Computer Network Technology and Application

Chapter 29 Appendix LAN Networking Equipment
1. Network card
1.1 The concept of network card

A Network Interface Card (NIC), also known as a network card, is a basic component of a network.A network card connects computers and transmission media in a local area network.

1.2 Network card classification
1. Classified according to the transmission rate supported by the network card
这是目前主要的分类方法。共有五类：①10Mbps网卡， ②100Mbps网卡，③10／100Mbps自适应网卡，④1000Mbps网卡，⑤10／100／1000Mbps自适应网卡。其中，10／100Mbp自适应网卡可同时支持10Mbps和100Mbps的传输速率，10／100／1000Mbps自适应网卡可同时支持10Mbps、100Mbps和1000Mbps的传输速率，它们均能自动侦测出网络的传输速率。

2. Classified according to the type of transmission medium supported by the network card

According to this classification method, network cards can be divided into four types: ① twisted pair network card, the commonly used unshielded twisted pair network card uses RJ-45 interface; ② thick cable network card, using AUI interface; ③ thin cable network card, using BNC Interface; ④ fiber optic network card, use F/0 interface.

Currently, due to the widespread use of unshielded twisted-pair cables, Ethernet cards that only provide RJ-45 interfaces are more popular.

2. Hub

2.1 Transponder

Repeater, also known as repeater or amplifier, works on the physical layer of OSI to realize the distortion-free forwarding of electrical signals.When the signal propagates along the transmission medium, it will produce attenuation and distortion. This attenuation and distortion will make the receiver unable to correctly identify the signal.The transponder can make the waveform and strength of the signal reach the required index by receiving, amplifying, shaping and forwarding.

The transponder is used to interconnect two network segments of the same type (for example: two Ethernet segments). Its main function is to extend the length of the network cable, expand the distance of the network segment and connect different transmission media, so as to realize the forwarding of bits.Since the network segments connected by the repeater still belong to the same subnet, in a sense, the repeater cannot be regarded as a real network interconnection device.

2.2 Hub (Hub)

Hub (Hub) is a kind of transponder and also works at the physical layer.It is a multi-port (8, 16, 24, 48 RJ-45 interfaces) repeater, which is equivalent to condensing the bus and repeaters of the bus network into a hub. The Hub can separate faulty sites, ensure the normal operation of other sites, and effectively improve the reliability of the Ethernet.

1. How the hub works
A hub does not process or inspect the traffic on it, but merely extends the physical medium by redistributing signals received through one port to other ports.All devices connected to the hub share the same medium, and as a result they also share the same collision domain, broadcast, and bandwidth.Thus the hub and its connected devices form a single collision domain.If a node sends out a broadcast message, the hub will spread the broadcast to all nodes connected to it, so it is also a single broadcast domain.

2. Working characteristics of the hub
(1) The hub is mostly used for small-scale Ethernet, and generally uses an external power supply (active), which can amplify the signal it receives.In some occasions, a hub is also called a "multiport repeater".A repeater is also a network device that works at the physical layer.

(2) All computers connected to the hub port are in the same network segment, and the information sent by the computer on each port is amplified, shaped and sent to all ports in the form of broadcast by the hub.

(3) The computers connected to the Hub still use the CSMA/CD method to compete for the use of bandwidth and share the bandwidth of the entire hub.At any one time, only one computer can send data, and only two computers can communicate (one sends, one receives).

(4) A stacked hub is the same as a single hub, and all ports on it belong to the same network segment.The use of hub cascading can expand the scope of network coverage.

(5) A maximum of 4 hubs can be connected in series with each other. If cascaded with twisted pairs, the length of each twisted pair should not exceed 100m. Therefore, the maximum length between two computers is 500m.All computers on the hub still belong to the same subnet and share the network bandwidth.

3. Bridges and switches
Both bridges and switches are network interconnection devices that work at the data link layer of the OSI model.The difference is that the bridge can realize the connection between the same type of subnets, and can also realize the connection between different types of subnets; and the switch is a network interconnection device designed to realize high-speed connections between the same subnets.

3.1 The function and application of the network bridge
A network bridge is also called a network bridge. As a network interconnection device, the bridge provides a LAN-LAN interconnection method and expands the LAN.

