1. Principle of Ethernet Switch
Ethernet switches, as the most widely used LAN hardware devices today, have always been familiar to everyone. Its popularity is actually due to the widespread use of Ethernet, as today's mainstream Ethernet equipment, there will be such equipment in almost all local area networks. Looking at the following topology, you will find that in the case of using a star topology, there must be a switch in the Ethernet, because all hosts are connected to the switch using cables to connect to each other:
In fact, in the earliest star topology, the standard cable centralized connection device is a "HUB (hub)", but hubs have problems such as shared bandwidth, conflicts between ports, because everyone knows that the standard Ethernet is a "hub". Conflict network" means that in a so-called "conflict domain", at most two nodes can communicate with each other. Moreover, although the hub has many ports, its internal structure is completely the so-called "bus structure" of Ethernet, which means that there is only one "line" inside for communication. If the device in the above figure is a hub, for example, if the nodes between ports 1 and 2 are communicating, other ports need to wait. The directly caused phenomenon is, for example, it takes 10 minutes to transmit data between the nodes connected to ports 1 and 2, and the nodes where ports 3 and 4 are located at the same time also start to transmit data through this hub, conflicts with each other, causing what everyone needs It will take longer, and it may take up to 20 minutes to complete the transmission. That is to say, the more ports on the hub that communicate with each other, the more serious the conflict, and the longer it takes to transmit data.
This kind of problem does not cause a big problem in a small Ethernet, and it can work very well, but if the traffic on the network increases or the number of connected nodes is large, the "conflict" will seriously affect the network. Performance. For example, when we explained the principle of Ethernet in the first chapter, we explained the problem of optimizing the "conflict domain". At this time, we need to be able to isolate the "conflicting" equipment, and the switch can complete this function.
When the switch is connected, each port can communicate at the same time, which means that there is no conflict between the ports, and it can also be used to isolate conflicts. So, what kind of principle causes the switch to achieve this capability? Let's take a look at the picture below: