Introduction to layer-3 Switching

In order to adapt to the challenges brought by the deepening of network applications, in the past 20 years, the Network has developed rapidly in the speed and network segments. In terms of speed, it provides users with higher bandwidth: the speed of the LAN has increased from 10 Mbit/s to 100 Mbit/s. Currently, Gigabit Ethernet technology has been widely used. At the same time, FDDI and ATM technologies provide users with more options to increase the network speed. There has also been a qualitative breakthrough in the network segment: it has evolved from an earlier LAN with shared media to the current one. The exchange LAN technology enables dedicated bandwidth to be exclusive to users, greatly improving the efficiency of LAN transmission. It can be said that in the network system integration technology, the first-layer interface and the second-layer exchange technology that are directly oriented to users have obtained satisfactory answers. However, as the core of the network, the router technology that acts as the interconnection between networks has no qualitative breakthrough. Traditional routers are software-based and have complex protocols. Compared with lan speed, their data transmission efficiency is low. At the same time, it serves as the hub for interconnection between network segments (subnets and virtual networks), which makes traditional router technologies face severe challenges. With the rapid development of Internet/Intranet and the wide application of the B/S (Browser/Server) computing model, cross-region and cross-network business is growing rapidly, the industry and users are deeply aware of the bottleneck effect of traditional routers in the network. It is imperative to improve the traditional routing technology. In this case, a new routing technology came into being. This is the layer-3 switching technology: it is a router because it can operate on the layer-3 of the network protocol, it is a type of routing that understands a device and can act as a route decision. It is a switch because it is extremely fast and almost achieves the second-layer switching speed.

Evolution of Network Integration Technology

Network technology continues to evolve and evolve with the needs of applications. In the early 1980s s, when the first generation of LAN technology began to be applied to internal enterprise networking, its applications were mainly limited to host connection, file sharing, and print sharing, the 10 Mbit/s channel shared by multiple users can meet the requirements.

As the network grows, more and more online users are available. In particular, users' applications have switched to customers/servers, high-traffic applications, Intranet Web access, and real-time audio/video services. At that time, the network system was no longer competent, as shown in the following figure: HUB is a communication device based on shared media, and it is a first-level device. User Data Collision Detection and error re-transmission greatly reduce transmission efficiency. The bridge can be used to minimize network segments and reduce collision domains to optimize LAN performance. It is a second-level device that can identify MAC addresses and intelligently forward information between local networks. However, it is a device that is transparent to high-level (Layer 3 or above) protocols and cannot effectively prevent broadcast storms. Vrouters play a key role in subnet interconnection, security control, and broadcast storm restrictions. However, their complex algorithms and low data throughput make them the network bottleneck. Aware of the above problems, the industry started from the HUB and bridge infrastructure that directly targets users and can form independent LAN devices, and revolutionizes network technology, the biggest change is to replace the HUB with LAN switches in the next generation of network system integration to improve network performance.

A large number of LAN switches were introduced in the network system integration mode in early 1990s. LAN switches are Layer 2 network devices. They can understand Layer 2 network protocols, such as MAC addresses. During the operation, the switch constantly collects data to create its own address table, which is quite simple and mainly indicates the port on which a MAC address is found, when the switch receives a data packet, it checks the target MAC address of the packet and its address table to determine the port from which the packet is sent. Instead of the HUB, any sender data will appear on all the ports of the HUB (whether or not required for you ).

The introduction of LAN switches enables dedicated bandwidth between network sites, eliminating unnecessary Collision Detection and error re-transmission, and improving transmission efficiency, in a vswitch, several independent and independent communication processes can be maintained in parallel. In the exchange network environment, user information is transmitted only between the source node and the target node, and other nodes are invisible. However, when a node sends a broadcast or multi-object broadcast on the Internet, or a node sends a MAC address packet that the switch does not know, all nodes on the vswitch will receive this broadcast message. The entire exchange environment constitutes a large broadcast domain. Industry insiders use a new term Flat Network to describe this environment: multiple switches Interconnect (stacked) to form a large LAN, but cannot effectively divide subnets. "Peer To Peer" is a fast and effective exchange on the second layer. However, broadcast storms can compromise network efficiency. Vswitches are much faster and cheaper than vrouters. However, layer-2 switching also exposes vulnerabilities: it cannot effectively solve broadcast storms, inter-network interconnection, and security control. Therefore, the virtual network technology on the vswitch is generated.

In fact, a virtual network is a broadcast domain. In order to avoid the broadcast storm caused by broadcasting on a large switch, we can further divide it into multiple virtual networks. In a virtual network, information sent by a workstation can only be sent to other sites with the same virtual network number. Other Virtual Network members cannot receive the information or broadcast frames. Using virtual networks has the following advantages:
Control broadcast storms on the network;
Increases network security;
Centralized management and control;
That is, the original intention of adopting the virtual network technology on LAN switches has indeed solved some problems. However, this technology also raises some new problems: with the upgrade of applications, network planning/implementers can divide users into different virtual networks in a switched LAN environment as needed. However, communication between virtual networks is not allowed, which also includes IP Address Resolution (ARP) packets. To communicate, you need to use a router to bridge these virtual networks. This is the problem of the virtual network: It is too slow to use a router. The speed of the switch is fast, but it cannot solve the problem of broadcast storm. Using the virtual network technology in the switch can solve the problem of broadcast storm, however, you must place a vro to connect virtual networks. Formed an insurmountable Strange Circle. This is the core of the network and the hub router. In this network system integration mode, routers are the core.
The role of a router is:
The network segment is micro-structured;
Network Congestion Control;
Network security control;
Interconnection Between subnets (virtual networks;
Problems arising from the use of routers as the core of the network vro increase the time required for layer-3 route selection, low data transmission efficiency increases, the complexity of moving and changing nodes increases, the price of routers is expensive, and the complexity of the structure increases the connection between subnets and virtual networks, which means the increase of router ports and the increase of investment.

