A, category
(a) the definition of computer network
1, the precise definition of computer network is not regulated, and
2, in which the good is defined as: computer network mainly by some common, programmable hardware connected together, but the hardware is not designed to achieve a specific purpose (for example: the transfer data or video signal), the programmable hardware can be used to transmit a variety of different types of data, and can support a wide range of application and the increasing,
3, according to this definition, we want to know:
(1) computer network connection by hardware, not limited to general computer, but includes smartphones,
(2) a computer network is not designed to transmit data, it can support a wide range of applications in the future (including all kinds of application),
(3) which says "programmable hardware" shows the hardware must contain a central processing unit (CPU),
(2) the computer network has the following several different categories:
1, according to the scope of the network are classified:
(1) WAN WAN (Wide Area Network) : the scope of dozens to several thousand kilometers, usually
(2) the MAN MAN (Metropolitan Area Network) : the range of about 5 to 50 km,
(3) LAN LAN (Local Area Network), confined to smaller range (such as km),
(4) Personal Area Network PAN (Personal Area Network) : the range is very small, about 10 meters,
If the distance between the central processor is very close (such as only in the order of one meter or less), is commonly called a multiprocessor system, rather than call it a computer network,
(3) according to network users are classified:
1, the public network (public network) : in accordance with the provisions, pay the fee can use network, also known as public sites,
(private network) : 2, private networks built for the needs of the special operation network,
Public networks and private networks (VPNS) can provide a variety of services, such as the transfer is the computer data, respectively is public computer network and special purpose computer network,
3, to the user access to the Internet network:
(1) Access the AN (Access Network) is also known as the local Access Network or residents Access Network, it is a kind of special computer Network, is used to connect to the Internet,
(2) access network itself is not part of the core of the Internet, also do not belong to the edges of the Internet, it was the first of a client system to the Internet router (also known as the edge router) a network, between
(3) from the scope of cover, a lot of access network or local area network (LAN),
(4) from the point of view of function, the access network is just a let users can connect to the Internet "bridge" role,
Second, architecture
(a) the formation of the computer network architecture
1, the computer network is a very complicated system,
2, communicate with each other two computer systems must be highly coordinated work, and this kind of "coordination" is quite complex,
3, "layered" large and complex problem can be transformed to a number of smaller local problem, which makes it easier to study and deal with,
4, the System Network Architecture SNA (System Network Architecture) : it is a in the United States in 1974, IBM announced according to the famous Network standard, the layered method after a lot of companies have been also introduced their company's Architecture with different names, but because of the different Network Architecture, equipment is difficult to connect with each other between different companies, in 1977, the international organization for standardization ISO set up specialized agencies to study how to make different Architecture of computer networks can be interconnected,
5, basic Reference Model of Open System Interconnection OSI/RM (Open System Interconnection Reference Model) : it is an attempt to make a variety of computer interconnected into a worldwide network of standard framework, referred to as the OSI, as long as you follow the OSI standard, a System can be located anywhere in the world, and also follow the standards of any other System for communication, but the OSI only had some theoretical research results, when advancing market fails, the reasons for the failure of it include:
(1) the OSI experts in completing the OSI standard no business driving force;
(2) of the OSI protocol implementation up is too complex, and the efficiency is low;
(3) of the OSI standard setting cycle is too long, and make the production by the OSI standard equipment can not be in time to enter the market;
(4) the different layers of the OSI also is not very reasonable, some features are repeated in multiple layers,
Note: the law (DE says jure) international standard OSI has not been recognized by the market, but the international standard TCP/IP won the most widely used, therefore, TCP/IP is often referred to as a matter of fact (DE facto) international standards,
(2) protocol and division level
1, the computer in the network data exchange must abide by the rules agreed upon in advance, these regulations specify the format of data exchange and synchronization problems associated (synchronous timing), and network protocol (network protocol), referred to as "agreement, means for the network data exchange and to establish the rules and standards or agreed,
