<<
Routers are the primary node devices of the Internet. They determine data forwarding through routing decisions. The forwarding strategy is called routing, which is also the origin of the router’s name. As the hub connecting different networks, router systems form the main脉络 of the TCP/IP-based Internet. It can be said that routers constitute the backbone of the Internet. Their processing speed is one of the main bottlenecks in network communication, and their reliability directly affects the quality of network interconnection. Therefore, in campus networks, regional networks, and even the entire Internet research field, router technology has always been at the core. Its development history and direction serve as a microcosm of the entire Internet research. At a time when China’s network infrastructure and information construction are in full swing, discussing the role, status, and development direction of routers in interconnected networks is of great significance for domestic network technology research, network construction, and clarifying the various specious concepts about routers and network interconnection in the network market.
The Role of Routers
One role of a router is to connect different networks, and another is to select the path for information transmission. Choosing a smooth and fast shortcut can greatly increase communication speed, reduce network system communication load, save network system resources, and improve network system throughput, thereby allowing the network system to achieve greater efficiency.
From the perspective of filtering network traffic, the role of a router is very similar to that of switches and bridges. However, unlike switches that work at the physical layer and physically divide network segments, routers use specialized software protocols to logically divide the entire network. For example, a router supporting the IP protocol can divide a network into multiple subnets, and only network traffic directed to specific IP addresses can pass through the router. For every received data packet, the router recalculates its checksum and writes a new physical address. Therefore, the speed of forwarding and filtering data using routers is often slower than that of switches, which only check the physical address of data packets. However, for networks with complex structures, using routers can improve the overall efficiency of the network. Another significant advantage of routers is that they can automatically filter network broadcasts. Generally speaking, the entire installation process of adding a router to a network is much more complex than that of plug-and-play switches.
Generally, interconnection of heterogeneous networks and interconnection of multiple subnets should be completed using routers.
The main job of a router is to find an optimal transmission path for each data frame passing through it and effectively transmit the data to the destination site. Thus, the strategy for selecting the best path, i.e., the routing algorithm, is the key to a router. To accomplish this task, the router stores data related to various transmission paths—the Routing Table—for use during route selection. The path table stores subnet identification information, the number of routers on the network, the name of the next router, and other content. The path table can be fixedly set by the system administrator, dynamically modified by the system, automatically adjusted by the router, or controlled by the host.
1. Static Path Table
A path table that is set up in advance by the system administrator is called a static path table. It is generally pre-set based on the network configuration during system installation and does not change with future changes in network structure.
2. Dynamic Path Table
A dynamic path table is one that the router automatically adjusts based on the operating conditions of the network system. The router automatically learns and memorizes network operating conditions based on the functions provided by the Routing Protocol, and automatically calculates the best path for data transmission when needed.
Router Architecture
Router Architectural Structure
From an architectural perspective, routers can be divided into first-generation single-bus single-CPU structure routers, second-generation single-bus master-slave CPU structure routers, third-generation single-bus symmetrical multi-CPU structure routers; fourth-generation multi-bus multi-CPU structure routers, fifth-generation shared-memory structure routers, sixth-generation crossbar switch architecture routers, and routers based on cluster systems, among others.
Router Composition
A router has four elements: input ports, output ports, a switching fabric, and a routing processor.
Input ports are the entry points for physical links and incoming packets. Ports are usually provided by line cards, with one line card typically supporting 4, 8, or 16 ports. An input port has many functions. The first function is encapsulation and decapsulation at the data link layer. The second function is to look up the destination address of the incoming packet in the forwarding table to determine the destination port (called route lookup). Route lookup can be implemented using general hardware or by embedding a microprocessor on each line card. Third, to provide QoS (Quality of Service), the port categorizes received packets into several predefined service levels. Fourth, the port may need to run data link layer protocols such as SLIP (Serial Line Internet Protocol) and PPP (Point-to-Point Protocol) or network layer protocols such as PPTP (Point-to-Point Tunneling Protocol). Once the route lookup is complete, the packet must be sent to its output port using the switching fabric. If the router uses input queuing, several input ports share the same switching fabric. Thus, the final function of the input port is to participate in arbitration protocols for common resources (like the switching fabric).
