A common old telephone system
Public switched telephone network is a network that transmits analog voice based on circuit switching technology. The number of telephones in the world has reached hundreds of millions, and it is still growing.
The only feasible way to connect so many phones together and work normally is to use hierarchical switching.
To sum up, the telephone network is mainly composed of three parts: local loop, trunk and switch. Among them, trunk lines and switches generally adopt digital transmission and switching technology, and local loops (also called subscriber loops) basically adopt analog lines. Because the local loop of P S T N is analog, when two computers want to transmit data through P S T N, they must realize the mutual conversion between computer digital signals and analog signals through M o d e m
P S T N is a circuit-switched network, which can be regarded as an extension of the physical layer. There is no upper layer protocol for error control in PS T N. After the two communication parties establish a connection, the circuit-switched mode monopolizes a channel, and when the two communication parties have no information, the channel cannot be used by other users.
Users can use ordinary dial-up telephone lines or rent special telephone lines for data transmission, and it is the cheapest to use P S T N to realize data communication between computers. However, due to the poor transmission quality and limited bandwidth of the P S T N line, and the lack of storage function of the P S T N switch, the P S T N can only be used in occasions with low requirements for communication quality. At present, the highest rate of data communication through P S T N does not exceed 5 6 K b p s.
X.25
X.2 5 is the "interface between data terminal equipment D T E and data circuit equipment DTE working in groups on public data networks" formulated by CCIT in 1970s. X. 2 5 officially became an international standard in March 1 9 7 6, and was supplemented and revised in June 1 9 8 0 and June 1 9 8 4. From the perspective of I S O/O S I architecture, X. 2 5 corresponds to the bottom three layers of the I reference model, namely the physical layer, the data link layer and the network layer.
The physical layer protocol of X. 2 5 is X. 2 1, which is used to define the physical, electrical, functional and process characteristics between the host and the physical network. In fact, few public networks support this physical layer standard because it requires users to use digital signals instead of analog signals on telephone lines. As a temporary measure, C C I T T defines an analog interface similar to the familiar R S-2 3 2 standard.
The data link layer of X. 2 5 describes the reliable transmission of data between user host and packet switch, including frame format definition, error control and so on. The X. 2 5 data link layer generally adopts high-level data link control HDLC (High-level Data LinkC o n t r o l) protocol.
The network layer of X. 2 5 describes the interaction between the host and the network, and the network layer protocols deal with issues such as packet definition, addressing, flow control and congestion control. The main function of the network layer is to allow users to set up a virtual circuit, and then send a data message with the maximum length of 1.28 bytes on the established virtual circuit. Messages arrive at their destinations reliably in sequence. X. 2 5 network layer adopts packet-level protocol (P L P).
X.2 5 is connection-oriented and supports switched virtual circuit (S V C) and permanent virtual circuit (PVC). A switched virtual circuit (S.V.C.) is established when the sender sends a message to the network requesting to establish a connection and communicate with a remote machine. Once the virtual circuit is established, data can be sent on the established connection, and the data can be guaranteed to reach the receiver correctly. X. 2 5 also provides a flow control mechanism to prevent fast senders from flooding slow receivers. The usage of Permanent Virtual Circuit (P V C) is the same as that of S V C, but it is pre-established by users and long-distance telecom companies through discussion, so it has always existed, and users can use it directly without establishing links. P V C is similar to a leased private line.
Because many user terminals don't support X. 2 5 protocol, in order to make user dumb terminals (non-intelligent terminals) access the X. 2 5 network, C C I T T has formulated another set of standards. The user terminal accesses the X. 2 5 network through a "black box" called packet assembler (PA D). The standard protocol used to describe the function of PA D is called X.3.
X. 2 8 protocol is adopted between user terminal and PA D; Another protocol called X. 2 9 is used between PA D and X. 2 5 network.
X.2 5 network is developed under the condition of poor transmission quality of physical links. In order to ensure the reliability of data transmission, it must carry out error checking and error retransmission on each link; Although this complicated error checking mechanism limits its transmission efficiency, it does provide a good guarantee for the safe transmission of user data.
The outstanding advantage of X. 2 5 network is that multiple virtual circuits can be opened on one physical circuit for multiple users to use at the same time. The network has dynamic routing function and complex and complete error correction function. X. 2 5 packet switching network can meet the data communication between terminals and computers, computers and local area networks with different rates and models. The data transmission rate provided by X. 2 5 network is generally 6 4 K b p s.
Defense data network
Digital data network (D D N) is a transmission network that provides data communication through digital channels. It mainly provides point-to-point and point-to-multipoint digital private lines or networks.