2, the network protocol is made up of syntax, semantics and synchronization of three elements, among them, the syntax and control index according to the structure or format of information, the semantics is to point to need a what kind of control information, finish what response, what kind of action and making synchronous implementation order refers to the event details, therefore, the network protocol is an integral part of computer network,
3, the agreement has two forms: one is to use the convenient for people to read and understand the text, but use, is to let the program code, the computer can understand the two different forms of agreement must be able to make a precise explanation on the network information exchange process,
4, by the development of the ARPANET experience has shown that for very sophisticated computer network protocol, its structure should be hierarchical, layered advantage is:
(1) between each layer is independent,
(2) good flexibility,
(3) the structure can be separated,
(4) easy to implement and maintain,
(5) can promote the standardization,
Its defect is:
(1) to reduce the efficiency,
(2) some functions will be repeated in different levels, and thus will produce extra overhead,
5, when layered though not the more the better, but also cannot too little, therefore, layer number of how many will be appropriate, layer number of too little words can make the protocol of each layer is too complex, and the layer number is too much, and in the description and combines the functions of each layer systems engineering tasks encountered many difficulties,
6, the main function of each layer to complete:
(1) error control: make the corresponding arrangement peer communication is more reliable,
(2) flow control: send the send rate must make the receiver could receive, not too fast,
(3) segmentation and reshipment: the sender to send blocks of data are divided into smaller units, at the receiving end to restore,
(4) the reuse and points to use: sending a few high-level session reuse a low-level connection, then points at the receiving end,
(5) the connection is established and release: establishing a logical connection before data exchange, data transfer after the release connection,
(3) deal with five layers architecture
1, the computer network system structure (architecture) is a collection of layers and the protocol of computer network, is the computer network system structure and components should be completed by the precise definition of function, and realize the (implementation) is follow the premise of this architecture with what kind of hardware or software to complete the function of the problem, therefore, architecture is abstract, and the implementation is concrete, is really in the operation of computer hardware and software,
2, the OSI seven layer protocol architecture is the concept of clear, theory is also relatively complete, but it is complex and not practical, TCP/IP is an application layer, transport layer, the layer and network interface layer of four layers architecture, but it doesn't the network interface layer on the bottom of the specific content, therefore, we often take a compromise, which integrated the advantages of the OSI and TCP/IP, adopts a kind of only five layer protocol architecture, it has the application layer, transport layer, network layer, data link layer and physical layer,
3, PDU (Protocol Data Unit) : Protocol Data Unit, the OSI reference model called peer level Unit of Data passed between the layer Protocol Data Unit of PDU, the term has been used by the number of OSI standard,
4, any two of the same level of the data (i.e., data unit and control information) by horizontal dotted line directly to each other, this is called "peer layer" (peer the layers), the communication between each layer protocol is actually in each peer transfer data between the provisions of,
Note: the following three layer called low-level protocols, and four floors above is called high-level protocol
(4) the entity, the agreement, the service and service access point
1, the entity (entity) said any hardware or software can send or receive information process,
2, the agreement is to control the two peer entities a collection of rules for communication,
3, under the control of the agreement, communication between the two peer entity makes this layer to up a layer provides services,
4, to implement the protocol, also need to use the services provided by the lower,
5, protocol and service is different in concept and implementation of the protocol ensures to provide services to up a layer, the layer of service users can only see and can't see the following agreement, namely the following agreement for the above service user is transparent, agreement is "level", the agreement is to control the rules of communication between peer entities, service is the "vertical", as a service is provided by the lower level to the upper layer through indirect mouth, and the upper is through the use of the services provided by the service primitives to obtain lower,
(5) the architecture of TCP/IP
TCP/IP architecture another way said: in fact, now the Internet use TCP/IP architecture has evolved, sometimes some applications can directly use the IP layer (network layer), or even directly using the network interface layer on the bottom of the