Application of vrouters and vswitches in practice

At present, the networking technology is becoming more and more mature. Two of the important technologies, vrouters and vswitches, are also widely used, and the upgrade of the routing technology is also promoted. As a matter of fact, I have to admit that the difference between a vswitch and a vro is getting increasingly blurred, and the functions between them are also beginning to penetrate into each other.

Not only does a layer-3 Switch have some "routing" functions that originally belong to a vro, but also has the "Switch" function for both the broadband and high-end enterprise-level routers. It can be said that it is mutual penetration, so some people predict that in the future, the vswitch and the vro may be combined into one, and I firmly believe this. From the technical point of view, it is not very difficult to achieve this goal, and it is also an urgent demand for users. On the one hand, the network structure can be simplified. On the other hand, users do not have to buy two expensive devices. Why not? However, for the moment, there is still a large area between them. Of course, this is not only reflected in technical theory, but also mainly in application. This article will fully explain the main differences between vswitches and vrouters in applications.

We know that the most basic functions and applications of a vswitch are to centrally connect network devices. All network devices, such as servers, workstations, PCs, laptops, routers, firewalls, and network printers ), as long as the vswitch port supports the port type of the corresponding device, it can be directly connected to the vswitch port to form a star network. The basic network structure 1 is shown in. In a star connection, the port connecting devices of the vswitch are equal to each other and can access each other unless there are restrictions), instead of like many friends who are just involved in the network management, the server connected to the vswitch is the highest level.

The actual star enterprise network is much more complicated than this. This complexity is not only manifested in how high network devices are, but also how complicated the configurations are, but more importantly, the complexity of network switching layers. In Enterprise Networks, only one vro and one firewall are required, but a vswitch is usually not a small network except for about 20 users ). If the number of users is large, such as hundreds or even thousands, you must rely on the cascade or stack extension connections of the switch. However, cascade technology and stack technology are also different, and their application scope is also different.

A switch cascade is an extension between a switch and a switch through an exchange port. This solves the problem of insufficient number of ports of a single switch, on the other hand, it also solves the connection between clients and network devices that are far away from the data center. Because the single-segment Switched twisted pair ethernet cable can reach 100 meters, each vswitch can be connected to an extended distance of 100 meters. However, this does not mean that any cascade operation can be performed. Because the line is too long, on the one hand, the signal attenuation on the line is also high. On the other hand, the lower-level switch still shares the available bandwidth of a port of the higher-level switch, the more layers, the lower the available bandwidth of the client, even though you may use a MB switch). This has a great impact on the network connection performance, we recommend that you deploy a maximum of three levels of switches, that is, the core switch-Level 2 switch-Level 3 switch. The Level 3 here is not to say that you can only allow up to three vswitches, but only three levels at a level. Vswitches connected to different ports on the same vswitch belong to the same layer. Therefore, several or even dozens of vswitches can be connected to each layer. The ports used for hierarchical connections can be dedicated UpLink ports or common exchange ports. Some vswitches have dedicated cascade UpLink ports, but some do not. If a dedicated cascade port is available, it is best to use it because its bandwidth is usually wider than that of a common exchange port, which further ensures the bandwidth of a lower-level switch. If not, you can only cascade through common switching ports.

Note that not only the ports used are different, but the cables used are different: the cascade ports must be used for common direct connections; the cascade cable using a common port is a crossover cable, just like a pair of two hosts. As for the stack of A vswitch, not all vswitches are allowed, but the stack module is required. The stack of A vswitch is connected through a dedicated stack cable instead of a switching port. Note that because the switch stack is usually placed in the same position and the connection cable is short, the stack of the switch is mainly used to expand the switch port, rather than to expand the distance.

At the same time, the switch stack can also increase the available bandwidth of each actually used switch port, because it aggregates the bandwidth of the backboard of the switch that is stacked together, in this way, the total backplane of the switch stack is the sum of the bandwidth of several stack switches. After the increase in the bandwidth of the backboard, if every port of the switch is used, this advantage is not very obvious but also effective, because it is impossible to communicate with every port at the same time ), however, if a vswitch port is idle, the effect will be more obvious because it can take full advantage of all the bandwidth of the vswitch. The stack connection port of the switch is usually in the D-Jack format. A switch has two such ports marked with "UP" and "DOWN", as shown in ), it indicates that it is used to connect up and up the stack, and cannot be connected to an error.

As we mentioned earlier, layer-3 switches also have certain "routing" functions to achieve connections between different subnets. However, it should be noted that its routing function is much weaker than that of a router. The routing function of a layer-3 switch can only be used for interconnection between networks of the same type. Generally, it is only the interconnection between LAN subnets and does not connect the LAN to the WAN or the Internet, because layer-3 switches support very limited routing protocols, after all, this is not its main function.

We know that on the LAN, a layer-2 switch identifies the packet sender through the source MAC address and forwards the packet according to the destination MAC address. For a data packet whose destination address is not on the local area network, a layer-2 switch cannot directly send it to the destination. It must be forwarded through a routing device such as a traditional router, connect the vswitch to the routing device. If you set the default gateway of A vswitch as the IP address of the router device, the vswitch sends the packet to be forwarded by the route device.

The routing device checks the destination address of the data packet and its route table. If the route table finds the forwarding path, the routing device forwards the data packet to other network segments. Otherwise, the data packet is discarded. A dedicated router is expensive, complex, slow, and easy to become a network bottleneck because it needs to analyze all the broadcast packets and forward some of them, and exchange route information with other routers, and these processes are all handled by the CPU rather than the dedicated ASIC ). A layer-3 switch can identify and forward data packets through a MAC address like a layer-2 switch, and can also forward data packets between two CIDR blocks as a traditional router. Traditional routers use software to maintain route tables, while layer-3 switches use dedicated ASIC chips to process route forwarding. Compared with traditional routers, the routing speed of layer-3 switches is generally 10 times faster or dozens of times faster.