This is the sixth day of my participation in Gwen Challenge
Transport layer communication Overview
Process of communication
At the IP layer, two hosts communicate with each other
From the transport layer, the real endpoint of communication is not the host but the process within the host
Host A communicates with host B by means that A program running on host A communicates with another program running on host B
Reliable channel
When the transport layer uses connection-oriented TCP, it is equivalent to a reliable channel
Unreliable channel
When the transport layer uses connectionless UDP, it is equivalent to an unreliable channel
| UDP | TCP |
|---|---|
| Connectionless protocol to provide connectionless services; The data unit it transmits is UDP message or user datagram. Support unicast, multicast, broadcast; Failure to provide reliable delivery; Simple and suitable for many applications. | Connection-oriented protocols that provide connection-oriented services; The data unit it transmits is TCP packets. Support unicast, do not support multicast and broadcast; Providing reliable services; Complex for most applications |
UDP protocol is used
- DNS – Domain name resolution protocol
- DHCP – Dynamic Host configuration protocol
- RIP – Routing selection protocol
TCP protocol is used
- HTTP – Hypertext Transfer protocol
- SMTP – Electronic Network transmission protocol
- FTP – File transfer protocol
Software ports and hardware ports
The abstract protocol port between the protocol stacks is the software port
Ports on routers or switches are hardware ports
A hardware port is an interface through which different hardware devices interact
Software port is an address where various protocol processes of the application layer interact with transport entities
Ports are marked with a 16-bit port that allows up to 65535 different port numbers.
Ports are used to identify processes in the application layer of the computer.
The same port number on different computers is irrelevant
Therefore, two processes in the computer to communicate with each other, both to know each other’s IP address, find each other’s computer, and to know each other’s port number, find each other’s computer application process.
Common familiar port number
53: DNS
25: SMTP
20/21: FTP
23: Telnet
80: HTTP
443: HTTPS
User datagram protocol UDP
The function of UDP
Reuse and reuse capabilities
Error detection capabilities
The characteristics of UDP
-
No connection, no need to establish a connection before sending data, reducing the cost and delay before sending data
-
Best effort delivery does not guarantee reliable delivery
-
In the packet oriented mode, UDP delivers a complete packet at a time without merging or splitting the packets transmitted from the application layer
-
Without congestion control, network congestion does not reduce the sending rate of the source host
-
Supports one-to-one, one-to-many, many-to-one and many-to-many interactive communication
-
The first overhead is small, only 8 bytes
Transmission control protocol TCP
The characteristics of TCP
- connection-oriented
- Ensure reliable delivery
- Only one to one support
- Provides full duplex communication
- Word oriented stream
Socket = (IP address: port number)
The socket = (192.168.1.10:8080)
TCP connection, the relationship between IP address and socket
A TCP connection is an abstraction provided by protocol software
The endpoint of a TCP connection is an abstract socket, i.e. (IP address: port number)
The same IP address can have multiple TCP connections
The same port number can also appear on multiple TCP connections
How reliable transmission works
Stop waiting protocol
Stop waiting. Stop sending each packet and wait for confirmation. Send the next packet after receiving confirmation.
Number. Number each packet and acknowledgement sent
Automatic retransmission request. Sender sets a timeout timer for each packet sent. If the timeout timer expires, the sender automatically retransmits the packet.
Simple, but channel utilization is low
Continuous ARQ protocol
The sender can send more than one packet at a time
Use the sliding window protocol to control the number and number of packets that both sender and receiver can send and receive
With each acknowledgement, the sender slides the send window forward
The recipient generally adopts the cumulative confirmation method
The rollback N method is used for retransmission
The header format of the TCP packet segment
Source port – 2 bytes.
Destination port – 2 bytes.
Ordinal field – 4 bytes. Each byte in the TCP connection data stream is numbered. The value of the ordinal field refers to the ordinal number of the first byte of the data sent in the paragraph.
Acknowledgement number field – The sequence number of the first byte of the next segment of data expected to be received from the other party
Data offset – up to 4 bits. Indicates the distance between the start of the TCP packet segment and the start of the TCP packet segment.
Reserved field – takes up 6 bits. Reserved for future use, the current value is 0
Emergency URG – A value of 1 indicates that the emergency pointer field is valid. It tells the system that there is urgent data in this message segment and to transmit it as soon as possible
Acknowledgment ACK – The acknowledgment number field is valid only when it is 1. When the value is 0, the confirmation number is invalid
Push PSH-receive TCP receives a packet segment when PSH=1 and delivers it to the receiving application process as soon as possible, rather than waiting until the entire cache is filled up.
