Shulou(Shulou.com)06/01 Report--

Internet architecture

Computer network is a group of computer systems connected by a certain form, which needs the support of four elements, namely, communication lines and communication equipment, computers with independent functions, support of network software, and the ability to realize data communication and resource sharing. Computer network has two reference models, OSI model and TCP/IP model, in which OSI model is the theoretical model, and TCP/IP model has become the industrial standard of Internet facts. nowadays, communication networks generally use TCP/IP protocol suite, while application programming uses socket sockets for programming.

The background of Network Architecture

The proposal of network architecture stems from the complexity and heterogeneity of computer network, which is embodied in the following aspects:

Different communication media-wired, wireless …

Different kinds of devices-hosts, routers, switches, multiplexing devices.

Different operating systems-- Unix, Windows …

Different software / hardware, interfaces, and communication protocols (protocols)

Different application environments-fixed, mobile …

Different types of services-time-sharing, interaction, real-time …

For complex network systems, what method can be used to organize the structure of the network reasonably to achieve:

The structure is clear

Simplify design and implementation

Easy to update tail maintenance

Strong independence and adaptability

Answer: divide and rule

A living example: logistics system

Logistics system organization:

User (sender)

User (pickup)

Courier (receiving mail)

Courier (pick up the letter)

Sub-inspector (sub-inspection)

Sub-inspector (sub-inspection)

Means of transport (departure)

Means of transport (arrival)

Transportation line

Features:

Each layer implements a specific service.

Through their own internal functions.

Rely on the services provided by your own lower layer.

The layered method is also used in the computer network. -- complex problems are divided into a number of smaller, single local problems, which are solved at different levels.

Hierarchical network architecture

In computer network, the collection of layers, protocols and interlayer interfaces is called computer network architecture. In other words: the architecture consists of three parts: hierarchical structure and functions of each layer, services and inter-layer interfaces, and protocols.

1. The layering principle of network architecture

Entity: any hardware / software process that can send or receive information.

Peer-to-peer layer: the same level of two different systems.

Peer entities: two entities in the peer layer of different systems.

Interface: an interactive interface between two adjacent layers, defining the operation between the two adjacent layers and the services from the lower layer to the upper layer.

Service: a capability of a layer and below that is provided to its adjacent upper layer through an interface.

Protocol: rules that must be obeyed by both parties in communication.

Hierarchical network structure diagram:

Figure 1

Any system in the network is organized according to the hierarchical structure.

In the same network, any two end systems must have the same level

Each layer uses the services provided by its lower layer and provides services to its upper layer

Communication only takes place between peer layers (indirect, logical, virtual), and non-peer layers cannot "communicate" with each other.

The actual physical communication is only done at the bottom.

Pn: layer n protocol, that is, the rules or conventions that must be followed when communicating between layer n peer entities

2. The advantages of hierarchical architecture:

(1) strong independence-low degree of coupling

The upper layer only needs to know what services the lower layer provides through the interlayer interface-the black box method.

(2) strong adaptability

As long as the services and interfaces remain unchanged, the implementation method of each layer can be changed at will.

(3) easy to implement and maintain

The complex system is divided into several sub-units with small scope and simple function.

Network system structure reference Model ISO/OSI

The Information processing Systems Technical Committee of the International Organization for Standardization (ISO TC97) established the subcommittee SC16 for open systems interconnection in 1978 and published the first reference model for open systems interconnection in December 1980.

The architecture of OSI reference model is called physical layer, data link layer, network layer, transport layer, session layer, presentation layer and application layer from low layer to high layer.

1. OSI architecture diagram

Figure 2

The function of the OSI network reference model says:

Layer name

Function

Corresponding problems

Application layer

Interface with user application process

"do what?"

Presentation layer

Conversion of data format

"what does the other person look like?"

Session layer

Synchronization of call management and data transmission

"who should speak and where to start?"

Transport layer

Reliable end-to-end data transmission

"where is the other party?"

Network layer

Packet delivery, routing, flow control

"go that way."

Data link layer

Error-free transmission frame between adjacent nodes

"how to take each step?"

