The second layer switches the end-to-end data exchange based on the MAC address of the second layer of data link layer and routing through the station table. Because the station table is built and maintained by the switch automatically, and the router is a third-tier device, its addressing process is based on IP address and through the routing table and routing protocol generated. Therefore, the biggest advantage of the second layer of the switch is the speed of data transmission, because it only needs to recognize the MAC address in the data frame, and directly based on the MAC address to generate a choice to the port of the algorithm is very simple, very easy to use ASIC dedicated chip implementation. Obviously, the second layer of the switch solution, is actually a "everywhere exchange" of the cheap solution, although the scheme can also be divided into sub-networks, limit broadcasting, build VLANs, but it has a small control capacity, flexibility is not enough, and can not control the flow of information points, lack of pan-convenient and practical routing capabilities.
The third layer switch is to complete the end-to-end data exchange directly according to the third layer network layer IP address. On the surface, the third layer switch is the forefather of the second layer exchanger and router, however, this combination is not a simple physical combination, but a logical combination of each director. The important manifestation is that when the first data stream of a source of information is exchanged in the third layer, the routing system will generate a mapping table of MAC address and IP address, and store the table, when the subsequent traffic of the same information source enters the switching environment again, the switch will be based on the address mapping table generated and saved for the first time. Direct from the second layer from the source address to the destination address, no longer through the third routing system processing, so as to eliminate the network delay caused by routing, improve the efficiency of packet forwarding, to solve the transmission of information between the network routing speed bottleneck. So, the third layer switch can not only complete the port switching function of the second layer switch, but also complete the routing function of some routers. The switch scheme of the third layer switch is actually a solution that can support the multi-level dynamic integration, although this multi-level dynamic integration function can be carried out by traditional routers and the second layer switch in some degree, but this kind of carrying solution not only need more equipment configuration, Take up more space, design more cabling and cost more, and data transfer performance is much worse, because in the massive data transmission, the router in the carrying scheme cannot overcome the bottleneck of routing transmission rate.
Obviously, the second layer switch and the third layer switch are based on the port address of the end-to-end exchange process, although this based on MAC address and IP address of the switch technology, can greatly improve the data transfer rate between the nodes, However, it is not possible to determine or dynamically restrict the port switching process and data traffic according to the application requirements of the port host, that is, the fourth level of intelligent application switching needs. The fourth layer switch can not only complete the end-to-end switching, but also can determine or limit its switching traffic according to the application characteristics of the port host. Simply put, the fourth layer switch is based on the Transport Layer Packet exchange process, is a kind of TCP/IP protocol application layer based on the user application Exchange requirements of the new LAN switch. The fourth layer switches support all protocols below the fourth layer, identify at least 80 bytes of packet header length, and differentiate the application type of the packet according to the TCP/UDP port number, thus enabling access control and service quality assurance for the application layer (TCP/UDP). So, instead of the fourth layer switch is the hardware network equipment, it is rather said that it is a software network management system. In other words, the fourth layer switch is a kind of software technology-based, hardware technology as a supplement to network management switching equipment.
Finally, it is worth pointing out that some people to some extent there are some vague concepts, that the so-called fourth-layer switch is actually added on the third layer of the switch with the ability to identify the fourth protocol port, only on the third layer of the switch to add some value-added software, and therefore not working in the transport layer, But still in the third layer on the exchange operation, but is more sensitive to the third layer of exchange, fundamentally denying the fourth layer of the key technology and role of exchange. We know that the second-level ieee802.1p field or the third-level Iptos field of the packet can be used to prioritize the packet itself, and we say that the fourth switch is based on the fourth Packet Exchange, which means that it can analyze the packet application type according to the fourth layer TCP/UDP port number. That is, layer fourth switches not only fully equipped with all the switching capabilities and performance of the third layer switch, but also support the intelligent features of network traffic and quality of service control that the third layer switch cannot have.
L2/L3/L4 Switch Introduction