Learning: the Ethernet switch understands the MAC address of the device connected to each port, maps the address with the corresponding port, and stores it in the MAC address table in the switch cache.

Forwarding / filtering: when the destination address of a data frame is mapped in the MAC address table, it is forwarded to the port connecting the destination node instead of all ports (if the data frame is a broadcast / multicast frame, it is forwarded to all ports).

Elimination of loop: when the switch includes a redundant loop, the Ethernet switch avoids the generation of loop through spanning tree protocol, and allows the existence of backup path.

In addition to being able to connect the same type of network, switches can also interconnect different types of networks (such as Ethernet and fast Ethernet). Nowadays, many switches can provide high-speed connection ports supporting fast Ethernet or FDDI, which can be used to connect other switches in the network or provide additional bandwidth for key servers with large bandwidth consumption.

Generally speaking, each port of the switch is used to connect an independent network segment, but sometimes in order to provide faster access speed, we can connect some important network computers directly to the port of the switch. In this way, the key servers and important users of the network will have faster access speed and support greater information flow.

Switch mode

The switch can be exchanged in the following three ways:

1) Straight through:

Through Ethernet switch can be understood as a vertical and horizontal line matrix telephone switch between ports. When a packet is detected at the input port, it checks the packet header, obtains the destination address of the packet, starts the internal dynamic look-up table to convert to the corresponding output port, connects at the intersection of input and output, and sends the packet directly to the corresponding port to realize the switching function. Because there is no need for storage, the delay is very small and the exchange is very fast, which is its advantage. Its disadvantage is that the packet content is not saved by the Ethernet switch, so it is unable to check whether the transmitted packet is wrong, and it can not provide error detection capability. Because there is no buffer, I / O ports with different rates can not be connected directly, and it is easy to lose packets.

2) Store and forward:

Store and forward is the most widely used way in the field of computer network. It stores the data packet of the input port first, and then checks the CRC (cyclic redundancy check). After processing the error packet, it takes out the destination address of the data packet and converts it into the output port to send the packet through the look-up table. Because of this, the store and forward mode has a large delay in data processing, which is its disadvantage. But it can detect the error of the packets entering the switch, and effectively improve the network performance. In particular, it can support the conversion between different speed ports and keep the cooperation between high speed port and low speed port.

3) Fragment isolation:

This is a solution between the first two. It checks whether the length of the packet is 64 bytes. If it is less than 64 bytes, it means that it is a fake packet, and then it discards the packet; if it is more than 64 bytes, it sends the packet. This method also does not provide data verification. Its data processing speed is faster than store and forward mode, but slower than straight through mode.

Switch application

As the main connection equipment of LAN, Ethernet switch has become one of the most popular network equipment. With the continuous development of switching technology, the price of Ethernet switch drops sharply, and switching to the desktop is the general trend.

If you have a large number of users, busy applications and various servers on your Ethernet network, and you haven't made any adjustments to the network structure, the performance of the whole network may be very low. One of the solutions is to add a 10 / 100Mbps switch to the Ethernet, which can not only handle 10Mbps conventional Ethernet data stream, but also support 100Mbps Fast Ethernet connection.

If the utilization rate of the network exceeds 40% and the collision rate exceeds 10%, the switch can help you solve some problems. The switch with 100Mbps Fast Ethernet and 10Mbps Ethernet port can operate in full duplex mode, and can establish a dedicated 20MBps to 200Mbps connection.

Not only the functions of switches are different in different network environments, but also the effects of adding new switches and increasing the switching ports of existing switches are different in the same network environment. Fully understanding and mastering the network traffic mode is a very important factor to play the role of switch. Because the purpose of using the switch is to reduce and filter the data flow in the network as much as possible, if a switch in the network needs to forward almost all the received data packets due to improper installation location, the switch can not play its role in optimizing the network performance, but reduces the data transmission speed and increases the network delay.

In addition to the installation location, if the switch is blindly added to the network with less load and less information, it may also have a negative impact. Affected by the processing time of packets, the buffer size of the switch and the need to regenerate new packets, it is better to use a simple hub than the switch in this case. Therefore, we can not all think that the switch has advantages over hub, especially when the user's network is not crowded and there is still a lot of available space, the use of hub can make full use of the existing resources of the network.