In terms of applications, the Internet and Intranet have developed rapidly, and the cross-network and cross-region B/S computing modes have been widely used, which puts forward higher requirements for routers. The high cost and low performance of a vro make it a network bottleneck. However, due to the need for interconnection between networks, it is indispensable and at the core of the network. It can be said that when the network technology develops to this step, the core of the network-the innovation of the router technology is no longer urgent. The real-world applications pose severe challenges to vro technology. In this case, the third-layer exchange technology is proposed.

Principle of layer-3 Switching Technology

A device with layer-3 switching is a layer-2 switch with layer-3 routing, but it is an organic combination of the two, it is not simply to overlay the hardware and software of the router device on the LAN switch.
From the hardware implementation point of view, at present, the interface modules of the second layer switch exchange data through the High-Speed Backplane/bus (the speed can reach dozens of Gbit/s, in layer-3 vswitches, layer-3 router-related Routing hardware modules are also inserted on the high-speed backplane/bus, in this way, the routing module can exchange data with other modules to be routed at a high speed, this breaks through the speed limit of traditional external router interfaces (10 Mbit/s --- 100 Mbit/s). In terms of software, layer-3 switches also have major initiatives, it defines the traditional software-based router software. The practice is:
1. For data packet forwarding, such as IP/IPX packet forwarding, these regular processes are implemented through hardware at high speed.
2. For layer-3 software, such as route information update, route table maintenance, route calculation, and route determination, Use optimized and efficient software. Assume that two sites using the IP protocol communicate with each other through the layer-3 Switch. When sending Site A, the IP address of the destination site is known, however, you do not know the MAC address required to send messages on the LAN. IP Address Resolution (ARP) is used to determine the MAC address of the target site. The sending site compares its IP address with the IP address of the destination site and extracts the network address using the subnet mask configured in the software to determine whether the destination site is in the same subnet as itself. If the destination site B and the sender Station A are in the same subnet, A broadcasts an ARP request, and B returns its MAC address, after obtaining the MAC address of the target site B, A caches the address and uses the MAC address to package and forward data. The second layer switch module looks for the MAC address table to determine whether to send data packets to the target port. If two sites are not in the same subnet, if Station A needs to communicate with Station C, station A must send an ARP (Address Resolution) packet to the "Default Gateway, the IP address of the "Default Gateway" has been set in the system software. This IP address actually corresponds to the layer-3 Switch module of the layer-3 switch. Therefore, when station A broadcasts an ARP request to the IP address of the "Default Gateway", if the layer-3 Switch module obtains the MAC address of the target Station B during the previous communication, then, send A reply to the MAC address of B to the sender. Otherwise, the layer-3 Switch module broadcasts an ARP request to the destination based on the route information, after receiving the ARP request, the target Station C replies to the MAC address of the layer-3 Switch module. The layer-3 Switch module saves the address and replies to the sending station. Later, when packet forwarding between A and C is performed, the MAC address packet of the destination site will be used, and the data forwarding process will be handed over to the second layer for exchange, so that information can be exchanged at A high speed. Layer-3 switching has the following outstanding features: the combination of organic hardware accelerates data exchange; the optimized routing software improves the routing efficiency; in addition to the necessary routing decision processes, most data forwarding processes are processed by Layer 2 switching. When multiple subnets are connected, they are only logically connected to the Layer 3 Switching Module. Unlike traditional external routers, ports need to be added, this protects users' investment.

Currently, layer-3 switches have been put into use in network integration, and their excellent performance has been widely recognized by users. However, as a new technology, there is still a long way to mature the layer-3 Switch. Like other new technologies, it is still necessary to standardize its protocols. At present, many manufacturers claim to have developed third-tier switches. However, they have been tested by international authorities with different practices and different performance. In addition, it may be based on the strategy of various manufacturers occupying the market. Currently, layer-3 switches can exchange routing IP addresses/IPX protocols, and cannot handle other dedicated protocols with certain application fields. Therefore, relevant experts believe that the layer-3 exchange technology will be the network integration technology of the next century. Traditional routers will be applied within a period of time, but they will be in the position they can do, that is, Wan interconnection, security control (firewall), and dedicated protocol heterogeneous machine interconnection at the edge of the network.

Related Articles]

• Changes in the application environment of layer-3 switches and Market Prospects
  

• Layer 3 Switch construction enterprise VLAN
  

• Analysis of Application Status of layer-3 switching technology in Enterprises