1. The main difference between a bridge and a hub

The bridge can be used to interconnect two independent subnets to realize the storage and forwarding of data frames; the hub is only used for the extension of the same subnet.

2. The protocol layer of the bridge
The bridge implements the protocol conversion between the data link layer and the physical layer of the OSI model, and is suitable for the interconnection between similar networks or networks that differ only in the lower two layers of protocols.

After the bridge receives a frame, it first checks it at the data link layer, then hands the frame to the physical layer, and forwards it to a different network.Before forwarding these frames, the bridge may make some changes to the header information of the frames in order to perform protocol conversion on the data link layer, but the user data carried by the frames will not be changed.The bridge implements the connection of the MAC sublayer, which is completely transparent to the LAN following the IEEE802 standard.

A bridge can connect multiple subnets, and their types can be the same or different. The bridge receives all frames in each subnet it connects and forwards them to the destination subnet.

3.2 Features of bridges

In the process of forwarding frames, besides the protocol conversion on the data link layer, the bridge also has the following functions and features.

(1) Address filtering
The bridge can identify various addresses and selectively allow data frames to pass through the bridge according to the destination address of the data frame.In fact, many bridge products currently have added various filtering functions, which allow users to set up and filter out frames that do not want to be forwarded, so as to ensure the security of the subnet.

(2) Frame limit
The bridge only performs the necessary frame format conversion to adapt to different subnets.The bridge discards information that exceeds the frame length limit of the subnet where the sink node is located.Therefore, when a bridge is used to support the interconnection between different LANs, a higher layer protocol is required to ensure the limitation of the length of the transmitted information.Another purpose of frame restriction is to maintain the independence of each subnet, not to allow control frames and information frames requiring responses to pass through the bridge, and to avoid broadcast storms.

(3) Monitoring function

The bridge participates in the monitoring of the subnet and the verification of the information frame.

(4) Buffer capacity

The bridge should have a certain buffer capacity to solve the problem of temporary overload of the amount of information passing through the bridge.That is to say, the bridge can solve the interconnection between subnets whose data transmission rate does not match.In fact, this buffering capability is necessary even if networks of the same rate are interconnected.

(5) Transparency

The introduction of the network bridge should not affect the communication capability of the original subnet, nor should it cause errors that cannot be detected by the sink node.

3.3 Path selection of the bridge
A bridge can directly connect two or more subnets.The process of frame forwarding is essentially the process of frame path selection. After path selection, the bridge sends the frame to the appropriate port.it is necessary.

There are two commonly used path selection methods, corresponding to two types of bridges - transparent bridges and designated path bridges.

1. Transparent bridge
A transparent bridge is also called a learning bridge or an adaptive bridge. The bridge dynamically maintains the node's address mapping table database; according to the address mapping table, the bridge decides to forward the received frame.

Because the operation of the LAN is completely independent of the bridge, it is called a "transparent" bridge.Transparent bridges are suitable for bus-type (such as: Ethernet, token bus network) or tree-structure network interconnection.

2. Source path selection bridge

The principle of the bridge comes from IBM's token ring.The source path selection bridge is also called the specified path bridge. The source node first judges whether the sent frame is sent to the local subnet or to other subnets, and then selects the exact path for frame transmission.

(1) Direct transmission

If the sending source node knows the exact path of the transmitted frame transmission, it can be transmitted directly.

(2) Find the path

If the source node does not know the path, it can send a broadcast frame with test function.The bridge that receives the broadcast frame first checks the RI (routing information) field in the broadcast frame. If the bridge is already in the RI——the frame has been forwarded by the bridge without any processing; otherwise, add bridge and port information, and forward the frame to other subnets connected to it that do not appear in the frame.After receiving the test frame, the sink node returns a response frame to the source sending node.The acknowledgment frame contains the required path information and is passed back along the path taken by the test frame.Due to the broadcast, the source node may receive multiple reply frames, usually through some algorithm to choose a (best) path.

The source path selection bridge can obtain the best path, but the sending of test frames increases the information flow of the network, which may form a broadcast storm and even cause network congestion.

3.4 Switches

If the same network is connected, the function of the bridge can be greatly simplified, thereby improving efficiency.An Ethernet switch is a simplified bridge used for interconnection between Ethernets.The characteristics of an Ethernet switch are as follows.