The switching fabric can be implemented using various technologies. The most used switching fabric technologies to date are bus, crossbar, and shared memory. The simplest switch uses a single bus to connect all input and output ports. The disadvantage of a bus switch is that its switching capacity is limited by the bus capacity and the additional overhead caused by shared bus arbitration. A crossbar switch provides multiple data paths through the switch. A crossbar switch with N×N crosspoints can be considered to have 2N buses. If a crosspoint is closed, the data on the input bus is available on the output bus; otherwise, it is not. The closing and opening of crosspoints are controlled by a scheduler; thus, the scheduler limits the speed of the switching fabric. In a shared memory router, incoming packets are stored in shared memory, and only the packet pointers are switched, which increases switching capacity. However, the switch speed is limited by the memory access speed. Although memory capacity doubles every 18 months, memory access time decreases by only 5% annually, which is an inherent limitation of the shared memory switching fabric.
Output ports store packets before they are sent to the output link and can implement complex scheduling algorithms to support requirements such as priority. Like input ports, output ports must also support encapsulation and decapsulation at the data link layer, as well as many higher-level protocols.
The routing processor calculates the forwarding table to implement routing protocols and runs the software for configuring and managing the router. At the same time, it also processes packets whose destination addresses are not in the line card forwarding table.
Types of Routers Routers can be seen everywhere in various levels of networks on the Internet. Access networks allow homes and small businesses to connect to an Internet Service Provider; routers in enterprise networks connect thousands of computers within a campus or enterprise; router terminal systems on the backbone network are usually not directly accessible; they connect ISPs and enterprise networks on long-distance backbone networks. The rapid development of the Internet has brought different challenges to backbone networks, enterprise networks, and access networks. Backbone networks require routers to perform high-speed routing and forwarding on a few links. Enterprise-class routers require not only a large number of ports and low cost but also simple and convenient configuration, along with QoS provisioning.
1. Access Routers
Access routers connect home or small business customers within an ISP. Access routers have begun to not only provide SLIP or PPP connections but also support virtual private network protocols such as PPTP and IPSec. These protocols need to be able to run on each port. Technologies like ADSL will soon increase the available bandwidth to homes, further increasing the burden on access routers. Due to these trends, access routers will support many heterogeneous and high-speed ports in the future, run multiple protocols on each port, and also need to bypass the telephone switching network.
2. Enterprise Routers
Enterprise or campus routers connect many terminal systems, with the main goal being to achieve interconnection of as many endpoints as possible at the lowest cost, further requiring support for different levels of service quality. Many existing enterprise networks are Ethernet segments connected by Hubs or bridges. Although these devices are cheap, easy to install, and require no configuration, they do not support service levels. In contrast, networks involving routers can divide machines into multiple collision domains, thereby controlling the size of a network. Furthermore, routers support certain service levels, at least allowing division into multiple priority levels. However, the cost per port for routers is more expensive, and extensive configuration work is needed before they can be used. Therefore, the success of enterprise routers lies in whether they provide a large number of ports at a low cost per port, are easy to configure, and support QoS. Additionally, enterprise routers are required to effectively support broadcast and multicast. Enterprise networks also need to handle various legacy LAN technologies and support multiple protocols, including IP, IPX, and Vine. They also need to support firewalls, packet filtering, numerous management and security policies, and VLANs.
3. Backbone Routers
Backbone routers enable the interconnection of enterprise-level networks. The requirements for them are speed and reliability, with cost being of secondary importance. Hardware reliability can be achieved using technologies used in telephone switching networks, such as hot backup, dual power supplies, and dual data paths. These technologies are almost standard for all backbone routers. The main performance bottleneck for backbone IP routers is the time spent looking up a route in the forwarding table. When a packet is received, the input port looks up the destination address of the packet in the forwarding table to determine its destination port. The shorter the packet or when a packet is destined for many destination ports, the cost of route lookup increases. Therefore, placing some frequently accessed destination ports in the cache can improve route lookup efficiency. Whether it is an input buffered or output buffered router, there exists a route lookup bottleneck problem. Besides performance bottlenecks, router stability is also a commonly overlooked issue.
