Shulou(Shulou.com)06/03 Report--

6.1 introduction

6.1.1 Network standard

1. Performance

two。 Reliability.

3. Safety

6.1.2 physical structure

Connection Typ

A network consists of two or more devices connected through a link.

A link is a communication channel through which data is transmitted from one device to another.

Point-to-point connection: provides a dedicated link between two devices

Multipoint connection: two or more specified devices share a link physical topology: refers to the physical layout of the network.

Two or more devices are connected to a link, and one or more links form a topology.

There are four possible basic structures: mesh, star, bus and ring.

In a mesh topology, each device has a dedicated point-to-point link to every other device.

In a star topology, each device has a dedicated point-to-point link that is only connected to the hub.

In a bus topology, a multipoint link is used and a long cable (called a bus) connects all the devices in the network. The node is connected to the bus using branch lines and connectors.

In a ring topology, each device has a dedicated point-to-point link that is connected only to devices on both sides. The signal travels from one device to another in only one direction along the loop until it reaches its destination. Each device in the ring is connected to a repeater. The advantages and disadvantages of the mesh topology ensure that each connection can transmit its own data load and eliminate traffic problems. If a link cannot be used The main disadvantages of the entire network will not be affected by the number of cables and the number of input / output ports required. The star topology is easy to install and reconfigure. The fracture of the simplified fault isolation ring will make the entire network unavailable. Bus topology installation simple bus cable applause or break will terminate all transmission ring topology with mesh topology, most of its advantages are cheap, easy to install, good scalability, the entire topology depends on a single point (hub). If the hub goes down, the whole network will not work.

6.1.3 Network classification

Local area network (LAN)

It is often privately owned, connecting devices in a single office, building or campus.

LAN is designed for resource sharing between personal computers or workstations. Shared resources include hardware (such as printers), software, or data. Wide area Network (WAN)

Metropolitan area Network (MAN) that provides long-distance transmission of data, image, audio and video information

A network with a size between LAN and WAN. Usually covers a town or a city.

6.1.4 Internet

Networks are all interconnected, and when two or more networks are connected, they become the Internet.

6.1.5 Internet

It is made up of thousands of interconnected networks.

6.2 TCP/IP protocol family

Allow Lans and WANs that use different technologies to connect to each other.

The TCP/IP protocol family is usually defined as five layers

Application layer 5

Transport layer 4

Network layer 3

Data Link layer 2

Physical layer 1

The figure above shows the layers involved when sending from device A to device B. it may pass through many routers, which only use the first three layers.

6.3 layer

6.3.1 Application layer: responsible for providing services to users

Allow people or software to access the network.

Provide support for e-mail, remote file access and transfer, browsing the World wide Web and other services.

The application layer is the only layer that most Internet users can see.

1. Client / server architecture

Each application consists of two separate but related programs: the client program and the server program.

Server-side programs need to run all the time. The computer running this program is called a server.

Client programs run only when needed

The communication between client and server programs is called process-to-process communication.

two。 Application layer address

When the customer needs to send a request to the server, it needs the address of the server's application layer.

URL (uniform Resource Locator) is used to identify a particular HTTP site.

The server application layer address is not used to send messages, but to help customers find the actual address of the server computer.

Every computer in the network has an address called a logical address or IP address

The server application layer address can help the client find the IP address of the server computer. The client process should already know the DNS (Domain name Server) address. These servers match the domain name to the directory of the IP address. The client prepares and sends messages to the DNS server, asking for the actual IP address of the server it needs. When the response is received, the client server knows the IP address of the desired server.

6.3.2 Transport layer: responsible for the logical transmission of messages between client and server processes

Establish logical communication between the transport layer of the client and the server computer.

1. Address of the transport layer (port number)

When a message reaches the server, it must be pointed to the correct process. At the same time, another address is needed to identify the server process, called the port number. Similar to sending a message to someone in a building, you need to know the address and room number of the building. The IP address is similar to the building address and the port number is similar to the room number.

Although the client port number can be temporarily specified by the computer running the client process, the Internet limits the range of temporary port numbers to avoid breaking the well-known port address range.

two。 Multiplexing and demultiplexing

Just like the doorman in a residential district, the resident can give the mail to the doorman, the doorman gives the mail to the postman (multiplexing), and the doorman distributes the received mail to the residents (demultiplexing).

The transport layer uses the port number (similar to the apartment number used by the doorman) for multiplexing and demultiplexing.

The server process uses the well-known port number, while the client process uses the temporary port number specified by the transport layer.

3. Congestion control

Traffic congestion may occur in the lower-layer network where packets are physically transmitted. This may cause the network to drop (lose) some packets.

Some protocols use buffers per process, and messages are stored in the buffer before they are sent. If the transport layer detects congestion on the network, it suspends transmission.

4. Flow control

The transmission layer of the sender can monitor the transport layer of the receiver and check whether the data packets received by the receiver are excessive.

5. Error control

During the transmission of a message, it may be damaged, lost, repeated or out of order. The transmission of the transport layer is responsible for ensuring that the message is received correctly by the destination transport layer.