Physical layer

Transparently transmit the bit stream on the physical node

"how to use physical Media"

Data flow of OSI Model

2. The function of each level of OSI.

(1), physical layer

The physical layer is the lowest layer in the OSI model, which is directly connected to the transmission medium, and its main function can be described as transparently transmitting the bit stream on the transmission medium connected to various computers. The service provided by the physical layer to the data link layer above it is to shield these differences from the data link layer, so that the data link layer only needs to focus on the task of data transmission on the link.

ISO defines the function of the physical layer in the OSI model as providing mechanical, electrical, functional and normative characteristics by establishing, maintaining, and releasing physical connections between data link entities, which are used for the transmission of binary bitstreams between data link entities. That is, the data units of the physical layer are bits.

(2), data link layer

The layer above the physical layer is called the data link layer, and its main function can be simply described as the error-free transmission of data on the line between two directly adjacent network nodes. The data unit it transmits is the protocol data unit (PDU) of the data link layer, which is called Frame. The fundamental purpose of this layer function is to achieve reliable data transmission on less reliable physical lines, that is, the data link layer provides reliable data communication between directly adjacent nodes in the network.

The concern of the data link layer is how to ensure that the data is transmitted correctly on the data link. If the data can be transmitted correctly on each link that makes up the channel, the feasibility and accuracy of information transmission can be ensured at least on the data link layer.

(3) Network layer

The computer network layer above the data link layer is the network layer. The main purpose of setting this level is to transfer user data from source to destination.

The main functions to be realized at this level.

First of all, the core function of the network layer is the choice of routing. Routing is a path from source to destination, and routing is to determine a transmission path for user data from source to destination. The data unit processed by the network layer, that is, the protocol data unit of the network layer, is called a packet or Packet.

Secondly, determining a route is not a simple task, and we need to take into account the heterogeneous networks where packets may go through different topologies, use different protocols and have different basic parameters.

Third, the link resources of communication should be shared, but the problem caused by sharing is that congestion may occur.

(4), transport layer

The transport layer, also known as the transport layer, is a layer above the network layer. It occupies an important position in the whole network architecture-it is located at the bottom of the resource subnet and directly connected to the communication subnet. it is the transfer layer of application-oriented services and communication-oriented services.

The main function of the transport layer is to provide end-to-end communication between the source and destination host processes. That is, the transport layer receives data from the session layer, decides whether to split it into smaller units according to the actual situation, and then passes it to the network layer to ensure that the data arriving at each other is correct. The data units processed by the transport layer are called messages

(5) session layer

The session layer is located above the transport layer and is used to establish a session (Session) relationship between user processes on two different computers. A session is defined as an information exchange between two user processes on different computers, usually for data transfer similar to the transport layer, such as transferring some data required by the user, including files, etc.

One of the services provided by the session layer is managing sessions. This feature includes establishing, using, and ending sessions between two user application processes on different computers.

The second service provided by the session layer is token management. A token is a special kind of data, and only the party who owns the token has the right to perform the operation.

The third service provided by the session layer is synchronization.

(6), presentation layer

The presentation layer is located above the session layer and is the sixth layer of the OSI model. It is used to perform some general information processing operations to reduce the complexity of the user's work. In the OSI model, the presentation layer is concerned with the syntax and semantics of the transmitted information, while the following layers are concerned with the correct transmission of information data, which is the obvious difference between the presentation layer and the following layers.

One of the typical general information processing operations performed by the presentation layer is to encode the information data using standard methods.

(7), application layer

The application layer is the highest layer of the OSI model, which provides a large number of application protocols to meet people's different network needs. Through the interfaces supported by various application protocols, network users can use all kinds of network services provided by these protocols and access various resources of computer networks. Based on these protocols, network applications can be further developed to meet their special needs.

In the OSI model, different protocols in the application layer provide information access means for specific network applications. As long as the two sides of the application conform to the specifications of a certain protocol, they can use the network services provided by the protocol.