4. Terabit Routers
Among the three main technologies for the future core Internet, fiber optics and DWDM are already mature and available. If there are no routers corresponding to the raw bandwidth provided by existing fiber optic and DWDM technologies, the new network infrastructure will not fundamentally achieve performance improvements. Therefore, developing high-performance backbone switching/routers (terabit routers) has become an urgent requirement. Terabit router technology is still mainly in the development and experimental stage.
Basic Router Protocols and Technologies
VPN
VPN (Virtual Private Network) solutions are one of the important functions routers possess. The solutions are roughly as follows:
1. Access Control
Generally divided into two protocols: PAP (Password Authentication Protocol) and CHAP (Challenge Handshake Authentication Protocol). PAP requires the login user to provide a username and password to the target router, and only if it matches the information in its Access List is login allowed. Although it provides some security, the user login information is transmitted unencrypted over the network and can be easily stolen. CHAP emerged to solve this by translating an initial random value and the user’s original login information (username and password) using a Hash algorithm to form new login information. This way, the user login information transmitted over the network is opaque to hackers, and since the initial random value is different each time, the user’s final login information will also be different each time. Even if the user login information is stolen once, the hacker cannot reuse it. It should be noted that because different manufacturers use their own different Hash algorithms, CHAP has no interoperability. To establish a VPN, the same brand of routers needs to be placed at both ends of the VPN.
2. Data Encryption
In the encryption process, the number of encryption bits is a very important parameter, as it is directly related to the difficulty of decryption. The Intel 9000 series routers perform most excellently in this regard, with over one hundred bits of encryption.
3. NAT (Network Address Translation)
Just like user login information, it is very insecure for IP and MAC addresses to be transmitted unencrypted over the network. NAT can translate legal IP and MAC addresses into illegal IP and MAC addresses for transmission over the network, and upon reaching the target router, translate them back into legal IP and MAC addresses. This process is somewhat like CHAP, and the translation algorithms have different standards among manufacturers, making interoperability impossible.
QoS
<<
digit after the closing quotation mark. Configuring BGP requires a deep understanding of user requirements, network status, and the BGP protocol itself, along with extreme caution. BGP operates at a relatively core position in the network; once an error occurs, the resulting damage can be significant! IPv6 Technology
The rapidly developing Internet will no longer be just a network connecting computers; it will evolve into an information communication infrastructure similar to the telephone network and cable television network. Therefore, the currently used IP (Internet Protocol) is becoming inadequate, and people are eagerly anticipating the next-generation IP, namely IPv6.
IPv6 is a version of IP, a transmission protocol at Layer 3 (Network Layer) of the OSI model within the TCP/IP Internet communication protocol suite. Compared to the widely used IPv4, proposed back in 1974, its address space expands from 32 bits to 128 bits. Theoretically, the number of addresses increases from 4.3×109 to 4.3×1038. There are two main reasons for the necessary transition from the current IPv4 to IPv6.
1. Due to the rapid development of the Internet, the number of addresses is no longer sufficient, making the effort and cost of network management unbearable. Address exhaustion is the primary reason driving the transition to the 128-bit address space.
2. As the number of hosts increases, the routing tables determining data transmission paths are constantly growing. Router processing performance cannot keep up with this rapid growth. If this continues, Internet connectivity will struggle to provide stable services. With IPv6, the number of routes can be reduced by an order of magnitude.
To make connecting many things to the Internet simpler and easier to use, IPv6 must be adopted. IPv6 achieves this by using four technologies: expansion of the address space, address structuring that reduces routing table size, automatic address configuration, and improved security and confidentiality.
