1. What is a computer network? Computer networks function What?
A: The computer network is the computer technology and communication technology combined with the product, it is a collection of autonomous computers interconnected. Independent means that each computer is independent, there is a further control a computer case. It means interconnecting the two computers to communicate, exchange data, rather than simply connected together by a cable.
Computer network features are: resource and data sharing, exchange of information, entertainment, and distributed computing.
3 . Packet-switched principles outlined
A: When a host has a message to send to another host, the message is first divided into several small blocks (for small message, may not be divided), add some control information (including receiving each data block front address of the party), which make up the header information. Header and data together form a group. A message may be divided into several groups. These packets are sequentially transmitted to the side of the packet switch connected thereto, the received packet switch the packet into the cache, in accordance with the control information packet header, sequentially forwards each packet, the packet will be passed to the next packet switch, it such as letters transmitted as packets passed down step by step, the final destination. When a message composed of all the packets arrive at the destination are then combined into the original message.
2-3 short answer
1. What is a network protocol? Network protocol consists of which of these components?
Communication network protocols are some rules in order to achieve both the communication and agreed. Specifically, the network protocol may be understood as being composed of three parts:
(1) Syntax: both control and data exchange format information communication.
(2) semantics: the meaning of each portion of the control information and the data represents.
(3) Timing: How to initiate communication; after receiving the data, what to do next.
4 . Description of the OSI reference model and TCP / IP Reference Model
OSI reference model, is divided into seven layers, from top to bottom for the application layer, presentation layer, session layer, transport layer, network layer, data link layer and physical layer.
Directly to the user application layer is used to achieve a particular application. The presentation layer is used for data format conversion of the application layer, the application layer data compression, decompression, encryption, decryption. Session layer used to manage the session, including the establishment of a session, maintenance and closure. Transport layer provides end to end communication between two processes (end to end) communications services. The main function of the network layer for routing packets. Be the interval between the sender and the recipient of the action of the other network, the network layer packet is to find a suitable route, such that the packets can reach the destination. Task data link layer frame is transmitted between adjacent nodes. The physical layer is the task of the bit 0 is transmitted from one end to the other end of the physical link.
TCP / IP model consists of four layers, from top to bottom is the application layer, transport layer, network interface layer and the Internet layer.
Application layer of the OSI reference model and the TCP / IP model is similar to the particular application including, corresponding to a number of application layer protocols, such as HTTP, FTP, SMTP, POP3 and the like. The transport layer also provides end to end communication services, transport layer TCP / IP system contains two protocols, TCP and UDP protocols. TCP protocol provides reliable end to end communications. UDP protocol is relatively simple, can not guarantee the reliability of the data, but more efficient. Internet layer corresponds to the network layer of the OSI reference model, it is the responsibility of the transport layer to its data to the destination, the middle may span multiple networks, the Internet layer to find the right path for the data. Internet layer is a key part of the TCP / IP protocol architecture, the main protocol is IP protocol. The network interface layer corresponds to the data link layer and the physical layer of the OSI reference model. TCP / IP model does not explicitly describe the network interface layer, but the use of a protocol to be noted that the host connected to the particular network, IP packets can be transmitted.
3-3 short answer
8. What is DNS, what is its role?
DNS refers to the Domain Name System. The Domain Name System is a distributed global database, there are many domain name servers, these domain name servers are interconnected to form a distributed system, mainly to complete the domain names into IP addresses work.
9 . Description of DHCP works
(1) client starts, the network broadcasts a DHCP Discover packet, looking for a DHCP server.
After (2) DHCP server receives DHCP Discover, the client sends a DHCP Offer packet to provide IP addresses to clients, subnet mask, default gateway information, domain name server.
(3) The client may receive one or more DHCP Offer packet (there may be multiple DHCP servers in the network), the client choose a DHCP server, and broadcast DHCP Request packet. The DHCP server must be specified to be selected in the DHCP Request packet.
(4) after being selected server receives the DHCP Request, response DHCP Ack messages, address allocation.
4-3 short answer
4. Description of TCP flow control principle.
The main window using the TCP header field to perform flow control. Window field in the header is used to notify each other of their current TCP receive buffer size, data transfer, the receive buffer will change at any time, when sending TCP acknowledgment or send data to each other, will now receive buffer case window into the header field, transmitted to the other party, the other party is received, the transmission data size not exceed the value of the window field, thus ensuring the receiver buffer does not overflow, TCP implements flow control so.
5. TCP is how to ensure the reliability of data transmission?
Mainly TCP acknowledgment and retransmission mechanisms to achieve reliable data transmission. Each party is to be transmitted TCP data number, the number is in bytes. Send bytes starting number is determined in the process of establishing the connection, and the other side has been informed by the three-way handshake.
If you receive the correct number of data, then it would send an acknowledgment to each other. After receiving confirmation, you can continue to send back data. After issuing a TCP segment, it starts a timer. If the timer expires but the confirmation has not come, then retransmits the first message.
If the data is lost midway, then the sender will not receive confirmation, it will retransmit timer expires.
If the data is an error, the receiver discards the packet, the sender will not receive the same acknowledgment, retransmission timer will timeout.
If confirmed lost, because the sender will retransmit timer timeout message, so the recipient will receive duplicate messages, the receiver discards duplicate packets, and sends an acknowledgment.
For the segment of TCP packets arrive out of order, TCP can buffer these packets out of order, which is lined uploaded to the application layer in the subsequent.
6. Why congestion control? How TCP congestion control?
When the influx of large amounts of data networks, network router may be partially due to overload forced to discard some packets congestion. At the same time, the sender will retransmit timer expires because those packets are discarded router, causing more serious congestion, the entire network to a standstill. Accordingly, the transmitting end to perform congestion control.
TCP believes that if there is a packet loss, make sure there is no time to come back, congestion occurs. A TCP connection maintained across a variable, called congestion window, for limiting the size and rate of injection TCP data to the network. When the good condition of the network, each acknowledgment can come back in time, the congestion window is increased, increasing network data injection; if the congestion is generated, there is no confirmation back in time, the congestion window is reduced to take the initiative to ease the pressure network .
After just establish a TCP connection, the TCP congestion window size is the maximum packet length MSS. The beginning of each receive a confirmation, the congestion window size is increased by one MSS, the size of the congestion window is 1,2,4,8, ... increasing exponentially, that is, each through a RTT, the size of the congestion window to double. This process is called slow start.
When the value of the congestion window reaches a certain large, the transmission rate is more than the capacity of the network is received, then the packet loss arises. Once packet loss, TCP will reduce the congestion window. If confirmed several consecutive packets are not received, indicating that the network congestion situation is more serious, then quickly TCP congestion window size is reduced to 1, and enters the slow start-up phase; while also setting a threshold value, this value is half the size of the congestion window when congestion occurs. After the slow start to exponentially increase this threshold value, an increase of the congestion window becomes linear growth. After that within a RTT regardless of how many received confirmation, an increase of only one MSS, until packet loss occurs again. If TCP received the acknowledgment of several consecutive identical, TCP congestion window size will be reduced by half, while increasing the congestion window into a linear growth in a RTT, increased congestion window at most only a MSS, until again packet loss occurs.