The transport layer can keep a copy of the message in the buffer (temporary storage) until it receives confirmation from the receiver that the packet arrives without damage and is in the correct order.

If there is no acknowledgement within the expected time, the sender resends the packet.

6. Transport layer protocol

UDP (user Datagram Protocol)

Complete multiplexing and demultiplexing.

Error control is carried out by adding a checksum to the packet. The receiver recalculates the checksum, and if the receiver concludes that the packet is damaged, it will drop the packet without notifying the sender to resend it.

UDP is fast and efficient, and can control flow or error on its own.

DNS server, which uses UDP to process video.

UDP is called a connectionless protocol because UDP does not provide logical connections between packets that belong to a single message. In other words, each package in UDP is a separate entity with no sequence number. For example, if you need to send a group of orderly parcels to your destination, the post office cannot guarantee that these parcels will be distributed in order. TCP (Transmission Control Protocol)

A protocol that supports all responsibilities of the transport layer.

Not as fast and efficient as UPD.

TCP uses serial numbers, confirmation numbers, and inspection and.

The sender also uses buffers, which provide multiplexing, demultiplexing, flow control, congestion control, and error control.

TCP is called a connection-oriented protocol because it provides a logical connection between the two transport layers. If the packet arrives in the wrong order or is lost, it will be retransmitted. The transport layer at the receiving end does not send out-of-order packets to the application, but retains all packets in the message until they are accepted in the correct order. SCTP (flow Control transfer Protocol)

It combines the advantages of UDP and TCP. It is suitable for real-time transmission of audio and video, as well as error control and flow control.

6.3.3 Network layer: responsible for sending a single packet from the source host to the destination host

Ensure that each packet goes from the source point to the final destination, and it may span multiple networks.

1. Network layer address

Packets from the client to the server and packets returned from the server require a network layer address. The address of the server is provided by the server, while the client address is known to the client computer.

The network layer uses its routing table to find the logical address of the next (router) and pass this address to the data link layer. Use this logical address required by the data link layer to find the data link layer address of the next router.

two。 Routing: refers to some or all of the paths of the snowtop packet

Because the Internet is a collection of networks (LAN, WAN, and MAN), packet delivery from source to destination may be a combination of several transmissions:

Source-to-router transmission, router-to-router transmission, router-to-destination transmission.

When a router receives a packet, it examines the routing table to determine the best route for the packet to its final destination. The routing table provides the IP address of the next router, and when the packet arrives at the next router, the next router makes a new decision.

Note: routers only use the first three layers of TCP/IP

The router updates the routing table according to the routing protocol. This protocol sends its own messages to all routers on the network, updating their information about routing.

Routing protocols are: RIP, OSPF and BGP

3. Network layer protocol

TCP/IP supports a primary protocol (IP) and several auxiliary protocols to help IP fulfill its responsibilities

IP (Internet Protocol)

IPv4 is responsible for sending packets from the source computer to the destination computer. Computers and routers around the world are identified by 32-bit IP addresses, represented by dotted decimal notation.

At the source of the message, IPv4 adds the source and destination IP addresses to the packets sent from the application layer. Auxiliary protocol

ICMP (Internet Control message Protocol): used to report a certain number of errors to the source computer. For example, if a router loses a packet due to congestion, ICMP can send a packet to the source computer to warn it of congestion.

IGMP (Internet Group Management Protocol): used to increase the multicast capability of IP. IP is essentially a unicast transmission protocol, one source, one destination. Multicast transmission refers to one source and multiple destinations.

There are other auxiliary protocols, such as ARP (address Resolution Protocol) and RARP (reverse address Resolution Protocol).

6.3.4 data Link layer: responsible for node-to-node transmission of data frames

1. Data link layer address

One device can find the data link layer address of another device statically or dynamically.

In the static method, the device creates a table with two columns for storing network layer and data link layer address pairs.

In the dynamic method, the device can broadcast a specific packet containing the IP address of the next device and ask the neighboring node with this IP address, and the neighboring node returns its data link layer address.

Data link layer addresses are not universal, and each data link protocol may use a different address format and size. The Ethernet protocol uses a 48-bit address and is usually written in hexadecimal format, such as 07:01:02:11:2C:5B. Data link layer addresses are called MAC addresses (physical addresses or media access control)

two。 Error control and flow control (some data link layer protocols are used in the data link layer)

The method is the same as the transport layer, but it is only implemented between the node emitting point and the node arriving point. Errors are checked multiple times, but none of them cover errors that may occur inside the router.

6.3.5 physical layer: performs the functions required to transmit a binary stream on a physical medium.

Responsible for the transmission of a single binary bit that makes up a frame from one node to another. In other words, the unit of transmission at the data link layer is the frame, and the unitary binary bits transmitted by the physical layer. Each bit in the frame is converted into an electromagnetic signal and propagated through a physical medium (wireless or cable).

The physical layer does not need an address, and the mode of transmission is broadcast.

6.3.6 layer summary

6.4 Internet applications (not expanded)

6.4.1 email

6.4.2 File transfer Protocol

6.4.3 remote login-TELNET

6.4.4 World wide Web

6.4.5 other Internet applications

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.