D D N consists of digital channel, D D N node, network management system and subscriber loop. The transmission media of D D N mainly include optical fiber, digital microwave and satellite channel. D D N adopts computer-managed digital cross-connect (D X C) technology to provide users with semi-permanent connection circuits, that is, the channels provided by D D D N are non-switched, and users have their own permanent virtual circuits (P V C). Once the user applies, the network administrator can change the routing or private network structure of the user's private line through software commands without going through the reconstruction and expansion project of the physical line, so it is very easy for D D N to connect to the line with the required bandwidth within the agreed time according to the user's needs.
The basic service provided by D D N is point-to-point dedicated line. From the user's point of view, renting a point-to-point private line is to rent a digital channel with high quality and high bandwidth. Users renting a point-to-point digital private line on D D N is very similar to renting a telephone private line. The difference between D D N private line and telephone private line is that telephone private line is a fixed physical connection, and telephone private line is an analog channel with narrow bandwidth, poor quality and low data transmission rate. D D N private line is a semi-fixed connection, and its data transmission rate and routing can be changed at any time as needed. In addition, D D N dedicated line is a digital channel with high quality and wide bandwidth, which adopts thermal redundancy technology and has the function of automatic detour of routing failure.
Let's introduce the difference between d D D N and X. 2 5 network. X. 2 5 is a packet-switched network. X. 2 5 network itself has three layers of protocols, and temporary virtual circuits are established by calling. X. 2 5 has the functions of protocol conversion and speed matching. It is suitable for mutual communication between user equipments with different communication specifications and different rates. However, D D N is a fully transparent network with no switching function. The main way to use D D N is to rent a dedicated line regularly or irregularly. Judging from the cost that users need to bear, X. 2 5 charges by byte, and D D N charges by fixed monthly rent. Therefore, D D N is suitable for data communication between LANs or hosts that need frequent communication. The data transmission rate provided by the D D N network is generally 2 M b p s, and the highest can reach 4 5 M b p s or even higher.
father
Frame Relay (F R) technology is developed from X. 2 5 packet switching technology. The introduction of F R is due to the change of communication technology in the past 2 0 years. Twenty years ago, people used slow, analog and unreliable telephone lines to communicate. At that time, the computer processing speed was slow and the price was relatively expensive. Therefore, a very complicated protocol is used to deal with transmission errors in the network, so as to avoid user computers from recovering from processing errors.
With the continuous development of communication technology, especially the extensive use of optical fiber communication, the transmission rate of communication lines is getting higher and higher, but the bit error rate is getting lower and lower. In order to improve the transmission rate of the network, Frame Relay technology omits the functions of error control and flow control in X. 2 5 packet switching network, which means that Frame Relay network can use simpler communication protocol when transmitting data and leave some work to users, which makes the performance of Frame Relay network better than that of X. 2 5 network, and it can provide the data transmission rate of 1. 5 million pounds per second.
We can think of Frame Relay as a virtual private line. Users can rent a permanent virtual circuit between two nodes and send a data frame through the virtual circuit, and the length of the data frame can reach 1 6 0 0 bytes. Users can also communicate between multiple nodes by renting multiple permanent virtual circuits.
The difference between the actual leased private line (D D N private line) and the virtual leased private line is that: for the actual leased private line, users can continuously send data at the highest data transmission rate of the line every day; For virtual leased lines, users can send data at the peak rate of the line within a certain period of time, but the average data transmission rate of users must be lower than the pre-agreed level. In other words, long-distance telecom companies charge less for virtual private lines than for physical private lines.
Frame relay technology only provides the simplest communication processing functions, such as the determination of the start and end of frames and the error checking of frame transmission. When a Frame Relay switch receives a corrupted frame, it discards it. Frame Relay technology does not provide acknowledgement and flow control mechanism.
Virtual circuit multiplexing technology is used in frame relay network and X. 2 5 network to make full use of network bandwidth resources and reduce user communication costs. However, because the frame relay network does not correct the wrong frames and simplifies the protocol, the time required for processing data frames by the frame relay switch is greatly shortened, the end-to-end user information transmission delay is lower than that of the X. 2 5 network, and the throughput of the frame relay network is also higher than that of the X. 2 5 network. Frame Relay network also provides a complete set of bandwidth management and congestion control mechanism, which has advantages over X. 2 5 network in dynamic bandwidth allocation. Frame Relay networks can provide virtual private lines with rates ranging from 2 megabits per second to 4 5 megabits per second.
Switched Multimegabit data service (realized by telephone network)
Switched multi-megabit data service (S M D S) is designed to connect multiple LANs. It was developed by B e l l c o r e in the 1980s and implemented in some areas in the early 1990s.