IPv6 inherits the advantageous aspects of IPv4 in routing technology, representing the future direction of routing technology development. Many router manufacturers have already invested significant effort into producing routers that support IPv6. Of course, IPv6 also has some notable and inefficient areas, and IPv4/NAT and IPv6 will coexist for a considerable period.
Router Configuration and Debugging
Routers play a pivotal role in computer networks, serving as the bridges of computer networks. Through them, one can not only connect different networks but also select data transmission paths and block unauthorized access.
Router configuration is not an easy task for beginners. Here, we introduce general router configuration and simple debugging for your reference when configuring routers, using the Cisco 2501 as an example.
The Cisco 2501 features one Ethernet port (AUI), one Console port (RJ45), one AUX port (RJ45), and two synchronous serial ports, supporting DTE and DCE devices, and supporting EIA/TIA-232, EIA/TIA-449, V.35, X.25, and EIA-530 interfaces.
I. Configuration
1. Configuring the Ethernet Port
# conf t (Configure the router from the terminal)
# int e0 (Specify the E0 interface)
# ip addr ABCD XXXX (ABCD is the Ethernet address, XXXX is the subnet mask)
# ip addr ABCD XXXX secondary (The E0 interface supports two address types simultaneously. If the first is a Class A address, the second is a Class B or C address)
# no shutdown (Activate the E0 interface)
# exit
After completing the above configuration, use the ping command to check if the E0 interface is normal. If it is not, it is generally because the interface has not been activated, a detail beginners often overlook. Use the no shutdown command to activate the E0 interface.
2. X.25 Configuration
# conf t
# int S0 (Specify the S0 interface)
# ip addr ABCD XXXX (ABCD is the IP address for S0, XXXX is the subnet mask)
# encap X25-ABC (Encapsulate X.25 protocol. ABC specifies DTE or DCE operation for X.25, default is DTE)
# x25 addr ABCD (ABCD is the X.25 port address of S0, provided by the telecommunications bureau)
# x25 map ip ABCD XXXX br (Mapped X.25 address. ABCD is the IP address of the peer router (e.g., S0), XXXX is the X.25 port address of the peer router (e.g., S0))
# x25 htc X (Configure the highest two-way channel number. X ranges from 1-4095, must be configured based on the actual number provided by the telecommunications bureau)
# x25 nvc X (Configure the number of virtual circuits, X must not exceed the actual number provided by the telecommunications bureau, otherwise normal data transmission will be affected)
# exit
After the S0 port configuration is complete, use the no shutdown command to activate the interface. If pinging the S0 port is successful, but pinging the mapped X.25 IP address (i.e., the peer router’s port IP address) fails, it may be due to the following situations: 1) The local X.25 address is incorrectly configured; reconfirm with the post office (X.25 address length is 13 digits); 2) The local mapped IP address or X.25 address is incorrectly configured; reconfigure correctly; 3) The peer IP address or X.25 address is incorrectly configured; 4) The local or peer routing configuration is incorrect.
If communication with the peer is possible but packet loss occurs, the following reasons are common: 1) Poor line conditions, or poor contact with the network card or RJ45 connector; 2) The value X for the x25 htc highest two-way channel number and the x25 nvc virtual circuit number X exceed the actual numbers provided by the telecommunications bureau. Higher values for both are better, but they must absolutely not exceed the numbers provided by the telecommunications bureau, otherwise packet loss will occur.
3. Leased Line Configuration
# conf t
# int S2 (Specify the S2 interface)
# ip addr ABCD XXXX (ABCD is the IP address of S2, XXXX is the subnet mask)
# exit
After the leased line port configuration is complete, use the no shutdown command to activate the S2 interface.