Reset RST – When the value is 1, a serious TCP connection error occurs and the connection must be released and then re-established
Synchronous SYN – A value of 1 indicates a connection request or connection accept message.
Terminate the FIN – Used to release a connection. When the value is 1, it indicates that the data on the sender end of the packet segment is sent and the transport connection needs to be released.
Window field – 2 bytes. The basis on which to set the sending window.
Check and – take 2 bytes. The scope of validation and field validation includes the header and data parts. When calculating the checksum, a 12-byte dummy header is added to the front of the TCP packet segment.
Emergency pointer field – 16 bits. Indicates the number of bytes of emergency data in the column.
Option field – Variable length. TCP initially provides only one option, the maximum packet segment length (MSS).
Fill field – this is so that the header length is a multiple of 4 bytes
The implementation of TCP reliable transmission
The send window indicates that the data in the window can be continuously sent out without receiving an acknowledgement
Receive window: Only data that falls into the window can be received
Send cache: The sender’s application process writes the byte stream to the TCP send cache
The send window is usually only part of the send cache
The role of the send cache
- The sending application sends data to the sender TCP to be sent
- TCP has sent data but has not received an acknowledgement
Receive cache: The receiving application process reads the byte stream from TCP’s receive cache
The role of the receive cache
- Data that arrives sequentially but has not yet been read by the receiving application
- Data arriving out of order
TCP traffic control
Ensure that the sending rate of the sender is not too fast, so that the receiver can receive data in time, and do not use network congestion
The sliding window mechanism is used to realize flow control on TCP connection
Controls the sending time of TCP packets
- The first mechanism is that TCP maintains a variable equal to the maximum message segment length MSS. As long as the data stored in the cache reaches MSS bytes, it is assembled into a TCP packet segment and sent out
- The second mechanism is that the application process of the sender specifies the segment of packets to be sent, that is, the push operation supported by TCP
- The third mechanism is to load the existing cache data into a message segment (length less than MSS) and send it when a timer expires.
TCP congestion control
Factors causing network congestion
- The capacity of the point cache is too small
- The link capacity is insufficient. Procedure
- The processing rate of the processor is too slow
- Congestion itself can exacerbate congestion
Difference between congestion control and flow control
Congestion control: Prevents too much data from being injected into the network to prevent overload of routers or links on the network. Is a global process that involves all factors associated with degrading network transport performance
Flow control: inhibit the rate at which the sender sends data so that the receiver can receive data in time. It’s point-to-point traffic control, it’s an end-to-end problem
Open loop control and closed loop control
Open loop control: in the design of the network, considerate in advance, strive to avoid congestion
Closed-loop control: based on the concept of feedback loop; Take control measures according to the current network running status.
Several measures belonging to closed-loop control
- Monitor network systems to detect when and where congestion occurs
- Communicate congestion information to a place where action can be taken
- Adjust the operation of the network system to solve problems
Monitor network congestion indicators
- Percentage of packets discarded due to lack of cache space
- Average queue length
- Number of packets for timeout retransmission
- Average packet delay
- Standard deviation of packet delay
Ideas to solve congestion
Add available network resources
Reduce user demand for resources
TCP congestion control method
TCP adopts window-based congestion control.
The TCP sender maintains a congestion window
The sender uses the congestion window to adjust the amount of data to be sent according to the congestion of the network
The size of the sending window depends not only on the receiving window but also on the congestion of the network
True send window value = Min(receiver window value, congestion window value)
Principles for controlling congestion Windows
As long as there is no congestion on the network, the congestion window can be larger to send more packets and improve network utilization
If congestion occurs or is likely to occur on the network, reduce the number of packets injected into the network by reducing the congestion window
Congestion judgment
The retransmission timer timed out. – The network is congested
Received three duplicate ACKS – indicates possible network congestion
TCP congestion control algorithm
Slow start
Congestion avoidance
Fast retransmission
Fast recovery
TCP transport connection management
The three stages of a TCP connection
Connection is established
Data transfer
Connection release
Client-server approach
The TCP connection is established in client server mode
The adoption process that initiates the connection is called the client
An application process that passively waits for a connection to be established is called a server
Three handshakes to establish a connection
- The client first sends a packet with the SYN flag to the server
- After receiving the packet, the server sends back a packet with SYN and ACK flags to confirm receipt
- The client sends back a packet with an ACK flag, which means I know you received it
Four waves to release the connection
- The client sends a FIN to indicate that I’m closing the connection
- When the server receives the FIN from the client, it sends an ACK to the client indicating that I know you want to close the connection
- The server sends a FIN to indicate that I am also closing the connection
- When the client receives the FIN from the server, it sends an ACK to the server, indicating that I know you want to close the connection, and the connection is released