To illustrate the usage of S M D S, let's look at an example. Suppose a company has four offices in four cities, and each office has a local area network. The company decided to connect the four LANs. One solution that can be adopted is to rent six high-speed private lines to connect the four LANs with each other, as shown in Figure 4-1a. This scheme is feasible, but the cost is too expensive.
Another method is to use S M D S, as shown in Figure 4- 1 B. We can think of S M D S as a high-speed backbone network between LANs, that is, one LAN is allowed to send messages to other LANs through S M D S, and the short-distance line between LANs and SMD S (as shown by the thick line in Figure 4-1b) can be rented from the telephone company. Usually this section of the line uses M A N's D Q D B protocol, but it is also feasible to use other types of protocols.
Figure 4- 1 Two different schemes for connecting four LANs Although most telephone companies provide services for continuous communication, the design of S M D S is for burst communication. In other words, sometimes one LAN must send data messages to another LAN quickly, and more often there is no data to be transmitted between LANs. Figure 4- 1 a The solution using leased lines has the following problems: once leased lines are used, users must pay a high monthly fee for each line regardless of whether they have been using these lines all the time. For intermittent communication, renting lines is a relatively expensive scheme, and S M D S is more competitive than it in cost. If there are n LANs, you need to rent n(n- 1)/ 2 long-distance private lines to connect them completely, and you only need to rent n short-distance lines to connect the LANs to the S M D S router.
Because the design goal of S M D S is to be used for communication between L A N and L A N, its data transmission speed must be high enough. The standard rate is 45mb/s, and the rate lower than 45mb/s is also feasible.
S M D S provides connectionless message transmission service. The format of safety management information is shown in Figure 4-2. The S M D S message has three fields:
A destination address field, a source address field and a variable-length user data field. The maximum length of user data can reach 9 1 8 8 bytes. The machine on the sender L A N sends the message to the telephone company's S M D S switch through the access line, and S M D S tries its best to deliver the message to the destination node, but it does not guarantee that the message is delivered correctly.
Figure 4-2 SMDS frame format
The source address and destination address include a 4-digit binary code and a 1 5-digit decimal telephone number. Each decimal number is individually coded as a 4-bit binary number. The telephone number consists of country code, area code and user number, which means that international services can be provided to users.
Whenever a message arrives at SMD's network, the first router of SMD is responsible for checking whether the source address of the message corresponds to the inbound line to prevent being cheated when charging. If the address is wrong, the message will be discarded; If the address is correct, the message will be sent to the destination node.
A very useful function of S M D S is broadcasting. Users can define a group of telephone numbers of S M D S and assign a special number to the whole group. Any message sent to this special number will be sent to all members of the group.
Another useful feature of S M D S is the address masking of inbound and outbound messages. To mask the output address, LAN1lan2lan3lan4lan4lan1lan2SMDS a) connect four LANs with leased lines b) connect four LANs with smds.
destination address
Number of bytes 8 8 ≤9 188
Source address user data
Users can specify a set of telephone numbers, thus limiting users to output messages only to the specified addresses (telephone numbers); Similarly, for input address masking, users can restrict incoming calls from external users by specifying a set of phone numbers.
Using this feature of S M D S, users can establish a private network.
The payload part of an S M D S frame can be any byte sequence, and the maximum length of this field is 9 1 8 8 bytes.
The data field of an SMS frame can carry Ethernet message, I. B. M token ring message and I.P. message, etc. That is, S M D S only transmit data from the source L A N to the destination L A N without modification (transparently).
S M D S handles burst communication as follows. The router connected to the user's access line contains a counter that increments at a fixed rate, for example, every 10μ s plus 1 ... Every time the router receives a message, it will check the value of the counter and compare it with the length of the message just received (in bytes). If the value of the counter is greater than the number of bytes of the message, the message will be sent immediately, and the count value of the counter will be subtracted from the number of bytes of the message. If the message length is greater than the counter value,
The message was discarded.
In fact, according to the counting frequency of 1 0 μ s plus 1, users can send data at the average rate of100,000 bytes/second, but the burst data rate may be higher than this. For example, if the idle period of the user's access line is 1 0 0 0 ms, the value of the counter is1000, so the user can send 1 K bytes of data at the transmission rate of 4 5 M b p s, and the transmission time required by the router is180 μ s. For the leased line with 100 000 bytes/sec and the same data with 1 K, SMD/S provides a slight delay for users' requirements for various data communication rates as long as the average data rate of users is always kept at the pre-agreed data rate. This mechanism provides a quick response for users who need to send data, while preventing users from using more bandwidth than they agreed to pay in advance.
Through the previous analysis, we already know that the data transmission rate supported by S M D S is higher than that of Frame Relay, but S M D S is connectionless.