4. Frame Relay Configuration
# conf t
# int s0
# ip addr ABCD XXXX (ABCD is the IP address of S0, XXXX is the subnet mask)
# encap frante_relay (Encapsulate the frante_relay protocol)
# no nrzi_encoding (NRZI=NO)
# frame_relay lmi_type q933a (LMI uses the Q933A standard. LMI (Local Management Interface) has 3 types: ANSI: T1.617, CCITT: Q933A, and Cisco’s proprietary standard)
# fram-relay intf-typ ABC (ABC is the frame relay device type, supporting DTE devices, DCE switches, or NNI (Network-to-Network Interface))
# frame_relay interface_dlci 110 br (Configure the DLCI (Data Link Connection Identifier))
# frame-relay map ip ABCD XXXX broadcast (Establish frame relay mapping. ABCD is the peer IP address, XXXX is the local DLCI number, broadcast allows broadcasts to be forwarded or routing updates)
# no shutdown (Activate this port)
# exit
After the frame relay S0 port configuration is complete, use the ping command to check the S0 port. If it is abnormal, it is usually because the port has not been activated; use the no shutdown command to activate the S0 port. If pinging the S0 port is successful but pinging the mapped IP address fails, it could be a configuration error on the frame relay switch or the peer side, requiring comprehensive troubleshooting.
5. Configuring Synchronous/Asynchronous Ports (Applicable to 2522)
# conf t
# int s2
# ph asyn (Configure S2 as an asynchronous port)
# ph sync (Configure S2 as a synchronous port)
6. Dynamic Routing Configuration
# conf t
# router eigrp 20 (Use the EIGRP routing protocol. Common routing protocols include RIP, IGRP, IS-IS, etc.)
# passive-interface serial0 (If S0 is connected to X.25, enter this command)
# passive-interface serial1 (If S1 is connected to X.25, enter this command)
# network ABCD (ABCD is the local machine’s Ethernet address)
# network XXXX (XXXX is the IP address of S0)
# no auto-summary
# exit
7. Static Route Configuration
# ip router ABCD XXXX YYYY 90 (ABCD is the peer router’s Ethernet address, XXXX is the subnet mask, YYYY is the corresponding WAN port address of the peer)
# dialer-list 1 protocol ip permail
II. Comprehensive Debugging
After all router configurations are complete, a comprehensive debugging session can be performed.
1. First, activate the router’s Ethernet port and all serial ports to be used. The method is to enter each interface and execute no shutdown.
2. Add a default route to the hosts connected to the router (using the central router’s Ethernet address). On Unix systems, this is done under the superuser account by executing: router add default XXXX 1 (XXXX is the router’s E0 port address). A default route must be added to every host; otherwise, normal communication will fail.
3. Ping the local router’s Ethernet port. If it fails, the Ethernet port might not be activated or might not be on the same network segment. Ping the WAN port; if it fails, no default route has been added. Ping the peer WAN port; if it fails, the router configuration is incorrect. Ping the host Ethernet port; if it fails, the peer host has no default route added.
4. Add a gateway (static route) on the X.25 host’s leased line card. On Unix systems, this is done under the superuser account by executing: router add X.X.X.X Y.Y.Y.Y 1 (X.X.X.X is the peer Ethernet address, Y.Y.Y.Y is the peer WAN address).
5. Use Tracert to trace the route to identify the unreachable network segment.
Q: What is a Gateway?
——A: A gateway is a crucial component of network connection equipment. It not only has routing functions but can also convert between two different protocol sets, thereby enabling interconnection between different networks. For example, a NetWare LAN can access an IBM SNA network through a gateway, allowing PCs using the IPX protocol to communicate with IBM hosts on the SNA network.
Q: What is a Switch?
——A: A switch, also called a switching hub, regenerates information, processes it internally, and forwards it to designated ports. It possesses automatic addressing capabilities and switching functions. Because a switch transmits each information packet independently from the source port to the destination port based on the packet’s destination address, it avoids collisions with other ports. Therefore, switches can simultaneously transmit these information packets without interfering with each other, preventing transmission collisions and improving actual network throughput.
Q: What is Cascading?
