Method and device for determining network congestion

Abstract

The embodiment of the present invention discloses a method and device for determining network congestion. The method includes: calculating the transmission loss delay caused by the transmission medium and the network throughput caused by equipment processing during the process of transmitting the message between the source end and the destination end Delay: According to the transmission loss delay and the network throughput delay, calculate the source-destination delay standard value; if the current network average delay is greater than the source-destination delay standard value, it is determined that the current network is congested. The device includes: a delay calculation module, a delay standard value calculation module and a congestion determination module. In the embodiment of the present invention, by calculating the transmission loss delay and network throughput delay, the source-destination delay standard value is further calculated, and the source-destination delay standard value can objectively and quickly judge whether there is congestion between the source end and the destination end , without relying on the experience and ability of network engineers.

Description

technical field [0001] The embodiments of the present invention relate to the field of communication technologies, and in particular to a method and device for determining network congestion. Background technique [0002] With the widespread introduction of cloud computing and big data in the industry, these new technologies put forward higher requirements for network bandwidth and delay. Since the DCN network is usually based on the TCP / IP protocol, the protocol itself has inherently insufficient flow control capabilities, and it is easy to Node congestion occurs, manifested as delays, TCP retransmissions, timeouts, and even discards, while at the business level, it manifests as slow business transactions, and generally has characteristics such as continuity and business perception lag. Considering that in the development process of DCN network of most companies, the introduction of multiple manufacturers has problems such as various hardware brands, different support featu...

AI Technical Summary

Problems solved by technology

[0002] With the widespread introduction of cloud computing and big data in the industry, these new technologies put forward higher requirements for network bandwidth and delay. Since the DCN network is usually based on the TCP / IP protocol, the protocol itself has inherently insufficient flow control capabilities, and it is easy to Node congestion occurs, manifested as delay, TCP retransmission, timeout or even discarding, and at the business level, it manifests as slow business transactions, and generally has the characteristics of persistence and business perception lag

Considering that in the process of DCN network development in most companies, the introduction of multiple manufacturers has problems such as various hardware brands, different support features, and private protocols, it is difficult to conduct global analysis through data flow sampling technologies such as Sflow and Netflow

[0003] At present, when network engineers detect and deal with congestion, they usually combine alarms and manual investigations, and complete the process through the following processes: receive alarms such as port bandwidth exceeding the threshold, log in to the device to confirm information (logs, routing tables, port status, etc.); Investigate a large number of port traffic monitoring chart information, check the ip flow information of the core equipment, and the positioning success rate is low; through manual investigation, communicate whether there is a large traffic timing task in the source-destination center of the congested link; detect the network at the site where the congested link device is located Data packets, locate the source and destination IP to limit

[0004] In the process of realizing the embodiment of the present invention, the inventors found that the existing method involves many network nodes and lines due to the global nature of network links, and it is difficult to judge congestion. And indirect means such as locating the source and destination IP to judge network congestion. Congestion judgment lacks objective standards and relies heavily on the experience and ability of network engineers. At the same time, the process of congestion judgment involves alarm information analysis, manual troubleshooting and a lot of communication work, so there are The entire processing process is too long and depends on the experience and ability of engineers. The analysis and judgment time is too long, resulting in too long processing time for the entire congestion, and even the situation where the problem has not been located even after the obvious business impact

Images