server through two network cards.

3, both have the situation

In this example, server 1 is connected to the switch through T5, and server 2 is connected to the switch through two network cards. SW1 and SW2, SW3, and SW4 are connected by T1, T2, T3, and T4. T1, T2, T3, T4 and T5 provide maximum bandwidth for each PC to access data from the server.

In this example, if spanning tree is allowed and trunk T1 and T2 are in the same spanning tree group, one of the trunks (T1 or T2) becomes a backup trunk for SW2. If spanning tree is disabled, trunk T1 and T2 must be configured in separate VLANs to keep the device functioning properly.

The trunk connection between switches and the trunk connection between the switch and the server are different. Applications can be done by selecting switch options or server options.

Fourth, we will support

1. Port-based Trunk principle:

According to the hardware function of the switch chip, combined with the actual application, the hardware register can be read and written to meet our application requirements. In this way, it is written to the hardware registers after it is configured. This method is static. It requires two switches to be configured correctly, and the connection is correct to work properly.

Implemented features:

Can implement trunk configuration;

Load balancing is implemented in each port in the trunk;

Support link backup;

Automatic reconfiguration after link failure can be achieved;

Enables trunk connections between switches and switches and servers;

Configuration:

In a port-based trunk configuration, you need to specify the ID number of the trunk. What are the port members in the trunk? The status of the trunk is allowed or disabled and is the connection between the switches or the connection between the switch and the server.

Issues to be aware of: Since port-based configuration is limited by hardware, there are certain requirements when configuring the trunk. For example, when we use the switch chip of all ports of aller125, there are only two trunks at most, and each trunk consists of 2 to 4 ports. Moreover, each trunk can only be composed of members of the first four or the last four ports, for example, it can only be composed of 1, 2, 3, 4 or 5, 6, 7, 8 instead of 2, 3, 6 Make up a trunk.

In a switch composed of Brodcom 5600 switch chips, each switch chip can be composed of 6 trunks, each trunk consists of 2 to 8 ports, but the members in the port can be arbitrarily combined without limitation.

2. Protocol-based Trunk

Principle: Protocol-based Trunks are implemented according to the requirements of the 802.3ad protocol. The main function is to run the LACP protocol on both sides of the trunk connection. The two parties exchange their own information and status with each other by sending LACPDU protocol packets to each other to finally reach a consensus. This approach is dynamic and can be dynamically configured and reconfigured. Moreover, certain error correction functions can be implemented when there is an error in the link.

Implemented features:

Implement trunk configuration;

Automatic load balancing;

Implement link backup function;

Automatic and fast configuration and reconfiguration;

Implement trunk connections between switches and switches to servers;

Configuration:

The configuration in the protocol-based trunk is basically the same as the configuration of the port-based trunk.

However, the user can select the time when the LACP protocol packet is periodically sent and the timeout period of receiving the LACPDU. These parameters are used to determine the timeout period during configuration and the detection time when a link error occurs. In addition, the priority of the system can be set to determine which system is the active party, and that system is passive to ensure the correct operation and automatic reconfiguration of the protocol. In addition, you can also specify some restrictions on the trunk application by setting the key value, such as limiting the maximum number of members of the trunk or the selection requirements of the constituent members.

Problems to be aware of: In the protocol-based trunk, there is no number of trunks and the number of ports included in the trunk, so you can arbitrarily select ports to form a trunk.

 Wana to pass your Cisco certification fast? SPOTO will be your best choice.