——A: Cascading is a method of connecting devices that need to be cascaded through their cascade ports using twisted-pair cables, thereby increasing the number of ports<<
0, but instead of splitting data into chunks, it splits the bits of each byte across multiple disks. This increases management overhead, but if one disk fails, it can be replaced, and data can be rebuilt from parity and error-correcting codes. RAID 5 handles all read/write operations. It requires three to five disks to form the array and is best suited for multi-user systems that do not require critical features or perform very few write operations.
Other less common RAID types:
RAID Level 1 is disk mirroring — everything written to disk 1 is also written to disk 2, and data can be read from either disk. This provides instant backup but requires the highest number of disk drives and does not improve performance. RAID 1 offers the best performance and fault tolerance in multi-user systems and is the easiest configuration to implement, making it most suitable for financial processing, payroll, banking, and high-availability data environments.
RAID Level 2 was developed for mainframes and supercomputers. It can correct data without interrupting operation, but RAID 2 tends to have high data checking and error correction rates.
RAID Level 4 uses large data stripes, allowing records to be read from any drive. Because this type lacks support for multiple simultaneous write operations, it is rarely used.
RAID Level 6 has seen almost no commercial use. It extends RAID 5 by using a second parity scheme distributed across different drives. It can withstand multiple simultaneous drive failures, but performance — especially write operations — is poor, and the system requires an extremely complex controller.
RAID Level 7 features a real-time embedded operating system as the controller and a high-speed bus for caching. It provides fast I/O but is expensive.
RAID Level 10 consists of an array of data stripes, where each stripe is a RAID 1 array of drives. It offers the same fault tolerance as RAID 1 and is aimed at database servers that need high performance and redundancy but not high capacity.
RAID Level 53 is one of the latest types, implemented like a Level 0 data stripe array, where each segment is a RAID 3 array. Its redundancy and fault tolerance are the same as RAID 3. This is beneficial for IT systems requiring RAID 3 configurations with high data transfer rates, but it is expensive and has low efficiency.
WINS Service
WINS is short for Windows Internet Name Server. WINS provides name registration, renewal, release, and resolution services for NetBIOS names. These services allow the WINS server to maintain a dynamic database linking NetBIOS names to IP addresses, significantly reducing network traffic burdens.
I. Why Do We Need the WINS Service
By default, the NetBIOS name of every computer on the network is updated through broadcasts. That means if there are n computers on the network, each computer must broadcast n-1 times. For small networks, this does not seem to impact network traffic much, but for large networks, it increases the network burden. Therefore, WINS is especially important for medium and large enterprises.
II. How WINS Works
As mentioned above, the WINS server provides name registration, renewal, release, and resolution services for clients. Below is a detailed explanation of how these four basic services work:
1. Name Registration
Name registration is the process by which a client obtains information from the WINS server. In the WINS service, name registration is dynamic.
When a client starts, it sends a name registration request (including the client’s IP address and computer name) to its configured WINS server. If the WINS server is running and no other client computer has registered the same name, the server returns a successful registration message to the client computer (including the name registration’s Time-To-Live — TTL).
Just like IP addresses, every computer must have a unique computer name; otherwise, communication is impossible. If the name has already been registered by another computer, the WINS service will verify whether that name is still in use. If the name is in use, the registration fails (sending back a negative acknowledgment); otherwise, the registration can proceed.
2. Name Renewal
Because the client is assigned a TTL, its registration has a limited lifespan. After this period expires, the WINS server will delete the name’s registration record from its database. The process works as follows:
(1). After 1/8 of the TTL has elapsed, the client begins continuously trying to renew its name registration. If it receives no response, the WINS client repeats the renewal attempt every 2 minutes until half of the TTL has passed.
(2). When half of the TTL has elapsed, the WINS client will attempt to renew its lease with the secondary WINS server, following the same process as with the primary WINS server.
(3). If it still has not succeeded after half the time has passed, the client returns to its primary WINS server.
During this process, whether communicating with the primary or secondary WINS server, once the name registration is successfully renewed, the WINS client’s name registration is granted a new TTL value.