(1) The same frame format is processed, and the exchange speed is fast.Since the switches are interconnected with the same type of network, there is no need for frame format conversion when data frames are forwarded, thus greatly improving the data exchange speed of the switches.

(2) With a small number of address tables, the table look-up speed of the switch is improved.

(3) Because multiple pairs of ports can communicate at the same time, it supports multiple independent data streams and has high throughput.

(4) Hardware switching is used internally, and the switching speed is fast.

(5) Like a general bridge, it has the function of dividing subnets (at the data link layer).

(6) Provide a certain storage capacity.In order to avoid the collision of forwarded frames at the output port, the switch is often configured with a certain buffer for buffering input or output frames, so that the interconnection of networks with different rates can be realized (adaptive capability).

（7）每个端口独享指定的带宽。如10M／100M／1000M的交换机，每个端口可独享10Mbps／100Mbps／1000Mbps的网络带宽。

3.5 Layer [-] Switching
Layer [-] switching technology is Layer [-] switching technology plus Layer [-] forwarding technology.After the division of the network segment in the LAN, the management of the subnet in the network segment has always relied on the router. The appearance of the three-layer switch has solved the network bottleneck problem caused by the low speed and complexity of the traditional router.

1. Types of Layer [-] switches

Layer [-] switches can be divided into two types: pure hardware and pure software according to their data processing methods.

(1) Three-layer technology of pure hardware

This method is complex in technology and high in cost, but it has fast speed, good performance and strong load capacity.

The principle of pure hardware is to use ASIC chips to search and refresh the routing table in hardware.

After the data is received by the interface chip of the port, the corresponding destination MAC address is first searched in the layer-[-] switching chip, and if found, the layer-[-] forwarding is performed, otherwise the data is sent to the layer-[-] protocol.

In the three-layer engine, the ASIC chip searches the corresponding routing table information, compares it with the destination IP address of the data, sends the ARP data packet to the destination host, obtains the MAC address of the host, and sends it to the second-layer chip, and the second-layer The chip forwards the packet.

(2) Software-based Layer [-] switch technology

This technique is simpler, but slower and not suitable for the backbone.

The principle is to use the CPU to use software to search the routing table.After the data is received by the port interface chip, the corresponding destination MAC address is first searched in the Layer [-] switching chip, and if found, the Layer [-] forwarding is performed, otherwise the data is sent to the CPU. The CPU continues to search for the corresponding routing table information, compares it with the destination IP address of the data, sends the ARP data packet to the destination host to obtain the MAC address of the host, and then sends the MAC address to the second-layer chip, and the second-layer chip forwards the data packet .

Due to the slow processing speed of low-cost CPUs, the processing speed of this three-layer switch is relatively slow.

2. Layer [-] switching principle
Assuming that two sites A and B using the IP protocol communicate through a layer-[-] switch, the process of layer-[-] switching is as follows.

(1) When sending station A starts sending, it compares its own IP address with that of station B to determine whether station B is in the same subnet as itself.

(2) If the destination station B and the sending station A are in the same subnet, carry out Layer [-] forwarding.

(3) If the two stations are not in the same subnet, the sending station A needs to communicate with the destination station B.The sending station A wants to send an ARP (address resolution) packet to the "default gateway", and the IP address of the "default gateway" points to the layer-[-] switching module of the layer-[-] switch.

When the sending station A broadcasts an ARP request to the IP address of the "default gateway", if the three-layer switch module has known the MAC address of the station B in the previous communication process, it will reply the MAC address of B to the sending station A.Otherwise, the layer-[-] switching module broadcasts an ARP request to station B according to the routing information. After receiving the ARP request, station B replies its MAC address to the layer-[-] switching module. The layer-[-] switching module saves the address and replies to sending station A, and at the same time sends the MAC address of station B to the MAC address table of the layer-[-] switching engine.

(4) At this point, all the data packets sent by A to B are handed over to the Layer [-] switch for processing, and the information can be exchanged at a high speed.

Since layer [-] processing is only required in the routing process, most of the data is forwarded through layer [-] switching, so the speed of layer [-] switches is very fast, close to the speed of layer [-] switches, and at the same time, the price is much lower than that of the same router.

(End of this chapter)