The formulation process of ISO's OSI/ RM seven-layer reference model is delayed, the protocol architecture is too complex, and there is too much duplication of hierarchical functions. Therefore, the widely used TCP/IP protocol architecture has gradually become a de facto industrial standard followed by the majority of computer manufacturers and the computer science community. However, the OSI/ RM seven-layer reference model is still used in scientific research because of its complete content and clear structure.

TCP/IP network architecture

TCP/IP is a four-layer architecture, which are: application layer, transport layer, Internet layer and network interface layer. If user data is to be transferred from the source computer to the destination computer using the TCP/IP protocol, it must be processed by the above four-layer network protocol stack before it can be transmitted in the actual physical network. But in fact, because there is no specific content in the lowest network interface layer, the TCP/IP protocol architecture is only described in detail in the application layer, transport layer and internetwork layer.

1. Application layer

The function of the application layer (Application Layer) is to provide network applications for users and to provide applications with the ability to access other layer services, that is, to send users' data to the layers under the TCP/IP model and to provide network interfaces for applications. Because the TCP/IP model assigns all application-related content to the application layer for processing, there are a large number of applications and protocols in this layer.

2. Transport layer

The transport layer is responsible for providing reliable, end-to-end data transmission between two host processes, that is, communication between application processes on one host and application processes on another host.

Two transport layer protocols, Transmission Control Protocol (TCP) and user Datagram Protocol (UDP), are defined in the TCP/IP model, which provide two different data transmission services.

(1) Transmission Control Protocol

Transmission control protocol TCP provides connection-oriented services to ensure reliable end-to-end data transmission.

(2) user Datagram protocol

The user Datagram protocol UDP provides connectionless services. It can ensure the efficient transmission of independent data packets, the network overhead is small, and the robustness of information transmission is strong.

3. Internet layer

It can be said that the internetwork layer is the core layer of the TCP/IP model, which is mainly responsible for the interconnection of all kinds of networks supporting TCP/IP protocol. Specifically, the core function of the internetwork layer is routing, that is, addressing according to the IP address of the destination host and choosing an appropriate path for data packet transmission. However, the IP protocol of the Internet layer provides best-effort delivery services, that is, when packets pass through the network, they may go wrong or even lose because of network congestion or other failures. Moreover, the IP protocol only has limited error detection capability, and the error control function of the data packet must be completed by the transport layer protocol.

4. Network interface layer

In the TCP/IP model, the network interface layer is located at the bottom of the whole model, and is responsible for receiving IP packets from the internetwork layer and sending IP packets to the network transmission media, and receiving data streams from the network transmission media, extracting IP packets and submitting them to the internetwork layer.

The TCP/IP standard does not define a specific network interface protocol, and its purpose is to include all the protocols that enable the TCP/IP protocol stack to communicate with the physical network, so as to enhance the flexibility and adaptability of the TCP/IP model for various networks.

Comparison of OSI and TCP/IP Architectur

First of all, the application layer of TCP/IP model includes the functions of application layer, presentation layer and session layer of OSI model.

Secondly, the transport layer and network layer of the two models can almost completely refer to each other, indicating that the end-to-end host at the bottom of the resource subnet

The transmission between processes and the transmission between the network nodes at the top of the communication subnet should be clearly separated.

Third, the TCP/IP model has only one undefined network interface layer, while the OSI model completely defines the data link layer and the physical layer.

The TCP/IP model does not clearly distinguish the concepts of service, interface and protocol, while the OSI model does detailed work, which meets the specifications and requirements of software engineering practice.

The TCP/IP model is dedicated and is not suitable to describe any protocol other than the TCP/IP model, while the OSI model is a general standard model framework that can describe any protocol that conforms to the standard.

The TCP/IP model focuses on the interconnection of heterogeneous networks, while the OSI model begins to pay little attention to this point.

The TCP/IP model provides both connection-oriented and connectionless services, while the OSI model begins to consider only one kind of connection-oriented service.

The TCP/IP model provides a strong network management function, but the OSI model only considers this problem later.

Please indicate the source of the original article reproduced by Xu Hao, School of Software, University of Science and Technology of China.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.