WINS Service Introduction (II)
3. Name Release
During a normal client shutdown, the WINS client sends a name release request to the WINS server to release the mapping of its IP address and NetBIOS name in the WINS server database. Upon receiving the release request, the WINS server checks whether that IP address and NetBIOS name exist in its database. If they do, the release proceeds normally; otherwise, an error occurs (the WINS server sends a negative response to the WINS client).
If the computer does not shut down normally, the WINS server will not know that its name has been released, and the name will not expire until the WINS name registration record reaches its timeout.
4. Name Resolution
Clients need the WINS server to resolve names in many network operations. For example, when using shared files on other computers on the network, the user needs to specify two things: the system name and the share name, and the system name needs to be converted into an IP address.
The name resolution process works like this:
(1). When a client computer wants to resolve a name, it first checks the local NetBIOS name cache.
(2). If the name is not in the local NetBIOS name cache, it sends a name query to the primary WINS server (sent once every 15 seconds, for a total of three attempts). If the request fails, it sends the same request to the secondary WINS server.
(3). If both fail, name resolution can proceed through other methods (such as local broadcast, lmhosts files and hosts files, or DNS).
III. WINS Server and Client Requirements
1. Microsoft’s Requirements for Servers
(1). Provide at least one primary WINS server and one secondary WINS server to ensure fault tolerance.
(2). A single WINS server can process nearly 1,500 name registrations and about 4,500 name queries per minute. Therefore, we strongly recommend that one primary and one secondary WINS server can support up to 10,000 clients.
(3). If the WINS server is not on the same subnet as the clients, router performance must be considered.
2. Client Requirements
Almost all Microsoft clients supporting network interconnection can be WINS clients. The following lists clients that can work with WINS:
*Windows NT Server 3.5x, 4.0
*Windows NT Workstation 3.5x, 4.0
*Windows 9x/me/2000
*Windows for Workgroups with TCP/IP-32
*Microsoft Network Client 3.0 for MS-DOS
*LAN Manager 2.2c for MS-DOS
DOS-based clients can also use WINS servers for name resolution, but you must add static entries for them in the WINS server.
Q: What is “DNS”? What is its Chinese name?
A: DNS, simply put, stands for Domain Name System, translated into Chinese as “域名系统” (Domain Name System).
Q: What is the purpose of DNS?
A: In a TCP/IP-based network environment (such as the Internet), DNS is a very important and commonly used system. Its main function is to convert easy-to-remember Domain Names into hard-to-remember IP Addresses. The network host that performs this DNS service can be called a DNS Server. Basically, we usually think of DNS only as converting Domain Names to IP Addresses, and then using the found IP Address to connect (commonly known as “forward resolution”). In fact, the function of converting IP Addresses to Domain Names is also quite commonly used. When logging into a Unix workstation, the workstation performs a reverse lookup to find out where you are connecting from (commonly known as “reverse resolution”).
Q: How does DNS operate?
A: DNS operates using a hierarchical structure. For example: the Domain Name of the HIT Zidingxiang BBS is bbs.hit.edu.cn. This Domain Name, of course, did not appear out of thin air; it was delegated from .edu.cn. .edu.cn, in turn, was delegated from .cn. Where did .cn come from? The answer is from “.”, which is the so-called “root domain”. The root domain is already the highest level of the Domain Name hierarchy. And the “.” layer is managed by InterNIC (Internet Network Information Center). This is how Domain Names worldwide are delegated down level by level.
Q: When I look up a Domain Name, how does DNS find its IP?
A: For example, suppose today we are querying the Domain Name (making a DNS query) for bbs.hit.edu.cn. The DNS Server handles it like this:
(1) The computer you are using (maybe a PC or a workstation) sends a question to the DNS Server configured for this computer, asking: What is the IP address of bbs.hit.edu.cn?
(2) This DNS server first checks if the answer is in its cache. If it is, it returns the answer. If not, it starts searching from the very top. The DNS Server must have