Transmission control protocol tcp data packet transmission method, transmission device and system

Abstract

The embodiment of the invention discloses a transmission control protocol TCP data packet transmission method, transmission device and system. The sending method includes: obtaining a first round-trip delay for sending a TCP data packet in the network, and determining a second round-trip delay; if the first round-trip delay is greater than the second round-trip delay, then determine with a first algorithm The congestion window of the first congestion window is used as the first congestion window; if the first round-trip delay is less than or equal to the second round-trip delay, the congestion window determined by the second algorithm is used as the first congestion window; the TCP data Packets are sent with the first congestion window. Through the technical solution disclosed in the present invention, the current congestion window obtained from the first round-trip time delay is increased to the first congestion window in one step, which can better meet the throughput requirement of the business and utilize the network bandwidth more effectively.

Description

technical field [0001] The embodiments of the present invention relate to the field of network technologies, and in particular to a transmission control protocol TCP data packet transmission method, transmission device and system. Background technique [0002] Network congestion (congestion) refers to the fact that there are too many packets transmitted in the network, but the resources of the storage and forwarding nodes in the network are limited, resulting in a decrease in network transmission performance. When network congestion occurs, there will generally be data loss, increased delay, and decreased throughput. When network congestion is severe, it will lead to congestion collapse. With the rapid increase in business volume of high-throughput applications (such as online audio and video playback applications), the amount of data to be transmitted through the network has also increased sharply, requiring the network to maintain high throughput. If the congestion control...

AI Technical Summary

Problems solved by technology

For the Reno algorithm, when a TCP packet loss is detected, the congestion window is immediately reduced to half the size. In the congestion recovery phase, the congestion window is increased by 1MSS after the RTT of data transmission in each round of congestion window (that is, the increase is 1MSS) , it takes a long time to restore the congestion window from half the size when the TCP packet is lost; taking the network with a network bandwidth of 100 megabytes and a delay of 100 milliseconds as an example, when the throughput rate is close to the network bandwidth, the congestion window value is about 863MSS, Reno The algorithm needs 431 rounds of RTT to restore the congestion window from half the size when the TCP packet is lost, which takes about 43.1 seconds

Compared with the Reno algorithm, the CUBIC algorithm has been improved in terms of congestion window growth. The CUBIC algorithm will record the congestion window when the TCP packet is lost, and increase the window in an exponential manner similar to slow start when the recorded congestion window is not reached. When the congestion window is recorded, the growth step of the congestion window is greatly reduced. After a period of time, the growth step of the congestion window is readjusted to an approximate exponential rapid growth. If it is only maintained for this period of time by chance, the CUBIC algorithm is still After a period of time, the rapid growth of the congestion window will inevitably lead to the loss of more TCP packets when the network is congested again, causing further deterioration of the network condition

[0006] Therefore, the above two algorithms (including the Reno algorithm and the CUBIC algorithm) both have the following disadvantages in TCP congestion control: the congestion window is increased according to a preset fixed value, and the current good network bandwidth cannot be effectively used, and even when adjusting the congestion window It is possible to make an adjustment strategy that is diametrically opposed to the actual network situation, which affects the throughput requirements of the application

Images